4 Seconds Loading Time Is Maximum For Websurfers

nieske writes "Of course we all want webpages to load as fast as possible, but now research has finally shown it: four seconds loading time is the maximum threshold for websurfers. Akamai and JupiterResearch have conducted a study among 1,000 online shoppers and have found, among other results, that one third of respondents have, at one point, left a shopping website because of the overall 'poor experience.' 75% of them do not intend ever to come back to this website again. Online shopper loyalty also increases as loading time of webpages decreases. Will this study finally show developers of shopping websites the importance of the performance of their websites?"

Comments

tabs

[Kuciwalker]

It takes a lot longer than 4 seconds for a Slashdot story to load, particularly with the new AJAXy discussion system. I usually open up several things at a time in different tabs, which decreases the average loading time since I can read one thing while another loads. What browser were these people using?

Re:tabs

[From A Far Away Land]

There ought to be a lot more consideration given to dialup users this study finds. Bling might draw people back to the site, but only if it takes a second to load.

Re:tabs

[Agent00Wang]

I'm sure that a study of "1000 shoppers" had to include mostly IE users.

Re:tabs

[dabadab]

For me /. is real fast - I use lynx :-)

Re:tabs

[Gr8Apes]

Remove Slashdotter plugin....

Get a faster pipe.

or, just load multiple stories and deal with it.

Personally, I like the new feature, because I no longer have to reload an entire page to get to 1 comment.

Re:tabs

[Overzeetop]

You must not post often. I usually post, then go get a cup of coffee or *gasp* - go do work - while the page loads.

Re:tabs

[Total_Wimp]

I usually open up several things at a time in different tabs, which decreases the average loading time since I can read one thing while another loads. What browser were these people using?

1. The average user goes about doing things in an average way. That tells me they use IE, and even if it was IE7 (which isn't average quite yet) that they would do all their shopping on a single site from a single tab. Telling them there's a workaround is nice, but in the end is about as likely to affect the average user as Firefox is. (no flames please; firefox is not used by the "average" user and you know it)

2. Why should I have to work around things in order to get a good experience? Why should someone who wants me to use their site expect me to jumpt through hoops to do so?

3. One of the focuses of the article as commerce. You can't really click through the checkout process in multiple tabs.

4. Some of this is easy to fix. Why should I have to click through 3-4 different page loads to buy stuff? Why not just put everything on one page? Why load pages up with extra crap that has to download? There are ways to do fast loading sites, as long as it's a priority for those building them.

TW

Re:tabs

[Ash Vince]

The thing to remember is that most of the rest of the world have better internet connections than the US do now so pages load quicker for us.

Here in the UK slashdot is near instantaneous over 24 Megabit ADSL. 2 Megabit ADSL accounts are given away for free in the UK now with most phone connections. The slowest account people actually pay money for is 8 Megabit ADSL.

As for all the people saying they still use dialup, why? Here you can get better net connections than 56kbit using a mobile phone (3G - UMTS).

To me the idea of waiting 4 seconds for a page to load is monsterous, expecially if the next page I clicked took just as long even though half the images were already cached.

Re:tabs

[Kuciwalker]

I can't live without Slashdotter, I'm on a blazing fast university connection, and I do the latter. I wasn't even complaining, I was just observing that > 4 seconds loading time can be acceptable if there's a way to amortize it over multiple page loads.

Re:tabs

[maxwell demon]

I guess your employer denies you a browser with tabs in order to have you get at least some work done. :-)

Re:tabs

[Psiren]

Here in the UK slashdot is near instantaneous over 24 Megabit ADSL. 2 Megabit ADSL accounts are given away for free in the UK now with most phone connections. The slowest account people actually pay money for is 8 Megabit ADSL.

What?! What?!!!! If you're referring to the offers from people like Carphone Warehouse, it's far from free. Virtually everyone is still paying for their ADSL. And I'm paying for mine, which is currently running at 512K, because thats all the line supports. 8Mbit/s is just a dream until BT upgrade to their 21 Century Network (yeah, good luck with that). Maybe if you're living in London you have more options open to you, but the rest of the country is still far behind that.

And FYI, I'm on a 100mbit JANET connection and Slashdot is far from instantaneous. Personally, I think you're talking out of your arse.

Re:tabs

[Hoi Polloi]

"4. Some of this is easy to fix. Why should I have to click through 3-4 different page loads to buy stuff? Why not just put everything on one page? Why load pages up with extra crap that has to download? There are ways to do fast loading sites, as long as it's a priority for those building them."

I want plain text descriptions, PRICES, and thumbnails. If I want big pretty pictures I can click on the item for more info. I also want more items per list, not the 10 or so that some places have. It makes scanning the selection impossible. Don't make me go to the checkout to see the price either. If you don't show it up front it discourages me from considering it.

Re:tabs

[aplusjimages]

But you are looking for news, which I'm sure people have more patients for than when they do for shopping. I've sat on sites looking for news that took about 30 seconds to load, but I don't think I would have that type of patients for shopping. Like real shopping I want to get in and get out.

Re:tabs

[Kijori]

News? On Slashdot? You must be new here!

Re:tabs

[Phisbut]

I'm sure that a study of "1000 shoppers" had to include mostly IE users.

Even when I'm using Firefox, I don't use tabs to shop online. Tabs are useful on slashdot, and to view online documentation, but not to shop. You can't open several tabs at the same time, one to view the description of the item, one to add the item to your cart, one to fill the shipping address, one to fill the credit card information, etc.

Online shopping is a linear process and tabs can't help that.

Re:tabs

[init100]

It takes a lot longer than 4 seconds for a Slashdot story to load, particularly with the new AJAXy discussion system.

Slashdot does not use AJAX, just some Javascript, for the new discussion system. In essence, the only thing it does is hide all posts until you click on its header to expand it. The posts are there anyway, loaded along with the rest of the page. That way, it uses about the bandwith of the nested option, while presenting it as a dynamic threaded view. If they used AJAX, it would (probably) send an asynchronous query to the Slashdot servers asking for precisely the post you try to open. That does not happen, I checked tcpdump myself.

Had they really used AJAX, a comment thread might have been a lot quicker to load initially, but slightly slower loading each post.

Re:tabs

[compro01]

As for all the people saying they still use dialup, why?

because internet service in my area is practically a monopoly. the phone company refuses to run DSL-capable line the 3000 ft from the highway into the village and install the nessesary hardware. they're making buckets of cash off of raping us for our dial up ($60 a month for 180 hours/month of net time, plus the "unlimited long distance" required to be able to get that 180/month plan), not to mention the overage charges they pull if you go over the 180/month, wheras you can get the cheapie 1.5mbps DSL for $15/month in town.

the only other options in the area are wireless high speed (similar to Wi-max), but for that you need to buy the antenna and gear upfront ($250) and satalite internet, with is not an option as a. our satalite provider doesn't do internet and b. it would be useless in any case due to the lag from satalite (stupid laws of physics).

Re:tabs

[wgaryhas]

I guess you have never wanted to have the descriptions of several items up and compare them? In my experience, it is only linear when you are ready to pay; searching and comparing products benefits greatly from tabs.

Re:tabs

[NotTheNickIWanted]

You can't open several tabs at the same time, one to view the description of the item, one to add the item to your cart, one to fill the shipping address, one to fill the credit card information, etc.

Online shopping is a linear process and tabs can't help that.

I disagree.

I routinely use tabs while online shopping, most commonly to open product descriptions in a new tab will leaving a product index unmolested in another. Additionally I do not recall any of the sites from which I have made purchases getting confused if I open a new tab to view a product description, add that item to my "cart" from the extra tab, then close the tab and continue browsing from the previously loaded index page.

I agree that the checkout is in most cases a linear process, but examining items and selecting them for purchase does not have to be.

Re:tabs

[LifesABeach]

It is amazing to see what happens when one does not take the time to read history. In the early 1970's, Delta Airlines commissioned a study about the patience of a customer calling to make a reservation for a flight. The result was 3 Seconds. Computers back then had to have that much speed to be acceptable for a global airline reservation system. I have been using that rule as a measure for user patience; I thought everyone did.

Re:tabs

[Overzeetop]

Actually, I so use FF, and often have to open up a few tabs to make my comments in the time I use for reading /. (damned limits on posting).

And, for the record, I own the company. I can surf /. as long as I want, as long as I don't mind making no money.

Re:tabs

[Bromskloss]

It takes a lot longer than 4 seconds for a Slashdot story to load, particularly with the new AJAXy discussion system. I usually open up several things at a time in different tabs, which decreases the average loading time since I can read one thing while another loads. What browser were these people using?

I'm all with you. Actually, I was just about to post a comment mentioning Slashdot load time and the advantage of tabs -- then I saw yours. Take this as an informal up-moderation. (I have no mod points.)

Re:tabs

[Ash Vince]

I was reffering to the offers from Carphone Warehouse, Orange, SKY, NTL and a few others. They are usually free if you sign up to whatever other service they offer for the same price they used to charge for that on its own.

Personally I know a few people using the talktalk offer from Carphoneware House.

By the way, I live in Manchester (Moss Side) and we have 24 megabit ADSL here. Not exactly the most expensive area but it is close to the city centre though, albieit on a different exchange.

And how come you are still on 512K? I didn't think anyone sold those accounts anymore as even the crappy old BT network supports 1Meg ADSL. Try speaking to you ISP and they might give a you a free upgrade if you threaten to switch.

The reason it is so slow on Janet (UK Accademic Network) is because of all the other students downloading Pr0n

Re:tabs

[cortana]

It sucks living in the sticks; however why do you think the phone company has a duty to sell your DSL services at a loss?

Re:tabs

[UnBlinking]

Meanwhile, Akamai chokes and gags for half a minute before announcing:
http://www.akamai.com/html/misc/requirements.html
"To view the advanced features and functionality of akamai.com, Adobe Flash Player 8 is required."
...thus assuring me that the page (presuming I would still waste time there) would require at least a minute to load, itself.
Good work, Akamai!

Re:tabs

[chaoticgeek]

I use tabs when shopping online. Add everything to my cart that I want then fill out all the junk they need. Its fairly simple, I do this on TG.com, Amazon.com, many computer supply stores, ect.

Re:tabs

[compro01]

not really that they have a "duty" to, but i've really been going "WTF?" since they extended said lines to a similar town (population of 100 vs. population of 175) 15 miles down said highway. i mean, they extend the lines 15 miles to another town, but can't extend them less than 1 to mine.

Re:tabs

[cortana]

Perhaps all 100 of those people had registered an interest in signing up for ADSL service. Try looking your phone exchange (and theirs!) up at The Broadband Resource. See what your 'ADSL prereg' status is.

Re:tabs

[tigerflag]

"As for all the people saying they still use dialup, why?"

1. Because it's possible to get 24/7 dialup for $5-8 per month, while even the slowest DSL costs about $35 for the line and the ISP. Believe it or not, a lot of people can't afford more than dialup.

2. Because about half the people in the USA can't get broadband even if they can afford it. I live in a city of about 150,000 people. Five miles down the street I could not get cable or DSL. I moved down the road a bit and now I have DSL.

I did quite allright with dialup. If I didn't have an ecommerce business I would have stayed with it. Using a hosts file cuts down on 99% of the advertising that slows page loading, and my speeds were often faster than my office's cable setup.

Re:tabs

[stunt_penguin]

Pfft, shopping online is one of the areas where tabs are most useful, for flicking between 2-3 sites, or a number of offers on different sites.
I think it's fair to say that there are very few things that are done online that are strictly linear.

Re:tabs

[webwidejosh]

Odd, most my pages load faster in IE. I primarily use Firefox but often use both. I notice that from the command line, IE loads and displays a webpage faster.

Re:tabs

[Psiren]

I was reffering to the offers from Carphone Warehouse, Orange, SKY, NTL and a few others. They are usually free if you sign up to whatever other service they offer for the same price they used to charge for that on its own.

See here for some of the problems Carphone Warehouse are having. The marketing is good, but when you dig a little deeper things are not always as they appear. My Dad was told he could receive free Sky broadband, but when he went to order, they wanted 17.99 a month for it. Seems the free broadband wasn't in his area yet, not that that stopped them from encouraging him to order it.

And how come you are still on 512K? I didn't think anyone sold those accounts anymore as even the crappy old BT network supports 1Meg ADSL. Try speaking to you ISP and they might give a you a free upgrade if you threaten to switch.

I was on a fixed rate 1mbit line. I switched to the rate adaptive and it's dropped to around 512K. I'm in the process of trying to get that sorted, but I'm not ever going to get anything more than 1mbit with current technology. No LLU here yet.

The reason it is so slow on Janet (UK Accademic Network) is because of all the other students downloading Pr0n

No, we're not even close to saturating our 100 mbit line. I admin the network here so I keep a pretty close eye on traffic.

Re:tabs

[Chandon Seldon]

What's the upload speed to go with those downstream speeds (24, 8, 2)?

I mean, sure, begin able to have more Google Video pages open at the same time is neat, but most of the interesting applications of decent bandwidth (like, say, sending someone a file) are constrained by the lower of upload and download speeds.

Re:tabs

[rapidweather]

I came up with the name "rapidweather" for my site(s) because I designed most of the pages in the day when 14.4 dial-up modems were being used. I have weather pages, and got tired of waiting for the TV weatherman to finally appear some 20 minutes or so in the newscast, with that "killer tornado on main street" that he promised to tell us about at the beginning of the show. Most TV weathermen vary the actual time of their appearance onstage, to keep you guessing, and expose the viewer to the maximum amount of useless commercials.
The internet came along, and I got busy and made the weather pages you see on rapidweather.com
My weather maps page is designed to load quickly, although I went overboard and provided too many weather map links. Idea was to provide maps, and backup maps, etc.
Very popular page when something like Katrina comes along.
I have watched usatoday.com go from a fast loading page to the one we get now, although it is not really all that bad. One page I provide is this one, designed to give the user a nice selection of decent pages, all tested and approved in their loading speed and usefulness. One might ask why I don't host all of my pages on rapidweather.com. Well, I like to keep some of them right were they are, so the "return visitors" can find them. Also, geocities is reliable, and does provide me with some statistics about the visitors to the pages, and I find that interesting to review from time to time. If I get a page that no one comes to, I might close it, and redirect to something similar in my collection.
In my knoppix remaster, (see screenshots link below), I provide a local copy of "web.html" as a built-in start page for all three browsers, Opera, Flock and Firefox. Start up the browser, and that page appears, ready to go to work. The idea behind "web.html" is to keep the user out of a "walled garden" like they get with using msn.com or aol.com for a "home page". I often take a look at what the "webby awards" has chosen, and see if any of those sites will be something I will want to put on "web.html".

-- Rapidweather

Re:tabs

[DeathElk]

Why still use dial up? Well the Australian rural coastal town in which I live has only had ADSL for 2 years or so. Before that it was ISDN (expensive, slow and only available close to exchange), 2 way satellite (very expensive) or dial up. My dad lives 90 minutes north west of here. Dial up at 14400Kbps OR 2 way satellite for him. He opted for 2 way sat, which I think offers 128/64 Kbps for $120 per month (although when I had 2 way sat I could get >400Kbps ftp to aarnet.edu.au when using a linux usb driver to the transciever). NO static IPs and shocking, teeth grinding latency thrown in free.

Here, I can STILL only get max 1.5Mb ADSL at nearly $100 per month. Consider yourself very lucky. Dial up is still a necessity for many users.

Re:tabs

[compro01]

that site is not useful for me as it has no clue where i live. the only method i could find a match via was using the village name, which points to some town in the UK, which is not where i live. i live in Saskatchewan, Canada. not Nottinghamshire, UK.

Re:tabs

[mojine]

Set your user preferences to simple design and low bandwidth - speeds slashdot up greatly. Looks better, too.

Well..

[Rob T Firefly]

Four-seconds-is-the-most-time-I-would-ever-spend-r eading-a-Slashdot-article-or-comment-so-I-ask-all- posters-to-please-make-their-points-quickly-and-su ccinctly-in-small-manageable-doses-ooh-look-there- is-a-shiny-object-on-my-desk-gotta-go.

Who is conducting that study?

[Anonymous Coward]

Note that Akamai has a vested interest in this study. They would like to encourage more businesses to use their technology so that their sites load faster.

I am not saying that the study is biased, but one should at least consider that it is in Akamai's best interest to convince every site owner that they will lose customers if their site is not fast enough.

Re:Who is conducting that study?

[castoridae]

Could be. The figure I remember being a rule of thumb during the .com days was always 8 seconds - but that was a much easier target to hit unless you were doing something stupid (or had to do some really heavy lifting) on the backend.

Re:Who is conducting that study?

[SkunkPussy]

4s seems remarkably high!

I get frustrated by a delay of even 1s.

Re:Who is conducting that study?

[Gr8Apes]

Very true. However, with a decent pipe and proper website design, using Akamai will be slower.

I have worked with Akamai and speed is not a reason to use them, redundancy and capacity are. Akamai also has other issues, such as with websites that may have CMS of some sort running that changes often may have issues with their cache. It's usually a 5-30m (depending upon your customer clout) to get that cleared. Meanwhile, you can have a broken website. Fun fun fun.

Re:Who is conducting that study?

[castoridae]

I guess we're spoiled by broadband now! :-)

Re:Who is conducting that study?

[teh kurisu]

Exactly. I remember that the figure used to be 10s back in the days of dialup, but now that connections are faster we are getting more and more impatient.

What really annoys me are sites where the main text of the page seems to load last. Everything else, like image intensive navigation strips, logo headers, ads (if Camino's ad blocker hasn't caught them) etc. seems to slowly load first and then it's a couple of seconds before the main text appears.

Re:Who is conducting that study?

[Andrewkov]

That's nothing, I couldn't even read your whole post since it was too long.

Re:Who is conducting that study?

[Fozzyuw]

Note that Akamai has a vested interest in this study.

Agreed, also, I would say the old humorous adage of...

Statistics show that 70% of statistics are wrong.

I'd also think that this is more FUD, since of course people are going to say "I'll never return to that website because once it never loaded for me". Of course, if said website was Amazon.com, I'm going to go all in and say that they will most likely return and that they where just unhappy at the time.

Also, what where the survey questions? Where they worded to direct people to answer in a certain way? For example, where they written like this...Would you visit a website again if you had recurring bad experience with it?. Most experts agree that 4 seconds and more is a long time for a website to load and contribute this to a bad user experience. Would you agree?.

Finally, this is a bit of a moot point, but page load times are not always the fault of the web server. Case in point, my wireless router sucks (hey, I got it from best buy for like $20 after rebates, I'm not complaining) and it will often just drop the wireless signal. This makes it appear that when you try to access a website, when you where just accessing another website just fine, that the website takes a long time to load or more often than not, doesn't load at all, making you think there are issues with that site, when there is not.

I'd rather point out the obvious when it comes to the Internet... people expect to find broken sites. It's just the way everyone has grown up with the 'net. Crappy designed sites with stale content and broken links, multiple browsers (IE(win), IE(Mac) , Netscape/Mozilla, Safari) on multiple systems (Mac, Unix/Linux, Windows), security issues, internet issues with their ISP, web servers that go down, etc, etc, etc. People know that the nature of the Internet is volatile. Sure, they'll get angry and might never return if your site is a small 'mom and pop' shop, but when it comes to the 'major' sites such as CNN, MSN, Amazon, eBay, etc; then they will most likely return, even if it takes more than 4-seconds (which it does on CNN and if I remember right, they use Akamai).

Saying 4 second load time will make customers never return is slightly FUD (since it's subjective). Saying a customer won't return if they type their zip code into their address bar, and your site claims it's invalid (when it's not) will probably do the trick. That's something web developers can change. Load times isn't alway in their hands. But that's not to say they shouldn't review their code for efficiency in execution.

Cheers,
Fozzy

Re:Who is conducting that study?

[k12linux]

I have a problem with this kind of statistic when it's based on only 1000 people with no mention of how they were selected. Are they all 3Mbps broadband users? Dial-up users wouldn't even blink at a 4 second load time.

The fact that it was a "survey" and not a study where they actually watched users surf is a sticking point for me too. Most people are not very good estimating time in that range. Did they really leave a site never to return after they had to wait 4 seconds for a page load? Or was it really 10 or 15 seconds?

I agree with you. The fact that Akamai has a vested interest in the results is not encouraging either.

Re:Who is conducting that study?

[trianglman]

I agree with your points. As far as the questions go, if you download the whitepaper http://www.akamai.com/dl/reports/Site_Abandonment_ Final_Report.pdf [may require 'registration'] you will see that the questions aren't poorly written. For example, the main point of this survery, "Which of the following factors are most influential in your decision to continue shopping with an online store where you have shopped in the past? (Select all that apply.)" This question included 13 possible responses including "Site is easy to navigate", "Page loads quickly", "Personalized e-mail offers" and others. The real problem with this report is that it ignores the actual findings like the fact that navigation is most important to users and load times is down at 8.

Re:Who is conducting that study?

[93 Escort Wagon]

That's nothing, I

Next!

Re:Who is conducting that study?

[Fozzyuw]

Thanks for the link! Being at work, I didn't spend any time digging through links. Helps answer one of my questions. =)

Cheers,
Fozzy

No, really?

[Magada]

Let's all have a poll of our own now... how long does your fave shop's site take to load? Would you consider a switch?

Re:No, really?

[gr8whitesavage]

I hope there is a CowboyNeal option.

Re:No, really?

[SleepyHappyDoc]

Um, would you pay for CowboyNeal to come up in less than four seconds?

Re:No, really?

[Alaria Phrozen]

My favorite computer shop NewEgg's website can take as long as it wants. The price is right, the shipping is fast, and they've handled my RMAs very well. It's just.. _the_ computer shop. Who cares if it's slow sometimes?

If I need to wait more than 4 seconds on a website, I'll get up and throw some poptarts in the microwave. That takes about 7 seconds. Way faster than a stupid toaster. Seven seconds is a long time, but if you've ever tried 15 seconds of silence, oh god..

Agreed

[Nos.]

Its not so much that the first page takes 4 seconds, but realizing that probably every other page will take at least that long. With broadband, we've become used to near instantaneous page loads. If a site doesn't load in about 2 seconds, especially sites I frequent, it usually won't load at all.

Great

[UbuntuDupe]

Now that this research is in, I predict that all website designers will realize the futility of flashy designs and instead remake their sites to be more like Craigslist or Google. I'm predicting an end to Flash and Javascript.[/naivete]

Re:Great

[jsebrech]

Now that this research is in, I predict that all website designers will realize the futility of flashy designs and instead remake their sites to be more like Craigslist or Google. I'm predicting an end to Flash and Javascript.

It's amusing that google gets hailed for being a beacon of simplistic web design while most of what they build nowadays are extremely complicated ajax or flash web applications. Just about the only thing that is "light weight" is google's homepage, and that only in classic mode, because the new customized homepage gets heavy real fast.

Secondly, flash and javascript, when used appropriately, make web applications faster because they decrease the amount of data that needs to be transmitted and increase responsiveness. For example, I replaced a web application that used generated bitmaps in simple HTML pages to show vector drawings with one that uses javascript, flash and generated SVG files to do the same, and the amount of data transmitted dropped by a factor 10, as well as the entire application being much more responsive.

Flash and javascript are both a lot more powerful as technologies than HTML with server roundtrips. Because of that they get misused more often. But don't blame the hammer for making it easy to bash your skull in with it.

Re: Amazon.Com clearned this along time ago.

[trdrstv]

Amazon still codes their pages so they come up "fast" on a 28.8 modem. Ebay is the same. Where as some sites are sold by flashy experiances, they are not. They want to keep the barrier of entry low so you buy from them, and the whole process is fast and easy. To do otherwise is simply bad business.

Bullshit

[diegocgteleline.es]

I'm still on dialup, you insensitive clod! (really)

Re:Bullshit

[prolene]

Same here, and on modern websites it takes over a minute to load a website completely. But life goes on.

Re:Bullshit

[thejrwr]

I'm still on dial-up my self, 28k is my max on my lines, and flashy sites i tend to just stay away from them, but i tend to do other things when pages are loading or a script is a downloading

Re:Bullshit

[se7en11]

4 seconds is like lightning-fast on dial-up!

While I do agree for shopping sites 4 seconds (on broadband) is a little long, but I've seen both Amazon.com and TigerDirect.com take well over 4 sec. Even Gmail which I use everyday can take 4 sec or more to fully load up.

Lesson to learn from post: By spare computer to go next to the other 3 in mom's basement to reduce time waiting.

Only four seconds?

[BenoitRen]

Sounds to me like people are getting really impatient these days. I'm willing to wait up to 10 seconds to let a page load, and if it still hasn't, but is busy (instead of connecting again) I load another tab and occupy myself with something else.

However, four seconds sounds accurate for how long to wait until the page -starts- to load. If I have to wait longer than 4 seconds just to connect to a web server, I start to get impatient. If it takes much longer, I'll come back to it later and go do something else.

Re:Only four seconds?

[Rob the Bold]

Sounds to me like people are getting really impatient these days

I think it's experience, not impatience that leads to this behavior. If you get nothing for 4 seconds after trying to open a site, aren't the odds of it working right today going down with each additional second? If nothing has happened after 4 seconds, will 5 help, or 500? At that point, you start to assume the tubes are clogged at the other end.

If a person is shopping or reading the news, long waits between each action can be very irritating. Like turning newspaper or catalog pages with tweezers while wearing gloves.

I call bullshit

[Deagol]

If the hyper-caffeinated, sugar-popping, MTV-watching, blipvert-desensitized ADD kids of today can tolerate the glacially loading site known as MySpace, I'd bet dollars to donuts that the slightly older, credit card-holding demographic of online shoppers gladly tolerate more than 4 seconds on we sites, and do so without much prejudice.

Either the summary is totally off, or this 'research' is total bunk.

Re:I call bullshit

[ScentCone]

I'd bet dollars to donuts that the slightly older, credit card-holding demographic of online shoppers gladly tolerate more than 4 seconds on we sites, and do so without much prejudice

I think it all comes down to what the site is doing, and how readily available another, virtually identical site (or range of merchandise, at similar prices, etc) actually is. The more sites there are that present and transact the same things in essentially the same way, the more that things like raw speed differentiate one from another. The more unique something is (niche merchandise, a blog with a particular perspective), the more patience people will have. Those things are nearly impossible to quantify, and thus you get largely BS, context-less reports like the one being discussed. I think that the larger conclusion ("people are less patient than they used to be") is valid - but pretty hard to nail down, in terms of specific seconds, for specific demographics, on particular platforms, across particular pipes, under certain seasonal circumstances, blah blah blah.

Re:I call bullshit

[Jeff DeMaagd]

I don't think your argument holds much weight.

For one, MySpace loads up less than four seconds for me and I am using DirecWay.

Another is that MySpace is an entertainment site, not a commerce site, and I think that makes a difference in what people will tolerate. For some reason, people want to be there. A lot of people don't want to wait to buy things, it's not entertaining. For some reason, people will often wait an hour in line for entertainment, but only rarely would they be willing to wait an hour to buy something.

This old bag?

[mattwarden]

People are still spending money on stuff like this?

Here's a question: is gmail.com the same as brochureware.com? Would a user visiting a web(2.0)-based application have the same load time expectations as visiting an about page of a company's website?

Of course the answer is no. People with half a brain start to sound like a broken record here when they say "This has no value. It all depends on the site's audience, not a general audience.", but that's because the people behind studies like this never listen.

Re:This old bag?

[Magada]

Erm. No. Most people who want to buy stuff... well, guess what, they want to just buy the stuff they need and be gone and they never ever give a rat's arse if your site is sooperdooper Web 2.0 or XML-enabled or LAMP or SOAP ow whatever l'acronyme du jour is. User expectation is that everything should be instantaneous... and why the heck not?

Re:This old bag?

[mattwarden]

Can you say "tunnel vision"? Since when is the web just a bunch of shopping carts? I've been in web application development for 8 years and I have somehow made a living while only working on one shopping cart in my career. Really, all you've done is further my point that response time expectations depend highly on the content of the site.

Re:This old bag?

[Magada]

Content yes. Application/framework, no.

Re:This old bag?

[mattwarden]

No idea what your comment means, but I'll take a stab and say that you think "content" is "text." It's not, and that's not what I was saying.

Akamai provides content acceleration services

[mattnuzum]

Of course Akamai is going to say that... they're business model revolves around bringing data closer to web surfers in order to speed up busy sites.

That's kind of like two years ago (or so) when RedHat released a whitepaper saying linux has a lower TCO while simultaneously Microsoft released a whitepaper saying windows server has a lower TCO.

The only difference is, there's no one out there selling a service or product that slows down website access to provide a contrasting viewpoint. Well, none except maybe these guys.

Re:Akamai provides content acceleration services

[MtViewGuy]

Self-serving probably, but Akamai does have a point, though. With today more and more users surfing the Internet at broadband speeds, anything that slows down access is going to aggravate users to no end.

This survey was sponsored by...

[MagicM]

How convenient for Akamai. A survey done by them shows that webpages need to load faster or you could lose money. And you wouldn't want that, now would you?

About Akamai
Akamai® is the leading global service provider for accelerating content and business processes online. Thousands of organizations have formed trusted relationships with Akamai (...)

Disturbing...

[rasmusneckelmann]

One of the main reasons why I'm considering upgrading my 233-MHz laptop, is not because it's slow at doing heavy calculations (like Matlab, etc), but because it will soon to be impossible to surf the 'net. Not only are webpages growing larger and larger kB-wise, but they're also using increasingly more CPU resources when loading. Why is it necesary for my poor laptop to run at 100% CPU usage for a long time, just because I want to view a website? When gmail just came out it worked perfectly fast on my computer, but more and more javascript have been stuffed into it, so now it's almost useless for me. The tendency is same for many, many websites.

Re:Disturbing...

[vG_NeSS_Vg]

Just use the "Plain HTML" feature of Gmail. I can't provide simple one line answers for all the other websites you visit, and I'm pretty sure I don't want to see your viewing habits. I'm no NSA.

Re:Disturbing...

[Wierdy1024]

Also, remember just by changing browsers and disabling browser add-ons things can speed up a lot. When I was away (without my laptop), the only pc I could find was a windows 95 16MB ram pc, and after loading up firefox (took about 2 mins), the actual web browsing wasn't too bad.

Re:Disturbing...

[Bob of Dole]

Seconded on the plain HTML version. I'm on a 2GHZ box with plenty of RAM, and I still use the plain HTML version cause while it make require more network traffic, it doesn't freeze up my browser for several seconds whenever I load it.

Re:Disturbing...

[crabpeople]

Yes your absolutely right. The internet should be fully compatable with 1995 technology for the rest of time. The issue isnt that you DESPERATELY need an upgrade, it's that the web is broken and needs to be redesigned.

Re:Disturbing...

[Anonymous Coward]

Considering that everything down to smartphones and the Nintendo DS now has a web browser - yes, pages should either work directly with low-end computers or offer alternatives that do.

Poor Layout

[COMON$]

I think poor layout is more of the problem than loading times. In the late 90s and up to 2003 or so, websites were aweful from an artistic standpoint, but many sites have gotten better.

But for me the ability to sort through goods is the #1 priority. Yes I like to have a pretty site to look at but if I cannot find what I am looking for with a few simple queries then I am gone. Newegg is a fine example of a site where I can find what I want quickly. Tigerdirect is getting better. Dell is the worst. I have a theory that Dell is like many supermarkets, they rearrange stuff and make searching difficult so you see the maximum number of items before finding what you are actually looking for.

Web designers, if you want business, make it easy. I dont really think most people go to sites just to browse. Most of the time we are there with a purpose and as an ADD generation we want it quickly or we will move on.

Re:Poor Layout

[trianglman]

This was the main finding of the survey actually. Because it was produced by Akamai however, they ignored the top 7 reasons customers leave a site or won't come back, to focus on #8 - site load times. Navigation, security, and easy checkout/registration,etc. are the top three, only 40% of responders said page load times were a factor.

Re:Poor Layout

[jsebrech]

But for me the ability to sort through goods is the #1 priority. Yes I like to have a pretty site to look at but if I cannot find what I am looking for with a few simple queries then I am gone.

Same here. I get massively irritated by the dumb designs of web stores like amazon and itunes when it comes to searching through music. For example, most music web stores let you search by genre, but only have the most generic genres, and won't let you combine them. If I happen to like instrumental electronic post-modern rock music, then I should be able to search for just that, and not have to trawl through the thousands of products in the rock genre. They've generally tried to offset this lack of categorization and metadata by using recommendation systems based on past purchases, but these rarely recommend something I might actually want, filling the web store with even more useless litter that distracts me from getting to the products I might want to buy.

In short: web stores suck pretty badly, and they're not improving. But I doubt the page load time has much to do with it. Pageloads are only an issue because it takes you dozens of pageloads to get the product you want (unless you know it by name), instead of the 3 or 4 that it would take in a well-designed system.

Cdbaby.com is an example of a web store done well. The pages are simple, and the basic metadata and search system is powerful and detailed enough that I can get to something I like in half a dozen clicks or less. But there aren't many web stores like that.

Re:Poor Layout

[COMON$]

That is why I like newegg, they have that handy subcategory item. If I search for Video card for instance. The query comes up with all items that are tagged or contain video card in them. There is a simple subcategory view that comes up where I can select the details such as number of pipes, RAM, manufacturer, or chipset.

You would think online music stores would do the same thing. Select a major category like alternative, and beneath that there would be all the subcategories.

Maybe the online market is just so new that the developers dont know what works well so they just come up with their own system. Perhaps this is a new class that needs to be added to BS programs, just study it alongside queue theory.

Typical management issue

[error406]

In my experience developers are not the problem. The initial site is fast and responsive. The traffic grows and stuff keeps getting added until the point where the site needs either an upgrade in infrastructure, an interface redesign or at least a partial rewrite of the software. The management-level decision to authorize such changes has for as long as i've been developing website always taken way too long.

It can't be that simple.

[artifex2004]

I just can't imagine that someone who is used to using, say, Amazon.com, is going to blink, much less suddenly switch to another vendor, if they have a 15 or 20 second page load every once in a while. Now, sure, if they did it all the time, I'd start to wonder. But since a site like Amazon trades on the fact that it's a central clearing point for a vast selection of inventory, there's a built-in barrier to trying someone else based on the assumption that they are less likely to have it. There may also be barriers to switching based on unfamiliarity with alternate vendors, etc., but previous experience, if not outright customer loyalty and perception of being able to deliver the goods, really drive retention a lot more than how fast you can always get that page up.

Now, whether Akamai is being disingenuous or something else... I really couldn't imagine :)

What about performance of slashdot?

[jrest]

I just did a completely non-scientific experiment. I pressed reload three times on the main page of slashdot. It took 9, 9 and 10 seconds to get the complete refresh.
For good measure, I also tested with IE6. I got 12, 12 and 11 seconds.

What is the experience of other users?
My specs:

• office LAN connected to SURFNET in the Netherlands.
• FF2.0
• WinXP SP2

Performance on other websites:

• http://news.bbc.co.uk/ 2 secs
• cnn.com 2 secs

Regards

Re:What about performance of slashdot?

[twodot72]

I got 10, 11, 11 seconds.

(From an 8 Mbit ADSL connection in Sweden)

Re:What about performance of slashdot?

[SlashSquatch]

So you still bought the product. The product being hot air, the currency is your time.

Crappy summary of a virtually worthless survey

[Psykosys]

"before potentially abandoning" is the actual phrase used in their press release. As in, people think they would think about abandoning a site after this much time, and said so when taking a survey (but were not exposed to this load time to see what their actual reaction would be). So this is not nearly as clear or useful a metric as the summary would have it appear.

Flash? No thanks.

[Channard]

I'm on dial-up so I expect some long load times. However, if I get a flash webpage, I close it - I'm not messing about waiting an age for it to load. I'm not looking for some crappy 'multimedia experience' - I just want information, to buy something or whatever. I'm getting Broadband soon and I'll still be closing anything flash sites, no matter how fast they load. It's the web equivalent of powerpoint poisoning, and the worst thing is virtually every flash page I've seen hasn't been skippable.

Re:Flash? No thanks.

[Benson Arizona]

I agree and I have no bandwidth issues. Flash is no more acceptable than RealPlayer or any other "multimedia" infestation on my computer. When I see "flash plugin required" I leave the web-site, never to return.

Why I leave..

[nolife]

Wow, a web caching company has determined that customers abandon a site that is slow.

That is not an issue for me. My highest chance of leaving is determined by when in the buying process, the site provides total price including all shipping, handling, taxes, and acceptance of coupons codes. If they need my name and address I may leave depending on if they have a shipping link or general shipping info somewhere on the site that I can reference first. I will ALWAYS leave if they require CC or payment information before providing the total price or even a hint of shipping costs.
I guess they need my address prior to calculating shipping and handling charges if they do not have flat rates but a place to enter my just my zip code should be enough IMHO.

For a good example of providing a good experience is NewEgg. They includes the shipping costs right next to the product descriptions without even having to go to a cart first.

I view the delay or confusion of shipping and handling charges to be an attempt to hide a total cost or get you to get so far that they figure you will not back out. I will back out and take my business elesewhere.
Almost like the the Ebay sellers that charge $20 to ship a motherboard (at least they are up front about it though).

Re:Why I leave..

[clickety6]

I agree - and as a person who often buys from US sites to be shipped to Europe, it's even more annoying when you have to go through the whole process of giving names, CC details, etc. only to be told right at the end that they don't ship outside US... (or if they do, but the postage costs will be 5 x the cost of the item you're ordering (and way above what the US Postal service actually charges. Look guys, if you don't want to ship to Europe, just tell me up front! Don't wait until the last minute or charge me a ridiculous amount to discourage me!)

Re:Why I leave..

[vigilology]

Agreed, and also...

I hate sites that require you having to create an account with them, and having to go through the whole give email address, receive email, go to special URL or enter / create password THEN you can buy stuff rigmerole.

This is what I want, this is my address, here are my card details; now send it!

Ignores Parallel Processing

[Hillgiant]

Tabbed browsing largely negates this. It takes me more than 4 seconds to digest any given page, so even if I am looking at only one slow site, I can still flip back and forth between two tabs, reading the one while the other loads.

off-topic on "akamai"

[dreamyu]

This is totally off-topic here, but it's a small little detail anyway: Did you know that "akamai" is the hawaii. word for intelligent, clever?

Any web developers who didn't already know this

[Tetsugaku-San]

Don't deserve the title. I've been designing and developing websites since 1997 and I have ALWAYS worked to a maximum of 10 seconds for a page to download - even back in the day when everyone was on modems . People come to your site for a purpose, all the flashy crap designers love to put in is just a barrier between the user and that purpose.

How long to find something

[Programmer_Errant]

That's probably at least as important. I've visited websites that don't have efficient ways of finding what you're looking for. Either you have to drill down through too many submenus or everything is lumped together and you have to tediously browse through about 50 pages of selections. Provantage is a good example of not being very browsing friendly.

I must be part of that statistic

[Centurix]

I'm here waiting for the thread to load up, click reply and decide to wri.. oooh, shiny thing!

4 second load time

[weteko]

I thought the load time only was important when it came to porn. That is; the profit of the site is related to the amount of time the guy browsing it can hold his load and click all the random CLICK ME-adds.

And really. 4 seconds is a long time! Damn these people must have a good sex life.

Gmail??

[bogaboga]

In my situation, a 5MB/s internet link and a Windows 2K machine with the latest patches and Firefox 2.0, the Gmail interface takes more than 4 seconds. I am not impressed with this. But when I boot into Ubuntu, the same Gmail takes less time than on windows, which is good enough.

Speaking of which....

[Toreo asesino]

Stats for http://slashdot.org/ -

Request Count: 78
Bytes Sent: 50.871
Bytes Received: 436.121

RESPONSE CODES
--------------
HTTP/200: 78

RESPONSE BYTES (by Content-Type)
--------------
          application/x-javascript: 288.162
application/x-shockwave-flash: 22.517
                                        text/html: 17.348
                                        image/png: 11.410
                                          ~headers: 21.942
                                          text/css: 37.599
                            text/javascript: 9.026
                                        image/gif: 28.117

That certainly takes longer than 4 seconds.

Re:Speaking of which....

[aelfwyne]

Dunno, slashdot on my system loads in about 2 seconds. Maybe your connection is just slow?

Re:Speaking of which....

[Toreo asesino]

Nearly half a meg in 2 seconds? That'll be the wonders of caching, my friend.

Re:Speaking of which....

[ivan256]

Admittedly, I use Flashblock, but 400k in 2 seconds is no big deal these days. With a $29/month internet connection I can easily do that in 5 tabs at once. Most big name sites can easily push out 200k/second, and 15Mbit internet connections are cheaper than AOL these days.

Re:Speaking of which....

[N7DR]

What tool did you use to get that info? I'm serious... That summary looks really nice and I'd like to use it on some sites.

Re:Speaking of which....

[N7DR]

Never mind; it must be Fiddler, from www.fiddlertool.com.

FlashBlock

[Midnight Thunder]

This is one tool I really find useful for Firefox:

https://addons.mozilla.org/firefox/433/

If you can't stand flash, then its for you.

Re:FlashBlock

[Zantetsuken]

I'm not sure if you understood, but the gp meant sites that use the full 800x600 page width for flash content with no real html alternative. Whether he uses FlashBlock or not, he wouldnt have a very good experience on such sites. I love being able to use Flash, but for accessability to low-speed dialup, 64bit linux (I still havent heard of Flash player working in that), and screen-readers (for the blind, etc) you would have to practically redo the site in standard html and maybe even use javascript and possibly some php for the non-flash users. At that point, you're just as well off making just the html version...

Re:FlashBlock

[Midnight Thunder]

At that point, you're just as well off making just the html version...

That's an acceptable solution ;)

Re:FlashBlock

[jsebrech]

I love being able to use Flash, but for accessability to low-speed dialup, 64bit linux (I still havent heard of Flash player working in that), and screen-readers (for the blind, etc) you would have to practically redo the site in standard html

Not true. Flash 6 and newer can integrate with screen readers or be made self-voicing (in which case a screen reader is unnecessary), and it is generally quite possible to build accessibility into flash applications. Also, in my experience a web app redesigned from HTML to flash decreases in bandwidth needed. You can get a very complicated web app with thousands or even tens of thousands of lines of code in a flash movie that's less than 100 KB because of how efficient the format is compared to HTML and javascript (bytecode-compiled).

The basic problem with flash bloat is that most flash designers are just that, designers. They're not developers, so they don't really know the tool they're using or how to use it effectively. An example of efficient flash built by developers is google finance's charting. It loads fast, it is powerful and it isn't intrusive.

Re:FlashBlock

[Midnight Thunder]

An example of efficient flash built by developers is google finance's charting. It loads fast, it is powerful and it isn't intrusive.

I'll agree there, though this tends to be the exception rather than the norm. Google also provides a non-flash UI for those who don't have or want flash. I think in general Google does well on the accessibility front.

Speaking of AJAXian load times of favorite sites

[bberens]

Am I the only one who gets absolutely infuriated when it the g-mail on my google home page lags a second before coming up?

One thing I see a lot of developers do which really kills me is to actually load initial content with AJAX. This is the reason the Google home page is slow. Apparently other developers disagree with me, but I've always generated the initial load data server side on the original request and then used AJAX for updates only. AJAX shouldn't be generating your entire page layout from a call in the body onLoad.

Thanks,
bb

What the fuck are you talking about?

[Anonymous Coward]

What in the name of glorious fuck are you talking about? Microsoft was the original developer of XMLHttpRequest! IE has supported it longer than any other browser, include those put out by the Mozilla Project.

From the Wikipedia page about XMLHttpRequest:

The XMLHttpRequest concept was originally developed by Microsoft as part of Outlook Web Access 2000. The Microsoft implementation is called XMLHTTP and, as an ActiveX object, it differs from the published standard in a few small ways. It has been available since Internet Explorer 5.0 and is accessible via JScript, VBScript and other scripting languages supported by IE browsers.

The Mozilla project incorporated the first compatible native implementation of XMLHttpRequest in Mozilla 1.0 in 2002. This implementation was later followed by Apple since Safari 1.2, Konqueror, Opera Software since Opera 8.0 and iCab since 3.0b352.

I don't know what the fuck you're talking about, and I don't think you do, either. Of course Internet Explorer supports AJAX. If it didn't, serices like GMail wouldn't have even been created in the first place. You do realize that the vast majority of the people in the world use IE, right? You won't see a technology become as widespread as AJAX has become if Internet Explorer doesn't support it.

Re:What the fuck are you talking about?

[brunascle]

i stand corrected. IE does support AJAX

it looks like i was running into problems because, as i said, there is no XmlHttpRequest in JScript. that's why, if you look at the source code for Sarissa, the class XMLHttpRequest is defined inside an if block that checks if the browser is running IE.

in addition, feel free to test this code yourself, as i just did:

<script type="text/javascript">

function DoStuff()
{
try
{
alert(((XMLHttpRequest) ? "yes" : "no"));
}
catch (e)
{
alert("not even close.");
}
}

</script>

Re:What the fuck are you talking about?

[GoatMonkey2112]

That is true, but there is an easy workaround to get the exact same functionality from IE.
new ActiveXObject("Microsoft.XMLHTTP");

Re:What the fuck are you talking about?

[absoluteflatness]

So wait, what is your actual problem with the GP? It said that Microsoft created XMLHttpRequest first, with Mozilla making a compatible implementation later in 2002. Those are both true. Nothing in your post actually contradicts the post above. I assume you are trying to say that the Wikipedia article is wrong, and the W3C made the spec first, but that isn't true. The W3C's working draft wasn't published until this year.

Re:AJAX completely lacks performance.

[Bogtha]

AJAX suffers from some severe performance problems.

This is a nonsensical thing to say. It all depends on what you are doing. Ajax can significantly increase performance too. Remember when GMail was first launched? The #1 thing everybody said was that it was fast. Why? Because it used Ajax.

We looked into using several AJAX-based Web forum systems, from several different vendors. After trying them all, those of us who were working on the project were quite disappointed. The performance of the AJAX-based products was absolutely terrible.

Without mentioning what those systems were using Ajax for, there is zero useful information there. It's certainly possible that Ajax decreased performance in these cases, there's plenty of people throwing Ajax at things where it just isn't useful just because it's the buzzword du jour. On the other hand, there's also plenty of people using it as just another tool, and getting decent performance and usability improvements out of it.

In short: "Ajax completely lacks performance" == stupid. "Ajax harms performance when used to do [x], [y] or [z]" == useful information.

Re:AJAX completely lacks performance.

[AKAImBatman]

So, did it take 20 minutes to copy the file through that AJAX interface? (rolls eyes)

Friends, I would like you to meet the newest Slashdot troll. The "AJAX performance is terrible!" troll.

Unless, of course, you'd like to actually provide a few examples of these "AJAX-based Webforums" that suck so much?

Re: Amazon.Com clearned this along time ago.

[mac.man25]

Amazon still codes their pages so they come up "fast" on a 28.8 modem. Ebay is the same.

Right. So the reason that ebay comes up in 3-6 seconds on my 1.5Mb DSL connection is why? I'm sorry but I HATED ebay when I had dialup (less then 7 months ago) because it would take 20 seconds or more before i could view ANYTHING, not just all the pictures and stuff.

Ebay is by far the WORST site for load times of a "major" site on dialup.

Nielsen said it a long time ago

[rw712]

The (somewhat controversial) web-usability expert, Jakob Nielsen, has had much to say about response times. From his 1994 book, Usability Engineering, he states, "10 seconds is about the limit for keeping the user's attention focused on the dialogue." (reference: http://www.useit.com/papers/responsetime.html). You may have heard of him through his 2000 book, Designing Web Usability: The Practice of Simplicity. There has been a lot he has written that, in light of new methodologies, still makes good sense/practice.

No way

[lilfields]

Myspace has proven this theory, or study rather, wrong entirely. I've sat for minutes (at least it felt like it) waiting on a myspace page to load on a high speed connection...all because someone told me to look at this or that on the god foresaken website. Though I saw on CNBC where Myspace traffic was falling at a near rapid pace...and all I can say is...about time. I could rant and rant about Myspace, but it would only be treading water, and completely irrelavant to the article.

Mod Parent Up!

[TheNinjaroach]

Why he is sitting at score:0 I don't know. He's right, not friendly, but right. Wish I had a few to hand out today, because +1 informative seems to hit the spot on parent post.

Subject dependent surely?

[scotbot]

I bet the some people are happy to wait longer than 4 seconds for his favourite pr0n to download. This is slashdot after all.

No. It's registration

[gelfling]

Registration causes me to abandon online store sites. I have a credit card. We've been working out the bugs for 50 years. Credit card sales work pretty well already. I am not going to waste a lot of time registering for your damn website in order to qualify to spend my money with you. Sorry. If you really really really really really really want me to do your market research for you I demand a discount - maybe free shipping, which is my other pet peeve. It really doesn't cost $15 bucks to pick, pull and pack that widget. I'll trade a few seconds of load up time for that. Now get to work assholes. I'm the customer, not you.

Re:No. It's registration

[crabpeople]

"It really doesn't cost $15 bucks to pick, pull and pack that widget."

No it costs 15 bucks because of the many things that can happen to a package. Free returns, lost shipments, reships (say you forget to pick it up at the post office.. OOPS), large orders, courier damage - the list goes on. If a company charges everyone flat rate shipping, then you have to factor in and average out many many other costs than just the stamp, box and peanuts. You are basically paying for all the bad customers and fuckups of the entire postal system.

Re:I for one...

[maxwell demon]

Well, the difference is between doing something and just waiting. There are few people who would care if they have to drive 10 minutes longer to get something. OTOH it the extra time consists not in driving, but in waiting at traffic lights, they probably will object. Indeed I can imagine someone to drive 10 minutes longer in order to avoid 5 minutes waiting.

One of the very early improvements in browsers were that the page was displayed not only after being fully loaded, but the parts already loaded were displayed immediatly. I think good web design should use that by having the interesing content load first, and all the secondary stuff (navigation, etc.) later. Unfortunately many web sites seem to do quite the opposite.

Shows how times have changes.

[xutopia]

"Currently, the minimum goal for response times should therefore be to get pages to users in no more than ten seconds, since that's the limit of people's ability to keep their attention focused while waiting." - Jakob Nielsen March 1st 1007 - Source: http://www.useit.com/alertbox/9703a.html

Loading times have always been a factor. Instant is usually better imo :-D

Re:Shows how times have changes.

[trongey]

"Currently, the minimum goal for response times should therefore be to get pages to users in no more than ten seconds, since that's the limit of people's ability to keep their attention focused while waiting." - Jakob Nielsen March 1st 1007...

Dang, he figured that out nearly a thousand years ago, and nobody is paying attention yet. And they were still hand writing every page back then.

Re:Shows how times have changes.

[xutopia]

rofl... damn keobarid.

Load times?

[MeanderingMind]

Personally I don't care about load times unless they exceed more than 20 seconds. It's the time it takes me to actually understand/find anything that bothers me.

My college's student portal was like that. Sure it would load quickly, but it was a pain in the rear to navigate. The splash page was plastered with options, tickers, sidebars and crap that you could care less about. The categories to navigate through were even worse, and completely unintuitive.

That's why sites such as Google or Thottbot appeal to me. You load them and there aren't any bells and whistles flying everywhere distracting you from what you're actually trying to do. I'd easily take a page that loads in 20 seconds over one that loads in 4 if it simply cut the crap.

Thank God they've researched on it

[sarathmenon]

I've kind of realized this since I've first used hotmail back in '96. Luckily, someone finally has research to back me up. Now, why is this headlines on /.?

Research?

[kbox]

I'm guessing the Dept.of pulling random numbers out of your arse were responsible for the research.

"so how long do you think people wait for a page to load?"
"iuno... 4 seconds?"
"sounds about right..."

Re: Amazon.Com clearned this along time ago.

[blackjackshellac]

Not ebay.ca, it's like watching paint dry waiting for those bloody pages to appear. Not that I mind though, we hates ebay.

Re: Amazon.Com clearned this along time ago.

[AnotherDaveB]

Amazon still codes their pages so they come up "fast" on a 28.8 modem. Ebay is the same.

Doesn't sound like the Amazon and Ebay sites I visit on my 56K modem.

I ran a little test using Safari's show page test load window option from the debug menu, results below.

• Amazon UK:
  • Total (load) time: 79.275 seconds
• Ebay UK:
  • Total (load) time: 17.54 seconds

Good, I am so happy that someone spoke up!

[Hatchback Mustang]

All of these new web 2.0 sites are slowing everything down, they often look cluttered, put more stress on the browser, and overall make a crappy expirence. I am happy with 10-20% images and 80-90% text.

This is the best reason to get AJAX

[DaitanGio]

This is the best argument for getting old web application and ajax-ize them, reducing load time and getting a better user experience.

Sponsored by *Akamai*

[Colin Smith]

THE WEB CACHE COMPANY

Nuff said. Tagged as FUD. Jupiter Research go down as spin for hire.

Re: Amazon.Com clearned this along time ago.

[neowolf]

Hmm... 10 seconds load time for Amazon.com on a 1.5Mbit DSL line. That's just for the home page. Trying to do a search or actually buy something has been nothing short of agonizing ever since they added "A9". I'd hate to think of what it is like on a modem.

Re: Amazon.Com clearned this along time ago.

[localman]

Yet the only company that actually kept it below 4 seconds was Zappos.com. Yeah, I admit I'm patting my team on the back :)

Cheers.

Not developers

[Angst Badger]

Will this study finally show developers of shopping websites the importance of the performance of their websites?

Developers already know this. But at the end of the day, we're paid to implement the ill-considered plans of marketers and designers.

not so bad

[ThePepe]

I remember taking a web design class years ago at Purdue that called for a 3 second load time as people clicked away after that. Whats funny is that modems were still in heavy use at the time so you could pretty much only load some text and a small pic and there goes your 3 seconds.

Poor layout? What pisses me off

[Colin Smith]

Are web sites (and the corresponding fuckwit designers) who assume that a browser window is exactly 1024 pixels wide... I happen to run my process bar down the right hand side rather than at the bottom of the screen, this takes away perhaps 100 pixels and the number of web sites this fucks up is absolutely staggering. btw having the process bar at the side allows you to read the name of each window no matter how many of them there are, it makes far more sense than having it at the bottom.

It's not that simple - Read the report !

[SMSailor]

1. It doesn't say that 4 seconds is "four seconds loading time is the maximum threshold for websurfers." In fact it says that 32% of people (on broadband) are willing to wait up to 4 seconds for a site to load. 68% are willing to wait longer.

2. They gathered this data by asking people "Typically, how long are you willing to wait for a single Web page to load before leaving the Web site? This is a very inaccurate measure - most people have no idea how long 4 seconds is when answering a question like this (obviously they would if they stopped and counted it out). They just pick something that sounds reasonable. A true study would have instrumented some web sites and gathered abandonment statistics for various page times.

3. The "overall performance" that caused people to say they would abandon a site included "poor error messages" and "site crashes."

I absolutely agree that sites work well and be fast. However posts like the one above will result in the "4 second rule" being carved into stone. It will get repeated around the internet and be built into business plans. All of a sudden, 4.1 seconds will become unacceptable. It's not that simple. Also, please note that this study was commissioned by Akamai, whose business is website acceleration. They have a vested interested in convincing you that you need to speed up your web site.

Slashdot loses!

[HTMLSpinnr]

I just refreshed Slashdot's homepage and FasterFox (http://fasterfox.mozdev.org/) reported a load time of 9.545 seconds. Good thing my ADD is only mild or I'd... oooh, pretty flashing banner ads...

Not just shopping sites

[Illusion]

While Akamai might have a vested interest in the results from this study, its outcome is similar to non-public data I've seen elsewhere. At least for the general public on a large site, making the site faster directly results in users staying longer and returning more frequently. If they make money from their users in some way, either by selling them something or showing them ads, improving their page load times usually directly results in more money.

Run a squid proxy and cacheing nameserver

[Colin Smith]

Somewhere on your home LAN. Your web page latency will reduce substantially.

Re: Amazon.Com clearned this along time ago.

[localman]

Actually, Office Depot pulled it off too, but they were in a different category.

Cheers.

I guess I have too long an attention span?

[Teilo]

Is that before or after the "Looking up slashdot.org" message disappears? Seriously, for a site you've never visited, it can easily longer (much longer) than 4 seconds, just to look up the DNS record.

Adobe.com is the slowest of them all

[Fotmasta]

That site hasn't loaded in anything close to 4 seconds in the last eight years. I curse them to hell everytime I have to go there.

Bull is not far from the truth

[trianglman]

I downloaded and read the whitepaper that is linked from that article. The summary is an obvious skewing of the actual findings to focus on Akamai's business. A couple key examples are: "Which of the following factors are most influential in your decision to continue shopping with an online store where you have shopped in the past? (Select all that apply.)" - 65% said good navigation, page load time was 8th of 13 with only 40%. "Typically, how long are you willing to wait for a single Web page to load before leaving the Web site? (Select one.)" - 21% said 3 - 4 seconds, however 30% were willing to give the page 5 - 6 seconds (broadband numbers) and another 38% were willing to wait more than 6 seconds.

Re:AJAX completely lacks performance.

[inKubus]

It's pretty easy:

If you're application uses a lot of repeated screens and is really only a data-view and entry application, you should go AJAX. Because the slightly longer initial load time (to load ALL the interfacey stuff) is better than having to rerender the interface over HTML every time you change views.

If it's a step by step wizard type thing, or informational (think wikipedia) just get on with it using syncronous web pages.

Where AJAX fails is in the hands of inexperienced developers, where they won't allow the app to load almost everything before running. This is not always possible--something like google maps is a good example of this. You are going to have to load the maps as you go because there's too much data. However, google maps really relies on Images as data which is not the most efficient. They need to expand their client to render the maps itself from GIS info (obviously the satellite overlay will need to come from images).

Also, it fails when there is a high latency connection. However, a lot can be learned from past interfaces: feedback! Flash a div on the screen letting the user know it's loading, apply visibility:hidden when it's done. As long as the user knows that it's actually DOING SOMETHING and not just sitting there, they will give it the benefit of the doubt and wait. Test the connection latency at startup and then let the user know what you know. If you tell them in advance that they might experience poor performance because of their current connection, they are more likely to tolerate it.

Good interface design is a lot more than having it be fast. You have to keep the user informed of the current situation. It's not slowness that annoys people so much as not knowing what's going on. Early X windows had that problem for me also. Whereas in Windows when you click something the window immediately is created by GDI while the actual application loads, in X the appliation is started somewhere and then IT creates it's own window. So when you click on an Icon, it takes a few seconds of nothing (it seems) while we wait for the kernel to find on FS, allocate memory for and run the executable which does it's own init and then FINALLY pops up it's window. If you're running over a network connection, there's no disk noise to let you know anything is happening, so you are basically just sitting there wondering if you should click again. I don't know if it's still like this.

Anyway, my point is that there are a lot of tricks you can use to prevent user annoyance because it goes a lot further than some arbitrary time length.

Moo

[Chacham]

Wow, this is pretty neat.

For telnet a *much* higher level is used, at less than a second, or people get annoyed at the slowness.

Good luck buying a car

[bill_mcgonigle]

Good luck doing research for a car without Flash. It's abysmal.

Pssst...

[werewolf1031]

Gonna let ya in on a little secret. They got these things, see, they're called "servers". Websites live on these things. Sometimes, lots of websites can live on one server. Then, they got all these "users" trying to look at the websites on the "server", and sometimes things get messy, or at least a little congested. (See "Tubes" for more advanced technical info.)

Ok, all snarkiness aside, what it boils down to is simply this. It does not matter one damn bit *how* many zillion terraquads-per-nanosecond YOUR connection can handle, the *server* can only dish out so much per second to all the folks making page requests to sites on that server. Period. End of story. You may have an Internet connection so goddamned fast the packets travel back in time, it won't make any difference. Which reminds me, <looks around again> there's also this thing called "point of diminishing returns", and it's got tons of uses in all kindsa fields, not just Teh IntarTubes... but if ya wanna know more, you'll hafta buy a watch.

Re:Pssst...

[TriezGamer]

You may have an Internet connection so goddamned fast the packets travel back in time, it won't make any difference. But it certainly brings some strange news. I got an email from my boss next week telling me I'm fired for something I haven't done yet.

Not specific study

[zerosix]

I don't think this study is really useful at all and let me explain why. First of all there isn't any definition of what "Web page to load" really means. It could be they are talking about simply starting to get data, to completely load the website, or have it load enough to be useful. Also, I don't think these people taking this survey have a good concept of time. It's amazing how you can tell someone that something takes a certain amount of time and they will completely disagree with you and when they see it actually timed they can't believe it. I find it completely BS that people say they would never visit a site again if it took longer then a set time to load. Are these people keeping track of this information writing down/timing every site they visit? I think not.

For this study to be actually useful they would have needed to have individuals visit various sites and state whether or not the loading time was acceptable. The load times for the pages would have to be cataloged within the browser without the "consumer" able to know how long it really took. Only then might we actually get some useful information.

Re:AJAX completely lacks performance.

[tehcyder]

The "AJAX performance is terrible!" troll.

It's still quicker and better than Java though.

*ducks for cover*

Not for non US residents

[chord.wav]

For non uS customers, the most bummer factor is that almost every online shop accepts only US-issued CCs, no matter wether they are Visa, MC or AMEX. For Buddha's sake! You can't even get an iTunes account cause they accept only US cards!! And we are talking about Apple, not some little store.

Depends entirely on what...

[Kjella]

If I go to the front page of an online newspaper or similar site, I expect it to take long to load. If I click an article link, my patience is very thin because my target is simple article text, possibly with an illustration. Same with an online store. Also, there's a few tricks that lets the browser render it before it's loaded all the items - for example setting an image's height and width attributes. Not everyone has learned that yet. Also it depends on how much meaningful content there's on a page. If I have to visit a [break] new [break] page [break] for [break] every [break] sentence, I'm a lot less patient than if you just load it in one big honking page that I can scroll.

In short, measuring cost (time) without measuring benefit (content) is meaningless. If google's search page took four seconds to load, they'd be a dead duck. Other pages couldn't be rendered in four seconds with a Core 2 Quad and GigE, but are still highly successful. The pages you want to check is where the user asked you for something specific, in which case you'd better deliver ASAP without crapping up the page with everything he didn't ask for. Pages that are slow, I can live with. Pages that are slow, deliver little and waste time on meaningless stuff I don't.

Re:Depends entirely on what...

[TaoPhoenix]

Slashdot has linked to its fair share of stories in which two manuscript pages of text are spread over seven flash-ad filled pages. The Ad time on TV is about lost time through viewing ads; the lost time on ad-jammed news pages is equally lost waiting for page loads (I used to have a low end Satellite, and it took thirty ( ! ) seconds... and when it rained I WAS back to dialup. )

We all laughed because "those 300 baud modems back in the day could only handle small text files" - and now we spend our time looking for "information" ... which is small text files (and MAYBE a picture.)

Attention span

[javilon]

The reason for this is that Surfers attention span is also about four seconds, so if the page takes longer to load, they forget what they where doing...

Notice who sponsors the study

[melted]

Not to argue that nothing on the web should take more than 4 seconds to load (on _second_ load), but notice who sponsored the study - Akamai. It's like if Microsoft sponsored a study "proving" that Linux sucks.

Caching is your friend. If you cache, don't forget to version your stuff as well:

<script src="foo.js?d=md5sum-of-the-script"></script>

And do this with everything you cache - css, xml, xsl, whatever.

Why use dial-up?

[mks113]

As for all the people saying they still use dialup, why?

Because I'm in the middle of Africa?

I share a 750mb/s sattelite downlink with about 1000 people. It costs around $3500 a month. It wasn't much more expensive than the 256kb/s ISDN line that we used to use. (our uplink is still ISDN)

That covers the school, but us staff can get dial-up for about US$20/month which is cheap for this country.

Or I can use my GPRS phone at 20/- per megabyte (That's about US 28c/mb).

One thing I do look forward to about moving back to Canada is my DSL internet.

four second psychology

[epine]

This all begs the question: what is the psychology behind the three or four second user patience threshhold? My vote is that after half a dozen four second waits, the distract-me-from-life-by-buying-more-crap circuit begins to lose its grip, and a wee murmur of "what the heck am I doing wasting my life on this garbage" penentrates the dim folds of consciousness. This is not a case of humans engaged in rational activity. These are just shopaholics regulating their emotional state the same way most true geeks regulate their mental electrolytes through beverage consumption. You could do the same study on the coffee maker at work: fresh brew delay vs oh-screw-it-Starbucks-needs-my-money-more-than-I-d o.

Compare to elevators

[Hoi Polloi]

I read that 30 secs is approx the max time people are willing to wait for an elevator. I thought that was impatient until now. Though if you are going through dozens of pages and you can scan them be eye in a few secs then those delays become annoying fast.

Answer to a survey vs. Real world action

[The_Crowder]

The danger of conducting surveys is that the answer to a question often times varies from what a user will really do in the real world.

Take the topic of "...the New Threshold for Acceptability..." as an example, the survey asks the question: "Typically, how long are you willing to wait for a single web page to load before leaving the site?" 31% of users said that they were willing to wait less than 5 seconds for a page to load before leaving. I am willing to bet that if you were to evaluate the metrics of online retailers you would find that the number of people leaving your site after a sub 5 second page load is less than 31%.

On a similar note, this survey fails to tell us whether this is something a user is willing to put up with only once or on every page load. It is my suspicion that a one time page load of 5 - 6 seconds is not going to cause "A full one-third of online shoppers with a broadband connection..." to abandoned the page; however, I would be willing to accept that if every page took 5 seconds or more to load on a broadband connection you would see a large amount of users abandoning the site.

Finally, I found it very interesting that the survey limited the answers to 6 possibilities (sub 1s, 1-2s, 3-4s, 5-6s, 6+). In my opinion, they made a big assumption in choosing 6s as the threshold of acceptability...where did that number come from/why was it chosen as the cap? I think the survey would have been able to benefit from having a higher range, perhaps to 10s as suggested by Jakob Nielsen (http://www.useit.com/papers/responsetime.html)

Long load times...

[cr0sh]

I am going to post this here, and while it may seem off-topic, it really isn't, and I don't think it would be accepted as an "Ask Slashdot":

I work for a company developing thier website. Our website is currently based around old-school ASP, and while I would love to change this platform to something more up-to-date, it isn't going to happen any time soon (ie, I am one programmer and there are well over 100KLOC of ASP involved, not too mention tons of MSSQL-specific SQL that would need to be converted, mainly for date handling reasons). I am stuck on this platform for the time being. I did not choose this platform, that was done for me a long time ago.

We have more than enought processing power to handle our site, but we have an issue with long page load times - from the front site to deeper pages. I have isolated the problem to our heavy use of a function which essentially does a singleton query (ie, given a table, field, and record ID, it returns the data value in that field - ie; SELECT field FROM table WHERE id = ???).

This form of query/function (which is actually done in a database wrapper function - so only the function call is used everywhere - if I can speed it up, the entire site benefits) was implemented due to another issue, which, when I first got here and saw what was happenning, made me say "Wha?" and try to do it the proper way, only to get stung by the problem that this solution was to fix. What this problem was is as follows: If you try to issue a SQL select on our system to return multiple fields from a table (ie, SELECT field1, field2, field3 FROM table WHERE ...) - sometimes, the data returned could be accessed (ie, var1 = Recordset("field1"), var2 = Recordset("field2"), etc) just fine, but at other times (seemingly, maddeningly - at "random" intervals), only the first field would have data in it, and all the other fields would return empty string or null values! We don't know what causes this or why (I suspect some form of field ordering in the SELECT vs. the order of the fields in the MSSQL table - but I am not sure), but by using singleton selects, this problem is avoided (because there is *always* data in the first field), at the expense of speed. Unfortunately, at this point, even if I found out how to keep the problem from arising and could use the proper SQL queries, there are simply way too many places to change for it to be practical (it would still be a HUGE job to perform to mod that much ASP code).

From what I have googled and researched on the interweb, singleton queries using ADO recordsets should never be done - they are very slow. I have heard that using an ADO command object, either coupled to the SQL select with parameter fields, or by using a stored procedure, is a better and faster method than using a heavy recordset. However, I have not been able to get either of these methods to work properly - I either get no data back (empty strings), or errors. I am not absolutely sure if it is my stored procedure or what (although, it seems to return data OK if I run it under Query Analyzer). What is interesting, even though it fails and doesn't return data, I seem to get the same performance (timing-wise) as the regular singleton query (ie, using an ADO command object with a stored procedure seems as slow as using an ADO recordset object and a query). I am not sure if this is because of the failure of returning data, or if I did get it working, if it would run faster.

Seriously - I wish I could convert everything over to Linux, Apache, PHP (or Python), and MySQL or PostgreSQL - but that isn't happenning - we don't have manpower or resources to throw at such a project (plus, considering new development and fixes on the current codebase, it would be a constant game of catchup and patching, even if we have another developer handy to take up this issue alone).

Does anyone here have any reasonable ideas or solutions I could possibly try? Does anybody know why we experience the random data droppage when not using singleton selects? Does anybody have any ideas on how I could speed this beast up? Or, am I simply bailing water out of a sinking ship?

Comment removed

[account_deleted]

Comment removed based on user account deletion

why dialup?

[CAIMLAS]

Short answer: because the United States is a lot larger with a much lower population density in most areas than Europe, Japan, and pretty much every other Westernized country in the world. I'm not kidding when I say there are many people who live in locations where the closest grocery store is 20 miles away or more.

The UK has 93,278 sq miles of land - slightly smaller than our state of Oregon - whereas the state I live in - South Dakota - has 77,121 square miles. It is only the 17th largest state of 50. The UK's population density is 649 people/mile, whereas South Dakota has a density of 9.95 people/mile. There's a big difference there, one which the lack of wealth distribution/state telco funding can't or won't make up for - meaning the economy just can't support or rationalize high speed out to the boondocks.

By the way: the Western world is not "the rest of the world". Most of the world is still on dialup.

4 seconds on Myspace!!!

[Kuvter]

No wonder my page views have droped significantly on Myspace, maybe I should get rid of the huge background image that makes it so you can't read my text, and all those slide shows of picutes I have on there, oh and that automatically playing song, the video on there, and...

This could possibly be a good use for Apdex

[TheLoneGundam]

Apdex is an interesting way of scoring application time, which assigns response times a score from 0 to 1 depending upon whether they meet their "tolerance" goals or not. For any transaction, there's an number T which is the upper limit that users will tolerate for response time, and a number F, where F=4T, that is the number at which people consider the system to be failing. In the study mentioned it sounds like T=1(calculation of F is left as an excercise for the reader). Once you have these numbers, you collect statistics on the response in three buckets: total transactions, satisfied count where response is T and F. The Apdex number is (satisfied count plus half of the tolerating count) divided by the total responses. The advantage to this system is that it gives you a number telling you how well something is working in relation to its goals, and it works no matter what the units of time used are.

Fancy Capitalists

[Cyanara]

4 seconds? Pfft, sounds like web users need a good dose of humility, courtesy of my 2.5Kb/s dial-up shared over 3 computers! I reckon I might start a web server as well, waddya say?

It added insult to injury to find out that our neighbours were on a ADSL-enabled exchange, while we're on the outermost rim of ours. With the state's water supply between it and us, of course. So technically it was good news when it was announced that they were going to set up wireless broadband instead of small-ranged ADSL for us. It's just that time goes by, and that's all that's happened...

Sheesh

[Brandybuck]

Sheesh. It takes three of those four seconds for Earthlink nameservers to resolve your URL...

Re:ha!

[The MESMERIC]

It still may do.
Sometimes even if a page gets first loaded - if there is high complexity there: bad tables, HTML, CSS, convoluted scripts, etc .. you will still have to wait for all to render.

OK so next thing you will tell me your computer is a Cray XT3!

Re:Corrupt mayor from hell

[Z80a]

here i cant get my paws on a 4 mbps Cable connection thanks to the corrupt mayor of this city,that demands from the cable company a bribe do the cable reach here :/

Re:Corrupt mayor from hell

[DeathElk]

Man, that sucks. Surely your state has a telecommunications ombudsman or something similar? (I linked to dictionary.com 'cause I'm not sure what your local term is...) That mayor dude is holding back progress, man.