Tips for Selecting a Web Development Firm?

cyrano asks: "The organization I work with is looking for nothing less than a complete re-launch of its web site - upgrading from cobbled together static HTML and ASP pages to nothing less than a dynamic, database-driven site with a full-featured Content Management System and a secure eCommerce component. I have already collected proposals from several firms, each advocating the benefits of Java and Struts vs. ASP.NET vs. PHP...however, the technology used by each firm will only form a small part in my final decision. My true concern is ensuring that the firm I contract will be professional, cooperative, timely and will ultimately deliver their services as promised. What sort of questions should I be asking them, and what sort of warning signs should I look out for to make sure I find the perfect fit?"

Comments

prior work

[FidelCatsro]

Just a simple bit of advice , but have a close look at the company's portfolio .
There is no better gauge of skill ,a consistant quality is a fine gaurantee .
Personaly i would also do some background checks , perhaps contact your peers in the other firms for which the web deisgn/development company has worked for.
content first , style second (no mystery meat)

One warning sign...

[Singletoned]

I was in a discussion with our web developers and I said that the back end of the site needed to be easy for the administrators and moderators to use, which they disputed.

They said that they could just give everyone a day's training on how to administrate the site and then it didn't need to be easy to use.

I was so shocked I couldn't reply, and once I had gathered myself together I couldn't persuade them otherwise. The website still isn't easy to administrate.

Get references

[torninfinity]

Get references from them and actually check them out. Sure you'll get only what they want to give you, but if you ask the right questions you'll be able to sort through the fluff. The only other thing I would say is put in deliverables into your contract. Many times development has been hindered by lack of planning and set timelines.

W3C Validator and Browser compatibility

[Trevelyan]

I would check their existing work against the w3c validator
http://validator.w3.org/
and open them up in a selection of browsers.

Its is unlikely you will find any non trivial site that passes the validator, but you can see who's is worst.

You may not be that concerned about browser compatibility, but it does show how much care they take when writing their code (in case of dynamic) or html

You could go as far as putting it in their contract that their pages pass the validator.

But note that passing the validator is not absolute proof that the page is correct, just that the sgml/xml is valid.

I work for a development firm

[FictionPimp]

I work for a company that does web development (no I wont plug them here unless asked). I would say the most important thing is to look at previous sites done by them. Call some of their clients and talk to them. We give our clients a list of past clients to call.

You should also check their work in a few browsers (safari, Firefox, IE6, etc). There is no excuse for a professional website not working in all major browsers. Make sure you own the work (A lot of our clients had been screwed in the past by this).
Try to stay away from a company that pushes one language over another. We do whats best for the client, if that means php, asp, perl, java, or even .NET thats what is going to get done. The only thing we really push hard is use of CSS for all layout. It saves bandwidth and makes the site easier to modify in the future.

Finally, if you are also going to host with them, find out who owns their servers, and where they are located. Inquire about their backup system and make sure they have standards for uptime.

Some basic things

[dtfinch]

Web design - The current trend is away from flashy sites that slow or distract the user from what they're looking for and confuse search engines. Fluid layouts are good. Another trend is reducing the space between the top of the page and the content, and sometimes moving the navigation bar to the right to reduce mouse movement between the nav bar and the scroll bar and allow the content to be closer to the top left, drawing the user's eyes to the content and putting it higher up on the page for the search engines.

SEO - Search engine optimization. If they've researched it they should know the acronym. Most of the old techniques no longer work in Google. Some will get you penalties. No javascript based links that the search engines can't follow. Repetition is bad. Very similar pages are bad, often the case with product pages where a product has many similar configurations, each having its own page. Dynamic pages with long urls or a countless number of possible urls (like with session ids) are bad because search engines will only spider a small number of them to avoid getting thousands of really similar pages from one dynamic page. Aside from avoiding repetition and things that confuse the search engine, content is mostly king.

Security - They should at least know about sql injection.

Content organization - As a plus, they should be able to organize your site based on what your customers are most likely to be searching for, what you sell the most of. This decides things like front page links, what images best represent each category, and so on. It's not enough just to have an online catalog with every product having equal weight. It might help if they took microeconomics in college.

As for which language, php, java, or asp.net, it doesn't matter as much as the quality of the programmer. A lot of things seem to take much less effort in php though.

Advice from a consultant

[eddy the lip]

Context: I've been doing web development professionaly for seven years, the last four with my own small web development company. We've worked with other firms, and been called in to clean up other people's work (the latter more often than I'd care for - that kind of work is zero fun).

As many other have pointed out, language doesn't matter a whole lot. We do recommend open source platforms, for the reasons familiar to anyone that reads this site often, but the most important question about this is whether the tool will fit the job. I've told clients before that what they really want is a Microsoft solution (because it fits the requirements) and that they should really find another firm to do it (because I'd rather put a hot poker in my eye than work with ASP).

Portfolio is important, but there are a million ways of fluffing it. Maybe it was subcontracted work, maybe they happened to have a really good person working for them for a few months, and they left because the company sucked. Maybe they're a large company, and their portfolio is all A team work, but you'll be getting the B team. On the other side, we've done work that would never make our portfolio because the client insisted on a nuclear orange and blue color scheme, or 500 links on the main page.

Picking a good web development company is difficult, largely because a) most of them are truly horrible and inexperienced, and b) the important things are difficult to quantify. There's a few things that are immediate warning signs, though. These should be rampantly obvious, but this is Ask Slashdot, and I've encountered each of these from companies that a client thought looked good on paper:

• where does the design phase come in? If it's early in the process, run away. You don't know what the design requirements are until you've worked out the site architecture, at the very least. Design usually shouldn't happen until somewhere around halfway through the project. (We don't touch design until the information architecture is done, we've seen any print materials you have, and layout grids are produced)
• they tell you that as long as it works in IE, browser requirements don't really matter
• can you talk to the project lead? Many companies will have a suprememly non-technical person doing the client liason. While I think this is generally a bad idea, it's the way a lot of companies work. But if you can't talk to the person that's going to be making technical decisions, run away.
• print is not the web. There's still this strange idea floating out there that print designers are interchangable with, or even superior to, web designers. The web is a whole different medium, and requires different skill sets. If they try to sell you on their print designers, it's a bad sign. Similarly, they shouldn't be selling their web designers as print designers.
• they make any kind of guarantess about search engine optimization. SEO is this week's snake oil. There are common sense things you can do to improve your ranking, but if they're promising top ten rankings, they're blowing smoke.

Some of the things that you should look for (this list keeps growing, I had to stop early):

• do they make recommendations? There's usually some aspect of a site that can be improved by a slightly different approach than what's outlined in the RFP
• do they provide source files for creatives? You don't want to be asking them to hunt down some graphic two years from now.
• ask about process. There isn't necessarily a "right" answer to this, but they should have a design process in place, and be able to explain to you why they do things this way. They should be able to explain this in a way that's understandable to a non-technical user.
• ask about usability. Budget is going to be a factor here. It's not necessarily cost-effective to run the site through multiple usability labs, but at the very least they should know about general usability guidelines (keeping content above the fo