Monday, April 07, 2003
Public Sector slowdown ... on the web at least
Computer Weekly have some survey results this week that show government websites are slower than those in the private sector. Been here before of course, but the message ought to be getting through by now. We've had: low usage (nobody can find anyone that has used a .gov.uk site), poor availability (like buses, they are never around when you want them), poor design (even if you do find one and it's there, you can't figure out where anything is) and too big (unlike elsewhere, size doesn't matter). Now we have too slow. Seems like a full house to me. Except noone ever seems to say "too many". My personal belief is that the reason few people use government websites is that (a) they don't know what they are looking for, so don't know where to start and (b) if they do, by some chance, end up at one, it's often the wrong one and there's no easy way to get to the right one. If we had that right, I think the performance problem would be an issue worth cracking, but it's not right now - because ... Remember, this is taxpayer money being spent here. Every website has a cost, whether it's £1 or £1 million, having thousands of them makes for more than should be spent. As more surveys come out highlighting a weakness here or there, the combined cost of paying attention to the results and doing something about it across the entire domain soars. If the changes are made, of course. And given that this is the nth survey in the last 12 months, I doubt that they are. The present website guidelines are broken - it's not easy to get these kind of things right, it's even harder to get them right and then persuade the right people to do makeovers on their sites in record time to make them work. The over-riding guideline should be "if in doubt, don't build another website or use another domain name" Public sector people will tell you, I expect, that they spend a lot of time and energy trying to make their site work with a variety of browser versions which bloats the pages and makes them slow. That's certainly true - and it's one of the most frustrating things about developing web pages and applications. And then they want to be sure that everyone can find the service that they are after, so it's best to put as much as possible on the home page, right? That's just not going to make for accessible, fast pages. The work we're doing on ukonline right now showed us that we can get the load time down from 21+ seconds (where it is now if you try it) to perhaps 11-12 seconds through better design, image management and style sheets. Not 8 seconds, I agree (which is the traditional impatience measure, one that I imagine got forgotten about long ago when the number of browser versions in use went past two) - but checking the BBC website just now (via Site Confidence), it came out at more than 18 seconds. MSN was more than 38 seconds. Consolidate, rationalise, exemplify. - Consistency of design. If when the site is there, it looks similar to others, people will forgive slow load time, because they won't have to learn your site - Simple pages. Make it clear what's on the page and how to use it. The easier it is to find what you need, the less time you waste. - One domain name. Expect people to find what they need through search, either on your own site, google, ukonline or whatever. So don't confuse things with dozens of domain names - I expect the search engines pay no attention to what the name of the site is. Or do we need a few more surveys to tell us that it really is a supply problem, not a demand problem?
Posted by Alan at Monday, April 07, 2003