Speed Kills
Posted by: Carl Smith on Tuesday July 3, 2007
Last week was a crazy week for me, traveling all over the country conducting usability testing for an updated web app. Overall most thought that the new system would be better with one very serious requirement.
It had to be faster. From the beginning we have all known that speed is one of the top issues in web development. It doesn’t matter if we’re getting movie times or renewing our driver’s license. Speed is critical. Too often I think developers fail to test their systems in the environments in which they’ll be used. In the heat of the deadline and budget concerns they forget that the user won’t have their equipment on their network. As a result they never see speed as an issue.
The end result is that users complain about performance issues and developers scratch their heads not being able to replicate the problem.
I have actually heard users with no technical background try to explain problems they encounter.
“It’s the graphics they use for the buttons. Why don’t they just make them links?”
“I think there is some javascript that slows things down.”
“Maybe their servers are old?”
One user summed up the expectation for speed better than I could have. “Google searches the entire web and brings back what I want in seconds. I’m asking them to look at their own information and it takes minutes.” Minutes versus seconds may not sound like much. But you have to consider that some users will complete the transaction 60 times a day. That’s up to two hours a day lost. 10 hours a week. 40 hours a month. 480 hours a year. I don’t know what you bill hourly, but we couldn’t afford to lose that much time or money.
I know good developers want their products to work for users. It’s just that deadlines and budgets seem to get in the way. Most of us accept that there is never going to be enough money or time. So we end up being a little over budget and eating hours or renegotiating our deadlines. It sucks, but not as bad as a project failing miserably.
So all you developers out there, take the time to find out how your project works on a user’s machine in their environment. Because if everything falls apart once it’s in the hands of the users one thing is for sure. No one is going to give a damn about the budget or deadline.
Filed under: Usability
Troy Green
07.08.07 at 10:44pm
Question: For usability testings on web apps, how many users in general should a developer interview to gain a good understanding of what needs to be accomplished? For instance, let’s say the client has over 5,000 users for their application. How would you decide on which users to interview and how many?
Rhea
07.09.07 at 2:00pm
Great post! =)
Carl Smith
07.09.07 at 5:00pm
Hey Troy.
Everything I’ve ever read says 5 people get you 85% of the issues, and 20 people get you 99%.
Overall it’s not really important how many people will use the application, but how many user types. For example Amazon has to worry about grandmothers and teenagers. If both are priority users I would get 10 from each group. I’m sure other usability professionals would have their own bias.
For us, we always shoot for 10. We find we get great information at that level without killing the clients budget.
And thanks Rhea, you say the sweetest things.
Rhea
07.17.07 at 2:22pm
There was recently a discussion on comment spam and “great post” was deemed not worthy of posting. I argued the point that you’re not spamming anyone if you know the individual and were actually impressed by their post. In today’s busy world it’s difficult to respond with in-depth comments on every post that’s read, so do you not take the time to provide feedback or fight the urge to be seen as a spammer??
Carl Smith
07.17.07 at 4:26pm
Wow, I disagree with that phrase being content spam if you know the person.
It made me happy someone thought it was a good, nay, a great post!
So I thank you Rhea.
P.S. Once you were linked from our site did your page rank go through the roof or what?
Rhea Drysdale
07.17.07 at 5:27pm
lol, Google contacted me to talk about the peak in traffic. ;)
Troy Green
07.17.07 at 8:51pm
Hi Rhea-
Your post was much appreciated. I’m actually a newbie in web apps and your comment helped me feel more comfortable with asking more questions down the road that might be similar to my first post. Thank you! You guys have a good evening.
Rhea Drysdale
07.17.07 at 10:46pm
Troy, glad it helped! :)
Sparking a conversation online is a great way to meet new peeps, learn about some amazing resources and feel a little more human. If you make it to the next Refresh Jacksonville meeting introduce yourself, I’m still a newbie to this group.