Customer Discovery Done Well
Conducting customer discovery discussions is an art that needs guidance and practice to deliver useful insights.
Food recipes with a detailed background of a dish are interesting. But when it comes to cooking, scrolling to the instructions is annoying. So I will immediately share the core questions you can use in customer discovery interviews.
The Problem Interview Questions
These will help you obtain concrete insights and avoid unhelpful (sometimes misleading) compliments. They have been amalgamated from The Mom Test book, Ash Maurya, and Y-combinator Startup School.
Warm-up question: How does area/topic X affect your life?”
— Provides background information that can help you evaluate the relevance of the other answers. It also allows for expanding outwards if it turns out they do not have direct personal challenges in the area.
1) What challenges do you face with X?
— Finds out if there is an acknowledged problem.
2) Tell me how you deal with X today?
— Sets out the current jobs-to-be-done and existing solutions? (i.e. the competition)
3) Why is it important for you to do this?
— Understanding the core needs/motivations.
4) What is the hardest thing about doing that?
— Exposes the crux of the challenge.
5) Why is it hard?
— Improves understanding of the challenge.
6) When and how often do you have to do this?
— Finds the motivational triggers. The frequency of a challenge affects motivations for change. Infrequent problems need to be more substantial.
[Note: The best problems for startups to solve are frequent and intense.]
And the most important question:
7) What actions have been taken to find better solutions?
— Finds out if the problem is critical or low-level.
If little or no actions have been taken, then the idea is a no go for this entity.
Be flexible with the questions, as you don’t want to interrupt if a line of insights is evolving.
In responding to answers keep in mind: Compliments are a curse (deflect/ignore them) & Concrete is critical (people can’t “lie” about practical things they do)
Never say ever (i.e. avoid “Would you ever…” type questions as the answers are meaningless since they will be vague.
[Side note: These are also great questions to use when sparing with a founder. Ask them to pretend to be their own ideal customer and the process will be illuminating. It succinctly highlights what remains to be validated.]
The reason for this article
I have recently been mentoring a university student interested in pursuing a startup. This is for many months rather than the usual 1-to-1 sessions with existing founders. And it is particularly stimulating as it gives me a reason to more fundamentally review how I would approach the startup challenge from a blank page. Can I impart helpful knowledge learned from my own adventures that will reduce their risk? You can never completely remove risk but as with mountaineering you can practice and plan.
The first leg of the startup journey is the customer discovery process. At least it should be. Many first-time founders only pay it lip service. It usually takes one or two startup crashes before they take it seriously.
Customer discovery guidance is now concrete
It is 20 years since Steven Blank first published his seminal book on the topic of customer discovery: The Four Steps to the Epiphany. It is something of a tomb of a book that requires considerable effort to digest. But Blank’s call to “get out of the building” and talk to customers is widely acknowledged as something entrepreneurs must do. The challenge then is that without a concrete approach to guiding those customer discussions, the information gained can be unhelpful and even detrimental.
Ten years later in 2013, Rob Fitzpatrick published The Mom Test book which provided detailed practical advice on how to conduct customer interviews that provide accurate and useful feedback. Ash Maurya had also published Running Lean in 2012 which gave practical advice around implementing recommendations made in Eric Ries’s The Lean Startup book (itself inspired by Blank’s customer development principles).
Another ten years later in 2023, we now have helpful YC Startup School videos on how to organise and conduct customer discovery interviews. Here is one from last year by Gustaf Alströmer on How to Talk with Customers. This has some useful tips on how to contact people.
So ignorance on the topic of customer discovery is now no longer acceptable for a serious founder.
Problem spaces and not solutions
The customer discovery journey should first be approached by developing an obsession with a problem space. But it is necessary to start with a hunch. This points you in a research direction. The sooner you lock down a solution, however, the greater the danger of unconscious bias interfering.
Here is a collection of insights on how to pick an initial startup idea. Rather than waiting for the “perfect” idea, on some level it only needs to be good enough to get you going, practising, and learning. Like all the other high-risk careers (e.g. singing, acting, writing, sport, etc.) the only way to know if you are any good is to start doing. And, unless you are blessed by the fates, most likely your first startup or two will be practice runs.
[Side note: To decide whether a startup idea is worth pursuing it is worth contemplating the word worth. Something of worth has value. A startup is more likely to be sustainable if it can generate direct monetary worth. An impact startup can be worthy and thereby garner donor support. Ideally, the startup idea will have worth and be worthy. Investors nowadays are keener than ever to support startups that will also have positive environmental and societal impacts.]
Who and what gets you excited
Creating a successful startup is exceptionally hard. To sustain the enthusiasm you need to be both passionate about the space and, perhaps even more importantly, the people who will be your customers.
So before you even start the customer discovery journey you need to narrow down the search to spaces that excite you and the people who inspire you.
If you are genuinely enthusiastic about a space — and the people in it — you will find it easier to create social gravity to attract the necessary connections. Reaching out to people takes considerable psychic energy since you will be faced with a great deal of indifference and outright rejection. If you also need to build up your credibility as a knowledge expert (e.g. via blogging, hosting events, research, etc.) this takes considerable time and resources. If you don’t have the passion the challenge is likely to be overwhelming.
The above is the reason people talk about following your passion. But for a startup to be successful, it must also meet other demands: desirability, feasibility, viability, etc. Works of art can live and be valid with a small number of admirers but startups will die and the work fade away.
Refer again to the startup idea insights for further information on how to evaluate the quality of a startup idea.
Getting down to business
Let’s now assume that a number of worthwhile startup ideas have been considered. And a decision has been made to validate the most promising.
Define the startup hypothesis
The initial (problem) interviews are all about validating the startup hypothesis in relation to potential risks.
There are three critical risk areas being explored in the customer discovery process: Product risk: What are you solving? (Problem); Market risk: Who is the competition? (Existing Alternatives); Customer risk: Who has the pain? (Customer Segments).
[Side note: Risk validation is an expanding and ongoing process. The risks remain grouped within the Product/Market/Customer triarchy. For example, future questions to validate will be: How to identify early adopters? What is the minimum feature set needed to launch? Does the founding team have the technical capacity? Will customers pay for the solution? What price will they bear? etc.]
Together with co-founders or mentors, agree upon a startup hypothesis in a form like this:
We believe that Y people/companies might need something like X to achieve/solve Z, as measured by W.
Use “achieve” to describe a desired goal that the startup would try to help customers reach. Use “solve” to describe a specific problem that the startup is trying to address. The ultimate difference is of little consequence but there is a rule of thumb that during economic downtimes customers are looking to make savings (by solving problems), but in boom times they are looking to expand (by achieving new goals).
So now we know to reach out to Y people or people in Y companies. Always trying to evaluate who the solution would be for while paying attention to who are the purchase-decision makers.
Take a deep breath and commit to the practicalities
Aim to set up and conduct 30–50 interviews within 2–4 months. This requires considerable effort and also needs to be costed. But always keep in mind the Startup Rule of Pi: everything takes 3.14x longer than planned; costs 3.14x more than expected; and makes 3.14x less than you hoped. So chances are within 1–2 months you will not be able to organise and conduct ~40 interviews with relevant people unless you are already highly connected and respected in your space.
In regards to the above, it is better to subdivide the interviews into groups of 4–5 within different potential customer segments; i.e. 4–5 people show a clear and burning need suggests that there is potential. So there is no point in conducting 10 or more interviews in a segment that is not looking promising.
[Side track: I learnt about the Startup Rule of Pi from Moaffak Ahmed and it has proved more or less bulletproof. It is common for project managers to double provided estimations—especially in IT. I think the reason for the required additional pessimism with startups is that they also operate within contexts of great uncertainty, as well as overoptimistic ideas of efficiency.]
Acknowledge the costs of customer discovery
My driving instructor said; “Go in haste, repent at leisure.” This pithy statement holds up in so many parts of life. And losses related to startup endeavours can be particularly painful.
Albert Einstein said; “If I were given one hour to save the planet, I would spend 59 minutes defining the problem and 1 minute resolving it.” So why when it comes to startup development do most founders spend the equivalent of 1 minute on the problem defining stage?
The answer is that novice founders don’t fully appreciate how hard it is to do customer discovery effectively. And that it requires considerable practice and resources.
Anyone with a modicum of startup interest should have digested the idea of getting “out of the building” to speak with potential customers is essential. Many simply choose to ignore this advice and rush in all dev guns blazing. Occasionally they can get away with this approach: particularly if they are making something for their own use, or have enough internal development capability to keep on stumbling until they find product/customer fit.
Most will do some amount of customer discovery activities but it will be superficial and done badly. Those who try to do it properly might get discouraged by how much work it takes.
There is already a high chance you will find the customer discovery stage overly challenging, i.e. you are struggling to get enough interviews with relevant people, or it is taking too long. This is a “healthy” early sign that the problem space (or how you are positioning it) is not interesting enough to potential customers (i.e. the pain—if they even have it—is not frequent or intense enough). Or you don’t personally have enough social capital or credibility to engage the right people. If this is the case you will need to either rethink the problem space or invest more in building up relevant connections (e.g. start and engage in online discussions, conduct academic research, attend/host industry events, etc.).
Problem interview / Solution interview
Books like Running Lean (Ash Maurya 2012) advise founders to return after the problem interview for a separate solution-oriented interview (with a demo). Unless you have strong personal connections or discover highly-motivated potential customers, it is unlikely that you will get both a customer discovery and solution interview with the same people. So consider aiming for around 30–40 customer discovery interviews and leave 10–20 connections for solution-oriented interviews.
Ask for 30 minutes of their time: A few minutes for bookend pleasantries and 25 minutes for the core insight discovery questions.
Given the challenge of getting time with relevant people, it is not uncommon for founders to try and combine both the product and solution interview into one session. Only if you are satisfied the key insights into their day-to-day realities have been obtained, then introduce a provisional X solution idea. It is anticipated that solution X will evolve based on the feedback during the customer discovery process but try to be systematic about adapting it, e.g. after ten or so interviews otherwise the sample sizes will be too small.
Identifying customers via motivation, ability, and triggers
Maurya has helpful insights into the importance of customer triggers. Startups are about behaviour change, so you should consider Fogg’s behaviour change model: Behaviour = Motivation + Ability + Trigger (B=MAT). This is useful to think about when it comes to potential customers: Who has the motivation and ability (to use the solution) and “when” will they be triggered to try using it?
For example, a Japanese startup friend, Tatsu’s first big success was Mamari (a forum for pregnant ladies and new mothers). There is a clear life-situation motivation and trigger to try using the product: becoming pregnant for the first time.
Customer demographics improve the chances of finding people with shared needs and desires, but it is better to think of contexts in which people have similar needs and desires regardless of their demographics. Then think about what triggers the jobs-to-be-done that lead to their mutually desired outcomes.
The Problem Interview Questions
These are worth repeating here as they need to be learnt by rope so they can be yielded skillfully like a customer discovery ninja! And for reaching this far you get a bonus question 😊
Bear in mind the hypothesis: We believe that Y people/companies might need something like X to achieve/solve Z, as measured by W.
1) What challenges do you face with Z?
— Finds out if there is an acknowledged problem.
2) Tell me how you deal with X today?
— Sets out the current jobs-to-be-done and existing solutions? (i.e. the competition)
3) Why is it important for you to do this?
— Understanding the core needs/motivations.
4) What is the hardest thing about doing that?
— Exposes the crux of the challenge.
5) Why is it hard?
— Improves understanding of the challenge.
6) When and how often do you have to do this?
— Finds the motivational trigger. The frequency of a challenge affects motivations for change. Infrequent problems need to be more substantial.
7) What actions have been taken to find better solutions?
— Finds out if the problem is critical or low-level.
8) How do you measure if you are succeeding in achieving/solving Z?
— Finds out the key W metric that guides them.
Are you brave enough to quit before you start?
The killer hidden within the customer discovery process is that it is a costly process that someone pays for. And the more you pay for it yourself the more you will be tempted to ignore negative signals since you have already invested resources (time, social capital, psychic energy, etc.). The good old sunk cost fallacy (if you don’t know what that is just ask ChatGTP).
To set up and carry out a full-on customer discovery process (with 2–3 interviews a day) will not be possible with a full-time job, so it is likely you will have to quit your job. One option with remote working possibilities would be to extend the process. But you would need to either be open with your employer or take the risk that word wouldn’t get out that you are actively exploring doing a startup up. Tricky.
Imagine deciding to look for new employment after only a few months from announcing to the world you are actively exploring a startup idea. If you go ahead anyway and later fail at least you won’t look feckless. It is hard to pull the plug once you have publically taken the first steps. Perhaps this is a core reason why customer discovery is done so badly. And also why being a student is a great time to become a founder.
Remember: Go in haste, repent at leisure. So if you are not in a rush to become an entrepreneur, there are other ways to fund the customer discovery process that have additional benefits.
Customer Discovery via employment
A classic way to do customer discovery is to work within a space that you find interesting. In this way, you will get direct insights into the challenges while building up relevant network connections and credibility. Just keep your eyes and mind open to avoid getting sucked into the “this is the way it’s always been done” way of thinking. The founders of Uber had no experience in the taxi industry and the founders of Airbnb had no experience in the hospitality industry, so they could more easily approach the needs based on first principles.
Industry-experienced founders working on niche industry needs are more likely to succeed because they understand the context well and are already well-connected. The drawback of course is the “limited” market size.
Customer Discovery via student research
Academic research involving external people/companies is a good form of customer discovery: structured, supervised, and supported. With ostensible and declared efforts to avoid bias. And by default one needs to be highly interested in a problem space to commit to years of project research. The founders of Relex, Mikko Kärkkäinen, Johanna Småros, and Michael Falck, were all supply chain scientists whose research informed the foundation of the startup.
If you enjoy academic work and are obsessed with a space this is potentially a wonderful option. The major drawback from a startup founding point of view is that you are locked into a particular space for many years. And if your research indicates you are not working on something that would form the basis for a good startup idea then you are back to square one. So only consider this approach if the research and related qualifications are of primary concern.
Customer Discovery via doing a startup
A common way to “do” customer discovery is to dive into product development and fundraising only to find out along the way the real customer needs one should have been addressing. This is obviously a high-cost approach and few have the resources to try again. So not recommended. But push comes to shove, if you really want to become an entrepreneur it is better than doing nothing due to analysis paralysis.
If your team has technical capabilities and low burn you can get away with making many mistakes. Michael Seibel often recounts how Twitch started out as a failed personal 24/7 streaming service (Justin.tv) created for co-founder Justin Kan. And the problem being solved had to evolve over many years.
To conclude
If you are a first-time founder you will most likely pay superficial attention to advice like this and take some wild swings. But after a cold dunking in the harsh reality of startup life, you will undoubtedly come back to review the classic texts on this topic and try again after finding more solid ground the second or third time around.
That said you never remove the risk; you just reduce it. Hundreds of people die mountain climbing every year. But if you are compelled by the love of climbing mountains you can train, prepare, and plan (fingers crossed) to avoid becoming one of them.
And at the end of the day, you will always need to start out on the startup path knowing that there are risks ahead. And this will always require a leap of faith. The point of this article is to reduce the width of the chasm you have to jump. And better evaluate if you have the ability to jump it.
My grandparents’ had a coffee mug that I used to find intriguing. On it was written LOOK BEFORE YOU LEAP. That’s fundamentally what the customer discovery process is all about.
Further reading/watching:
How To Talk To Users | Startup School: Alströmer, G. (2023)
The Mom Test: Fitzpatrick, R. (2013)
Running Lean: Maurya, A. (2012). (Modified in this Leanstack blog update)
The Four Steps to the Epiphany: Blank, S. (2003)
