I work as a UI/UX designer. I have a good understanding of UI, but my UX is not as great. I spent a lot of time this week looking up topics on conversion, and ways to increase conversion rate. Most of the blogs sites I've been on, always relates back to the end user, which I understand.

My question is how to understand users with no user testing? Are there alternate ways that I can gather information on users, without doing user testing?

Also if there are any websites or references that might help with question that I have, please feel free to post them.

PS. My coding background is very basic.

    – cord
    Commented Apr 1, 2016 at 0:30
  • I have edited the question for you.
    – Michael Lai
    Commented Apr 1, 2016 at 0:40
  • Can you provide a reason why you want to understand user without doing user research/testing? Seems like a contradiction because if conversions ultimately relate back to the end user, then surely that's not something you should be omitting from your processes.
    – Michael Lai
    Commented Apr 1, 2016 at 0:43
  • A related question popped up today: ux.stackexchange.com/q/92197/21857 Commented Apr 1, 2016 at 16:48

5 Answers 5


Guess a lot

I know that sounds kind of silly, but that's what it comes down to. You can analyze

  • Sales figures
  • Seasonal volume spikes
  • Site analytics
  • Customer service feedback
  • Related case studies
  • and so on

But none of that will tell you the why behind the behavior you're seeing.

So you're left with one option: make smart, experience-informed guesses. Use A/B or multi-variate tests to see if the guess is moving you in the right direction. Keep doing that until you've optimized the last drop of conversion out of it.

If that works, you should have enough money to start user testing. Then you'll find a whole new batch optimizations.

  • What's with the down vote? We can't be honest anymore? Commented Apr 1, 2016 at 2:12
  • I'd call it 'karma' :D
    – Michael Lai
    Commented Apr 1, 2016 at 3:55
  • +1 - simply for A/B tests. Extremely useful for seeing what a user prefers without asking them.
    – DasBeasto
    Commented Apr 5, 2016 at 19:13

Depends on your definition of testing, because some UX practitioners would argue that even gathering analytics and using it to guide some of the design decisions is in a way doing user testing, so what you are talking about is probably doing testing without actual users being directly involved in the process.

The two things straight off the top of my head are: analytics and guerilla testing.

With analytics, the aim is to spot patterns and behaviours through a large volume of data that are indicators of underlying user interaction (or micro-interaction) design issues. Alternatively, you can use analytics to compare between different design concepts and approaches to find out which one creates the better user experiences (and hopefully therefore conversion rate, but keep in mind that there isn't necessarily a correlation between the two - just because users like your website doesn't mean they'll take out their wallet). There are lots of tools out there for doing analytics and user testing (Optimal Workshop is one that comes to mind, but there are plenty of others that are free/limited and will do the job without much expertise or setup).

Guerilla testing, when done in a way that doesn't focus the user's attention on the actual questions you want answered, is in essence a way to do user testing without directly involving the users. You are indirectly trying to infer information from them by some other method, and this could be through conversations or observation. What I like about guerilla testing is that the less formal and conversational you make it, the easier it is to get the information you want from the user. However, it does require that you are comfortable with talking with strangers (or observing them) without unsettling them in the process.

I would question or argue the value of obtaining user data that is not actually validated through a more rigorous or robust research methodology. But in your case the bottom line might just be to increase the conversion rate regardless of how it is achieved (be careful what you wish for though).

  • "some UX practitioners would argue that even gathering analytics and using it to guide some of the design decisions is in a way doing user testing" < Those practitioners would be wrong. Commented Apr 1, 2016 at 1:45
  • @plainclothes can you expand on that?
    – Midas
    Commented Apr 1, 2016 at 6:31
  • 1
    @Midas Analytics and user testing are complementary, but different things. Data points are events without a clear view of the process behind it. And you can only guess about the motivations that drove those processes. Directly testing with, observing, and possibly interviewing users will help fill in those gaps. Conversely, the mere fact that you are testing someone means that their behavior is to some extent tainted by the experience. You need both for the complete picture. Commented Apr 1, 2016 at 7:44

You can use some tools to understand the user behavior and so implement the modifications if necessary.

There are a lot of tools like that but to be simple and objetive, if you use the Google Analytics and Hotjar you'll do a nice combo to understand the users behavior.

With Analytics you'll have some informations like the visited pages, the time the users keep in your website, where they are going out, from where they are gettin in your website, what they most buy, what device they use and a lot more.

And with hotjar you'll know where the users are clicking, how they are navigating, you can create funnels to complement the analytics informations and do research questions.

But when you get a chance, do some real users testing just to feel how the users react to your product.


A/B testing. You're essentially user testing, but you're doing it live with pre-formed solutions and measuring to see which solution performs better relative to your conversion funnel.

For example, if you wanted to improve sign-ups on your landing page, you could run an a/b test on two different landing page versions (one original version, one new) to see which results in more sign-ups.

This, in effect, would skip user testing and you'd take a higher risk bet on a newly proposed design/solution. If the original version performs better, you could a/b test again with another new design. And then keep repeating this process which is likely much more costly in terms of time/resources. This is why user testing is important—so you take more informed bets :)


I assume when you say "user testing" you're talking about usability testing, watching people use your system as you look for problem areas.

If that's the case, then I'd say that you don't get to know users from that kind of research. Instead, you're learning about your system, where it succeeds and fails. You get to know users by observing them do their work in their environment. We call this shadowing, but some call it contextual inquiry. We do this before we start designing anything.

We're working on a business application, so we visit users at their desks and watch them work. They use our system some of the time, but they also use others. They keep scraps of paper and printouts on their desks. They compose messages in word processors instead of in our system. They get interruptions--phone calls, coworkers, etc. We ask why they're doing what they're doing, what happens to the records when they're done with them, how they receive their tasks, ...

So we're learning about users' tasks and goals and what hurdles get in the way. That is, we're learning about the users.

(The poster doesn't say why they want to avoid user testing, but I'd assume they don't want to do shadowing either.)

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.