Some Learnings on Customer Discovery at Early Stage Startups
Some thoughts…
I’ve been thinking a lot about how to effectively use customer feedback to guide high-level strategy early on in a company’s journey. Now, this isn’t how to use customer feedback to decide what features to build - it’s how to use feedback and what the market is saying to decide product positioning, core value drivers, and GTM strategy.
In this post, I wanted to share my thoughts on customer discovery => how I go about doing it, the pitfalls I’ve discovered, and some ways to frame the feedback received to hopefully avoid those pitfalls.
What does customer discovery look like at a startup?
Customer discovery at a startup is very different from discovery at a mature company. The answers you are trying to uncover are much more existential in nature and broader in scope. You also are less likely to have a large customer base that you can farm for these discovery calls. On top of that, many of the calls you have with potential customers are sales opportunities as well, so it’s hard to be truly “first principles”-driven when running customer discovery at a startup. So here’s how I define customer discovery at its highest level: talking to a broad range of potential customers in an open-ended way to find patterns in what value props resonate the most for a targetable segment. Let’s break that down.
Broad Range of Potential Customers: Typically when sourcing potential customers to talk to, I go very wide. Think “has marketing in their title and works in B2B SaaS” wide. This is because you want to identify new insights, so you want to go in with an open mind. This is also helpful to hone in on what your target market persona might look like. Does it resonate with all marketers, or just some who have a very specific problem or work at specific type of company.
Open Ended: Ok, this is a fun one. We’ve all seen these IDEO videos about their innovation process where really open-ended user interviews result in an amazing insight and all is well with the world. I have rarely seen this happen in early stage startup land for a very simple reason: you simply don’t have the funds and time to pay people to sit there and answer these very open ended questions while you figure things out. It’s already a challenge to get 30 minutes of time from really happy customers, let alone get more time from a random persona from a broad cohort. So what do I mean by open-ended? I mean asking targeted, but open-ended questions. The goal here is, you want to figure out if what you’re building provides value and is important enough for them to buy it now. If it’s not, you want to figure out why, and what they would buy instead. That’s why I believe for early stage startups, you should be showing your product during every call and refining how you talk about it with the learnings you uncover in each call to get actionable feedback. However, you should make sure to ask mostly open-ended questions. One question I like using is: after seeing our product, can you describe in your own words the value it provides?
Value Props: Here’s another important thing about customer discovery - you are not trying to figure out what to build in terms of features, you’re trying to figure out what problem to solve and for whom. That’s why a focus on value props is key, even if you’re showing off your product.
Resonate the Most: Everyone will always say something is cool if it’s new and doesn’t look like shit, pardon my language. However, does it resonate? I have a couple ways to judge whether or not something resonates: body language (are they excited), “yes and” (do they take what you described and run with it). And I also just straight up ask. “Would you buy this today, if not why not?” And I also like to ask people to stack rank their priorities, and see where our value prop fits in. You’ll find that people will say things are cool, but they aren’t immediate needs.
Targetable Segment: Beyond just whether or not people want what you are building now, you also want to figure out which segment you want to target. For example, let’s say really small companies want what you’re offering, but you don’t actually want to sell into those companies. Some reasons might be that they don’t have a high propensity to pay, it’s too costly to get them up and running, it’s too costly to find enough of them to build a big business, etc. However, let’s say the market that you want to go after likes what you’re building, but wants another additional feature. In these cases, the requests of your target market are more important and relevant than those of non-target markets, even if they like what you’re doing.
My Customer Discovery Process
I’ve done some form of customer discovery now at most of the startups I’ve worked at. Typically what sparks customer discovery is a need to figure out how to focus efforts on the right things - the right customers, the right positioning, the right problems. The process I follow is quite simple:
Figure out what questions we’re trying to answer and who we should talk to to answer them
Write up a questionnaire to guide the discussion. I’ll usually do 3 sections here: an open ended one about someone’s role and priorities, a demo of what we’re trying to get feedback on, and then a section to ask targeted questions to get feedback
I do not follow the questionnaire verbatim. I let the interview flow naturally, and make sure to cover all the topics
Post interview, I’ll do a little bit of thinking and try to tie the new interview up with old interviews. Depending on the results of the most recent interview, I adjust the next interview
Finally, if we’re trying to be a bit more numbers driven, I’ll create a little pivot table that breaks down responses by type of persona
The Pitfalls
Lagging Indicators: The greatest pitfall I think of for Customer Discovery at a startup is that the insights you discover tend to lag behind the questions you have. By the time you’ve had enough calls to develop a gut feel for an answer (this typically is around 20 calls for me), you’ve probably discovered some new questions you want answered.
Misleading Signals: The hardest part of Customer Discovery is figuring out who to listen to. It’s really easy to do the simplest thing => create a pivot table of the types of personas and their feedback, and then pick the group that has the most positive feedback ratio. I’ve done this before, and this can lead to the wrong outcomes, because the most positive feedback group might not be the group you actually want to go after.
Purposefully Vague Feedback: No matter how hard you try, it’s incredibly hard to make sure the feedback you’re getting is truthful and not filtered. Further, you’re purposefully keeping things open-ended so that you can hear how your potential customers talk about their problems. You cannot expect Customer Discovery to tell you what to do or what features to build.
Framing Customer Discovery Feedback
After having done customer discovery several times, I don’t find that being overly stats driven is helpful or effective. Instead, I like using the following process:
Filter out the feedback from customers that aren’t in your target customer group - you don’t want to be misled to build for the wrong type of customer
Group what people say into general statements
Keep an eye on key words that are repeated - you can incorporate this in your positioning
Write up some thoughts on how these findings impact strategy going forward
Beyond that, I think early on customer discovery is really more about developing a good gut sense and understanding the problem you’re trying to solve. Of course, you can also do customer discovery on specific features, but that’s a different kind of process.
A Modified Customer Discovery Process
I continue to believe that short, 30-minute sessions with a broad group of personas is a critical exercise, but I think Customer Discovery should happen every time a conversation happens with a customer. We should be thinking about whether or not this customer fits into the right customer profile, and ask ourselves those tough follow-up questions. If a customer says yes, why did they say yes and how can we find another customer like them? If a customer says no now, why did they say no? What would they have to see for them to say yes? Most critically, do you even want this type of customer to say yes to you eventually?
To close things out, some takeaways:
always be talking to customers
take what they say with a grain of salt, especially if they aren’t your target persona
focus on identifying the problem they are trying to solve, and make sure it’s high up in their priority list
don’t worry too much about trying to quantify qualitative feedback