Average Web Page Approaches 1MB

MrSeb writes "According to new research from HTTP Archive, which regularly scans the internet's most popular destinations, the average size of a single web page is now 965 kilobytes, up more than 30% from last year's average of 702KB. This rapid growth is fairly normal for the internet — the average web page was 14KB in 1995, 93KB by 2003, and 300KB in 2008 — but by burrowing a little deeper into HTTP Archive's recent data, we can discern some interesting trends. Between 2010 and 2011, the average amount of Flash content downloaded stayed exactly the same — 90KB — but JavaScript experienced massive growth from 113KB to 172KB. The amount of HTML, CSS, and images on websites also showed a significant increase year over year. There is absolutely no doubt that these trends are attributable to the death throes of Flash and emergence of HTML5 and its open web cohorts." If you have a personal home page, how big is it?

Comments

can't wait to see these on my phone!

[Anonymous Coward]

It's a good thing phone carriers don't limit your data consumption....

oh wait..

Not surprised

[kannibal_klown]

With the growth of Javascript libraries like JQuery for more UI features, more images, I can see it reaching that high.

Meanwhile, web developers don't care because more and more people are getting faster and faster broadband speeds. So as long as the page-load metric works OK on their rig or perhaps what the envision most of their viewers have... they think it's all OK.

Re:Not surprised

[Fujisawa Sensei]

Apparently it is because users are still hitting their websites.

Re:Not surprised

[Anonymous Coward]

Web developers don't care because the majority of their images/css/js is cacheable by each visitor (and most people have jQuery cached from the official site and many sites link to that directly). 1MB page but it's only 45k on the next visit.

Re:Not surprised

[webnut77]

Yup. Google helps us out here. If we're using offsite resources like that, there's a fair likelihood that it's cached in the user's browser even if it's the first time they've visited the site.

And then Google also gets to look at the referrer data.

Re:Not surprised

[matmota]

First select the "classic discussion system (D1)" under "Discussions" in your options (gear icon). Then, in the settings just below the summary pick the "flat" view instead of "nested".

Ad Content

[Anonymous Coward]

And how much of it is ads?

If you have a homepage

[burning-toast]

I have a homepage, and it's only 4.92Kb. Granted it is the "It Works!" page for CentOS which has all of the other text and icons and such but who needs more than that? Do people really have personalized home pages now that Facebook came about (other than some hobbyists or professionals who run a side business)?

I wonder what the average "Facebook" homepage size is... since that is what most people will be seeing regularly.

- Toast

Compression?

[s7uar7]

If the bulk of the increase is from javascript wouldn't turning on compression on the web server solve the problem? They're text files, they compress down massively.

Re:Missing data

[Anonymous Coward]

10 years ago online video was virtually nonexistent, and where it did exist it was never larger than 320x240. Pictures were equally low resolution and page formatting was minimal. Allowing user comments was rare, and user contribution based sites like YouTube and Wikipedia were nonexistent. Oh yea, and the "blink" tag was still popular. So yes, I would say the amount of information has increased significantly.

Larger Pages

[Master Moose]

And Less Content. .

I remeber the days when a site would include an 10 paragraph article on one page - Not 10 pages with a paragraph on each.

Now I can feel smug

[rbrander]

My home page remains where it has been since 1993 at the Calgary Unix Users Group: http://www.cuug.ab.ca/branderr ...clocks in at 9.2K, plus a 15K GIF and a 9.1K JPG (if you "turn on images" in your browser - remember when it was a realistic option not to?)

I have held the line, while Viewing With Alarm (VWA) the growth of web pages for the entire 18 years since. I wrote Bob Metcalfe when he had a column at InfoWorld 15 years back, and he was Viewing With Alarm the exponential growth in Internet traffic and predicting the "collapse of the Internet" (had to eat those words - literally) because of it. My letter pointed out that his column constituted 2K of text - that was all the generated content that was bringing in the readers, (unless you count the 10K gif of Bob Metcalfe, and I don't), and the page had an additional 100K of framing and advertising-related image GIFs. His reply was somewhat defensive.

This last year, I had occasion to travel on the Queen Mary 2, where all internet is via satellite at a minimum of 34 cents per minute with their bulk plan. How quickly I grew to resent the giant Flash blobs that would be automatically downloaded with every page of a newspaper so I wouldn't miss the animated ads for the latest in car buys. At QM2 speeds, I'd have to wait about two minutes before I even had an "X" mark to click on to dismiss the ad. I was rather quickly cured of almost any interest in the Internet content at ALL, I did my E-mail, checked the google news headlines (fewest high-byte ads), and logged off.

My point: 90% of mail is spam. So are 90% of web page bytes. We just don't call it spam. We call it "the whole outside frame around the news page that we try not to see, but keeps jumping around into our field of view".

No, it's not HTML5. It's just junk.

[Animats]

There is absolutely no doubt that these trends are attributable to the death throes of Flash and emergence of HTML5 and its open web cohorts.

No, it's not about HTML 5. A lot of it is about bloated content management systems and templates.

I was looking at a Wall Street Journal page recently, and I brought it into an HTML editor so I could eliminate all non-story content. The story required an HTML page with only 72 lines. The original page was over 4000 lines. It contained a vast amount of hidden content, including the entire registration system for buying a subscription. All that junk appears on every page.. Inline, not in an included file.

On top of that, there are content management systems which create a custom CSS page for each content page. So there's no useful caching in the browser.

Remember those people who said CSS was going to make web pages shorter? They were wrong. Look at Slashdot - bloated, slow pages that don't do much, yet consume CPU time when idle.

Re:No, it's not HTML5. It's just junk.

[Desler]

But Slashdot is now web 2.0, Ajax-enabled and social. Pretty soon it'll be "hosted on the cloud" and provide SaaS so it can win at buzzword bingo!

Re:No, it's not HTML5. It's just junk.

[rubycodez]

let's just use 1 acronym for all that, BWC. Bloated Web Crapware

Coded mine in notepad

[Anonymous Coward]

CSS is for prima donnas and Flash is for artistes. PHP is for chatterboxes and Perl is for psychics. Javascript is for the clinically insane, and Ruby is for hipsters. Drupal is for geeks and Ajax is for nerds.

I'll stick to plain-jane HTML, thanks.

slashdot = overweight

[Anonymous Coward]

Ironically posted on a website that is itself a bloated pig.

Video sucks.

[PeanutButterBreath]

10 years ago online video was virtually nonexistent, and where it did exist it was never larger than 320x240.

And now it is ubiquitous, HD and largely devoted to pointless things that would be skimmed over and disregarded in a fraction of the load time if left to text and still images.

Re:Video sucks.

[Tumbleweed]

TubeCat disagrees with everything about what you just said.

Re:Automatic shutdown at 95C

[Anonymous Coward]

In 5 more years there will be another layer of abstraction and 5GHz CPU's will be dragged to a crawl by a Super Mario Brothers clone.

... and it also sucks.

[warrax_666]

From a usability perspective:

• Blue background. Why? Are you trying to accomplish some artistic purpose we're not privy to?
• Why are the pictures laid out vertically rather than horizontally? Why is there lots of text to the right of the second picture rather than to the right of both pictures. That means that your contact info is obscured/invisible to potential readers -- it's also out of context in that place.
• Why do your anchors span multiple sentences rather than just a few semantically relevant key words?

In short: You fail web page design, so who the fuck cares if your page is 10K?

Re:... and it also sucks.

[mickwd]

"In short: You fail web page design, so who the fuck cares if your page is 10K?"

As a normal human being possessing the ability to read, I found his site perfectly accessible, and it gave me a decent amount of information about the guy in a quick, concise manner.

If I was to be snarky here, I would say something like:

In short: You fail meaningful criticism, and who the fuck cares if his "anchors span multiple sentences rather than just a few semantically relevant key words"?

More data, and less information

[Anonymous Coward]

What's worse is that the "payload" of text is less and less interesting. Bandwidth isn't the problem. I have more than enough bandwidth for these pages. When they hit the browser, they take forever just to render. There are a handful of web sites I still use, Slashdot among them. Most new sites I just back right up. If your site does that on day 1, it's not worth the bother. I'm not buying a new machine just to look at your crap web site that's probably just a rehash of every Internet meme.

We're well into the "nobody comes here anymore it's too crowded" and/or "57 million web sites and nothing on" stage.

Re:90KB of Flash

[Anonymous Coward]

Waaaah

Sent from pretty much any Android

Re:How Big?

[Pharmboy]

Well, I don't mind bragging about mine. I was 100k, but now has swollen to 150k this year. As to *real* servers, I try to keep our ecommerce pages below 250k for gateway pages. Until this year, I tried to keep them under 150k. Up until 2008, 100k was the target. Before 2003, 50k. This is kind of light, and a few pages bust this, but very few. Before 2000, I used to spend lots of time just optimizing graphics, now I just use some common sense, PS, and very little time.

What I have found is that the total k of data isn't as important as the number of items and hosts the page calls. I find I can make my pages faster by using image maps, which make larger images size (12 images 1 image of all 12 items) but load faster because it takes less connects. There are a few tools online that can help you figure out total load times. Nowadays, load time is NOT purely a function of the size of the data. If you can cut down on the number of GETS and cross domain GETS (ie: DNS lookups) you can radically cut down load time and reliability.

Also, pages that don't need to be dynamic, shouldn't be. Our gateway (to product categories) pages are generated as we update the site, and stored static. This allows them to be cached. It sounds old fashioned, but the fact is that it greatly increases perceived latency. I am amazed at how many websites are generated via PHP and SQL on the fly, yet aren't updated more than a couple times a day or less. That is a lot of wasted CPU cycles on the server, and a lot of wasted potential for caching, both locally and down the line. And yes, it makes your website load slower, making it seem like your pages are larger than they are.