Getting input from customers is vital to making sure we are building the right solutions. The tricky part is getting enough feedback to have a complete picture of the issues the majority of customers and users face. Ask too few users at too few companies, and the result is a picture very specific to their particular situation. Ask too many users at too many companies, and you spend so much time doing research and analyzing the data that by the time you finish, you’ve missed all your milestones and are over budget. You need to find the right balance.
Building standardized software is a bit like trying to create one meal that makes everyone happy at every time of day regardless of the weather and independent of dietary restrictions. That’s a really hard task. Just ask any airline.
If you take one company’s or user’s particular wish list too much to heart when building enterprise software, you end up with non-essential functionality cluttering the screen. Do that again and again and you get a massive case of feature creep. If you don’t get enough feedback from a wide enough range of customers, you miss opportunities to optimize the user experience and deliver ROI through innovation.
Let’s take an example. A product team in the area of say… human resources, only has time to visit two users at two companies. By observing the first user and asking her questions, the team discovers that her company requires her to name a colleague who will take over her tasks during her vacation. So it would really help to have a feature integrated into the application which checks the availability of that colleague. But wait; at the company of the second user, this is not required. So user number two doesn’t need that feature at all. What he would love to have is an easy way to pick the person who will approve his leave request, because at his company the approver changes regularly. In each case, the feature that one user would need would clutter the screen for the other user. What to do? Clearly, more users need to be involved to increase the data points to decide what to build and how. But how many more users?
Now before we go any further, I’d like to say for the record that doing a little bit of customer engagement is certainly better than none at all! Even this little exposure to only two end users has great potential to spark excellent ideas to solve other problems the team is struggling with or give perspective on aspects they hadn’t even considered before. It may also provide an invaluable opportunity to empathize with the users. However, in our concrete example, the team only has a small set of data points and the score was tied 1-1 between the “flexible approver” and “calendar integration of colleague” features.
So, back to our question − how many customers and users do you need to observe and interview to get sufficient input for standardized software? You didn’t really think I was just going to give you a number, did you? The answer is, as unfortunately is so often the case, it depends. Projects and products differ. But here is a rule of thumb: The more conflicting the first findings are, the more research you need to do. And that can mean that after interviewing five users, you are done because you keep hearing the same thing. But it may mean that you have to interview many more because the feedback is so diverse.
My words of advice are to stay flexible. Continuously review the data you have gathered in order to recognize early patterns and major changes in the feedback you are getting. Change your customer engagement plans accordingly so that you do as much user research as you need, but not more.