What and Why, and How they Fit
One of the hardest parts of working on this book has been trying to reconcile WA and UX practices and perspectives. I'm convinced that their sum is far greater than their parts. But how do they fit together?
I'll be presenting a keynote on this very topic a couple times next month—once in Philadelphia, the other in Hamburg—and we're covering it in the book. But let me offer a couple simple points for now.
Web Analytics really uncovers the what of a given context. The User-Centered Design methods we UXers rely upon explain that context's why. Let's say you're operating an online store. If 43% of users fail to complete the account creation process, that's what's going on. Contextual observation might explain why they're failing.
What and Why. It's hard to really justify focusing one without the other. And, in a nutshell, here's how they fit:
- Web Analytics needs User Experience: It's not much use to know what is happening if you don't know why.
- User Experience needs Web Analytics: You can't know why things are happening if you don't know what is happening. (Got some good discussion going on this point in Facebook after Tweeting it. Follow all that?)
OK, snarkers, so your reaction might be a big old "DUH!". Well, how many of your research and design groups have successfully integrated UX and WA? :-)
Comments
I couldn't agree with you more. I work on an internal team and have been asking for any kind of WA. After much asking it ends up we only have some rudimentary stuff. BUT it has given us some insight, notably these:
- Finding potential users to interview and/or survey.
- Where people spent the most time.
- Where they get stuck.
- Where performance was lacking.
With that basic info we now have some idea of where to start. Before this, the WA went into Powerpoint presentations to show usage. But no one actually discussed why things were the way they were. That is what UX brings.
Posted by: John Labriola | April 24, 2009 2:24 PM
Frankly, I never perceived the two to be separate practices for the reasons you described. Backtracking from your point, I see the two intertwined because in justifying WA practice, why does one need to know what is happening? Usually the answer to this is pretty simple (i.e. that 43% failure rate is bad for business) but it's surprising how few clients can pinpoint an initial what-why combination upon which to bring WA practice to bear.
As far as group practice and research focused around this, it unfortunately only happens by accident where I work now. Hope that changes soon.
Posted by: Matthew | April 24, 2009 2:27 PM
In the past, before performing any User Research my team would always have the client gives us any WA information they had available. This allowed us to know exactly what we ares we needed to focus on during the interviews and studies. It is a great tool for figuring out the focus and scope of research and design project.
Posted by: Brad Nunnally | April 24, 2009 2:32 PM
Great post. You can tell it's a good one when people react with a 'duh' - that's when you've said what they were thinking but couldn't put into words.
Posted by: Ryan Kennedy | April 24, 2009 3:11 PM
From personal experience I couldn't do my job without WA and UX. I often find myself working with clients who either a) have little budget or time for a full blown UX approach or b) focus all their attention on meeting business and technical requirements.
- I use WA (usually Goals/Conversions) to highlight potential issues (Houston we may have a problem) and focus the area of research.
- If that not quite cutting it I run some quick unmoderated remote usability tests (time on task, error/abandonment rates etc...)
This usually gets my clients listening. Then the observational stuff comes in to get the heart of the problem.
At UX Week 2008 Don Norman spoke of the UX community speaking 'executive' in order to gain greater acceptance in the upper echilons of organisations. If used wisely, WA often starts that conversation.
I'm not sure if it's a case of them playing nicely together. WA is something I use alongside other UX activities (diary studies, questionnaires, wireframes etc...). It's a valuable addition in my UX toolbox.
I'm really looking forward to the book. Indy Youngs Mental Models and Luke Ws Form Design are both books thats I learned loads from. I have no doubt I'll be buying the book. For me and for my some of my clients.
Craig
Posted by: Craig Melbourne | April 24, 2009 3:16 PM
My first comment is that metrics for any application (the "what" in your discussion) MUST be developed with the "why" in mind from the beginning. I consult on Business Process Management, which is all about the customer's experience of both the delivered product or service And the experience of interacting with the process to get to that outcome.
Most folks in my industry instinctively develop metrics that deal with the mechanics of process -- the flow. This is exceedingly important, but it's only part of the picture. It's analogous to viewing processes as highways and design of process as traffic engineering. Necessary? Yes. Sufficient? No. Business processes aren't highways. Ultimately, you have to be concerned on whether the destination was the correct one, the trip was safe, and the ride was enjoyable. These can all be measured with a little thought to context BEFORE metrics and feedback systems are developed.
Another related problem I've noticed is the adoption of "best practices" for metrics (or analytics) which always has the potential for misapplication. The story I made up for my clients involves brother and sister entrepreneurs. The sister raises and shears sheep for high end yarn production. A metric of paramount importance to her is the weight of the wool removed per sheep per shearing -- all of her other metrics are essentially designed to measure variables that optimize this result.
Her brother manages the same basic business process. He owns a chain of children's haircutting salons. He wanted to improve business results and figured since his process was the same he'd just adopt his sister's metric. He now has all of his store managers focused on maximizing the amount of hair removed per kid per haircut.
It's true. I swear.
Posted by: Dave Kenny | April 24, 2009 4:18 PM
I think this is an excellent approach.
In order to get the whole customer path you need qualitative and quantitative insights, from UX you can collect both, but thanks to WA those insights could be more powerful because you just don't track some users, you track the your site.
Another usefull source, focused on mobile devices are those software applications that entitle you to collect mobile phone usage data. This software applications run on the background and user does not have to interact with it. We have use it and we have got very usefull insights. We all know sometimes users said they do many things that really they don't. It's an objective source.
Posted by: Camilo Sabogal | April 25, 2009 5:28 AM