- Time: You’ve got two weeks to get the next design out. How do you have time to conduct ethnographic research, build personas and conduct a usability test?In many organizations, even a “quick” usability test still takes 3 months to plan, recruit, execute and report on. And that’s assuming you know who your users are and what they’re trying to accomplish. The result is all too often that we just don’t have time for “UX” or the usability test is conducted after the product launch.
Tip: You don’t have to use all UX methods. Try converting giant projects into manageable hypothesis (e.g. does it take users longer to find products in the new navigation?) and focus on the most important ones. Involving just a handful of users and collecting the right data can generate statistically valid answers to your hypothesis.
- Costs: Maybe you’ve got plenty of time to get feedback from users but you don’t have a usability lab, you can’t afford an expensive usability consultant, software and compensate users.
Tip: You don’t need a PhD, a white coat and Stanley Milgram testing lab to conduct a usability test. Remote testing tools like those from MUIQ, Webnographer, Usertesting, Loop11 and Optimal Workshop allow you to collect data for a fraction of the cost it did a few years ago. Disclosure : We have business relationships with these companies.
- Tools & Techniques: You’ve found some time and budget but can you really get reliable data from a handful of users rambling on about the design of your web pages? It would be good to get the discount without paying the price in terms of reliable data.Tip: You can obtain both actionable and statistically reliable results from a few users by asking the right questions and converting comments and actions into data.
- Finding Representative Users: Your target profile is female drivers over age 50 who have bought a BMW in the last 3 months or maybe it’s hardware engineers who live in Germany (while you live in the US). Finding these people, much less getting them to agree to an interview or session, is daunting.Tip: It’s always best to gather data from users that are representative of the population of interest. With remote testing we can now more easily gather international data, during odd hours and with less inconvenience to users. Even with these tools it can still be a challenge (think doctors, commodity traders and emergency workers). However, I’ve often worked on projects where there were no statistical differences in performance and attitudinal data across any of the eight personas.
People may be different, but their reactions are often more similar than we’d think. Keep in mind that rats and primates are clearly not humans, yet they are often used to determine what will happen to us. Sometimes there are close proxies for users. For example, I’ve found employees of enterprise software companies tend to have similar domain knowledge and attitudes about icons, terms and designs as the hard to recruit HR managers.
- Deliverables: Processes are meant to reduce variability, decrease defects and increase the quality of the product or service. Processes also have a way of getting in the way of improving the user experience. When you’re having meetings to prepare for meetings for a presentation of the user research… you might have a deliverable problem.Tip: One nice thing about not having time is that you’re forced to focus on the decision not the deliverable. It may sound like a cliché, but invest the most in actions that will impact your customers and users and less on lengthy documents and presentations. Sidestepping the process is bad, but sidestepping the user is even worse.
These are challenges for a reason. There aren’t easy and obvious on-size-fits-all solutions. If you’re looking for more solutions than problems, I will be hosting a webinar on Feb 28th 2012 in which I’ll provide some examples of how to work through these challenges.
I’ll walk through an example of how to use Lean UX methods and tools to squeeze as much meaningful and statistically reliable data under tight budgets and deadlines.