Insights for Procure-to-Pay and Finance Leaders

User testing at Tradeshift

Getting out of the building

A key principle in the Lean Startup approach that we subscribe to here at Tradeshift is to ‘get out of the building’ and learn from the users. We sat down with Laura Bonderup Jensen who heads User Research at Tradeshift and in the interview below you’ll get an introduction to how we go about getting user feedback.

Who are you and what do you do at Tradeshift?
Officially, I’m Head of User Research at Tradeshift. My task is to take the product we’re building and show it to our users. But more importantly, to make sure that we’re syncing our reality in the office with the one that exists among our users.

Why does Tradeshift do user experience testing?
When your team has its head down building something, you can easily start to become blind to any possible steps in the wrong direction. Worst case scenario, you not only build something they don’t understand how to use but they don’t even connect with it on a conceptual level and are left asking, “What is this tool for?”

What is the process here at Tradeshift?
We treat user experience testing like any scientific experiment — and that means beginning with a clear concept of what you want to test, and what your hypothesis is.

In practice we go through a number of different steps when running a new idea or feature through user research and testing.

If we are talking about implementing a new feature, we start out on the conceptual level. We have an idea and some assumptions about user behavior based on this idea and then go out to validate those ideas with our users.

This phase is very much about interviewing users, meeting small businesses, immersing ourselves in their world and developing an understanding that will make sense of the concept from a different perspective than our own.

These interviews provide a backdrop of understanding from which we can start usability testing the prototypes that the designers are working on. This is a very iterative phase. We do this usability testing in various ways.

What kind of form do these interviews take?
We sit down and test with our users, preferably in their own environment at their office. The idea is that the users really get a chance to engage with the product and provide qualitative feedback on specific feature.

We also do ‘plain’ usability testing. For this we use online services — there are some great tools out there that will record user sessions for playback and observation — which can be really powerful when you have an audio track to hear their reactions out loud and a video that shows their actual facial expressions.

What comes after that?
The last phase is when the new feature has gone live and we start tracking the larger patterns of data with real users. We now start using different kinds of tools, like our A/B testing tool that allows us to switch functionalities on and off, to track any change in user behavior. This is the first step where we really monitor the user behavior on a quantitative level, creating different types of experiments and exposing them to a large user base.

And then with those results we evaluate again. It is quite possible that the results from those experiments result in new questions and new ideas that we wish to investigate by talking to our users and then test those ideas again on a smaller scale.

What are the biggest challenges with user experience testing?
One big risk is that you lead the user in a direction with the brief and end up creating inauthentic results. This is always a fine balance as in any process.

That’s another reason it’s important to take every step possible to create the most natural circumstances. Get out in the field and meet the users in their own surroundings. Start by understanding who they are and their business — this also relaxes them and gives a more natural representation.

With that said I think anyone would be a fool to think that you can exclude yourself completely from the observation. There is always an element of participation from your side. Which is perfectly fine. And in that sense it is more important to be able to be flexible with the level of engagement from time to time.

Accept that sometimes you have to be more involved than other times. At the end of the day the most important thing is to get the users to talk and feel fairly relaxed — sometimes you have to reveal a little to get to that point.

How unusual are the patterns that come through your testing?
With Tradeshift, our aim is to become a universal platform for all businesses. That means we try and get feedback from every kind of user, from the most technically advanced through to the most casual.

As a result, you’re always seeing new and surprising reactions to the platform. People are an unpredictable force and just when you think you’ve seen it all, they’ll pull something out of the bag that you just couldn’t have expected.

But then the other thing to bear in mind is that it’s surprising how few users it takes to be able to track some really consistent patterns in behavior. How quickly that consensus appears is often unexpected.

What have you learned in your time observing users?
When it comes to conceptual testing, which is often an interview situation, much of it’s about trying to figure out the person you are sat in front of. There is no one size fits all and so being attuned to the results you’re generating can often depend on understanding that person. This can help you prioritize your findings.

The other side to it is that you have to be selective. You could fall into the trap of thinking you were testing to make sure everyone had a positive reaction every single time — in fact, negative feedback from the audience that isn’t the absolute priority can be just as valuable in helping you focus.

Finally, understanding that it is not an exercise in asking the users what they want. In the user research stage, it is more about understanding their reactions to the product and how they interact with it. That’s the way you learn more about the relationship between the user base and the stuff we create.

What recommendations would you give anyone who’s thinking about getting started with this approach?
With the Lean Startup movement and more and more people talking about the area, there’s a real wave of entrepreneurs starting to consider it for their business. My key piece of advice would be never to forget the importance of that personal context to your users.

It is time consuming but it is a process that cannot be rushed. If you don’t take the time to get to know your subjects, and solely rely on quantitative data or impersonal usability testing, you can guarantee your results will suffer. And you may not even be able to analyse why.