Lean Software Startup: Customer Development (lecture)
MA27 Lean Software Startup
Jan 20th, 2016
1. What is customer development?
2. Why is it needed?
3. Principles of customer development
4. Barriers of customer development
5. Customer development and this course
WHAT IS CUSTOMER
A method of validating demand for a (software) product by
developing hypotheses (”guesses”) and then testing them
objectively ”outside the building”. The goal is to build a scalable
startup by finding a repeatable sales model.
The conceptual framework is developed by Steven Blank,
and it’s most often applied to enterprise software startups.
Usually takes the form of semi-structured interviews
You want to avoid both false positives - i.e., getting the
impression your idea is good although it sucks; and false
negatives which is to conclude the idea is bad although in
reality it's not (=objectivity).
You want to create pull instead of push. For that, you need
to first find the right product and market (=research
purpose). (Meta-assumption: Market risk is higher than
RQS FOR CUSTOMER
• Is the problem you think of real? (problem validation)
• Does it make sense to solve the problem in terms of business
opportunity? (i.e., do enough people have this problem? Are
they willing to pay for solution?) (problem validation)
• Is your product a good solution for the problem? What are the
alternative solutions? (i.e., what is direct and indirect
competition?) (problem validation)
• Is the customer you think of the right one for you? (Is he
experiencing the problem you think? Is he willing to pay for a
solution?) (customer validation)
IN SOFTWARE STARTUPS,
MARKETING AND CODE ARE
A tech team needs to find a good
marketer that can translate customer
needs into technical requirements (or
do it themselves).
IS YOUR IDEA ANY GOOD?
VITAMINS VS. PAIN KILLERS
• Vitamin: a product that no-one hates, but does not love
either. Such a product will most likely fail, because people
don’t want to pay for it.
• Pain killer: a must-have product the customer cannot live
without (and is willing to pay for). If your product would
disappear tomorrow, the customer would be really sad.
• Most startup ideas are nice-to-have. They will fail
because of that.
• Willingness to pay is a must. Without revenue, it’s not
a business but a hobby (or non-profit).
TWO WAYS TO COME UP WITH A
• Deductive: spot a mega-trend or a lucrative market
from ”helicopter view”. (VCs care about this, since they
think of market size and big changes.)
• Inductive: spot a everyday problem of a customer and
solve it. (Then later on you will find out how scalable
business it is.)
WE ARE ALL BOUND TO
Hence, our ability
to come up with
ideas is limited.
As you gain more experience of a specific industry – often
referred to as domain-specific expertise – you are able to
recognize opportunities you never thought of.
That is why many students learn about business
development, work in the industry, and only then create
their own businesses.
(Keep all this stuff in mind,
even if you don’t plan to
start a company right now.)
PIVOTS AND PATH DEPENDENCE
• The end result is often different from what was initially
• However, the initial idea influences how the project will
• This is called path dependence
Consequently, startup parameters matter. Although a good
team can constantly improve, by choosing a bad idea it is
worse off compared to another good team choosing a good
THE DANGER OF
…even if you bootstrap,
life gets in the way!
1. DON’T FOCUS ON
Don't ask about your product, ask about their problem.
Wrong question: "We have this product A - would you use
it?". Right question: "Do you ever have this problem B?"
[that you think the product A will solve]
“Abstract your problem by a level. For example, if you
want to know whether someone will use a healthy lunch
delivery service, ask about ‘lunch’” (Cindy Alvarez)
2. LISTEN, DON’T PITCH
Pitching is for other times - you DON'T need to sell your
product to this person, you only need to hear about his or
“Shut up for 60 seconds. This is a LONG, LONG time and
it feels awkward. It also forces the person to go beyond the
short (and probably useless) answer and go into detail.”
Repeat what he or she says - oftentimes people think they
understand what the other person is saying, but they don't.
Only by repeating with your own words and getting them to
nod "That's right" you can make sure you got it right.
4. TAKE NOTES
Take notes or better yet, record the interviews. You don't
want to forget what was said, but without notes you will.
Sometimes looking back at your notes or listening to
audio recordings, you’ll realize new angles.
5. ”WOULD’VE, SHOULD’VE,
COULD’VE… BUT NEVER DID”
Don't ask "would you" questions, ask "did you" questions.
People are unable to predict their behavior, but they can
fairly well tell what they have done.
6. AVOID LOADED
Avoid loaded questions. False: "Is this design good?“ or
“We spent a lot of time doing this mockup. What do you
think about it?” Correct: "What do you think of this
design? (…) How would you improve it?"
7. AVOID ”YES OR NO”
Avoid yes/no questions. What would you learn from them?
Avoid yes/no questions. Whichever one the person
chooses, it's probably not useful for you. (Cindy Alvarez)
8. DON’T VALIDATE,
Focus more on disproving your idea rather than validating it. In
philosophy of science, this is called falsificationism. It means not
claim can be proved absolutely true, but every claim can be
proved wrong. Rather than wanting to prove yourself right (at the
risk of making a false positive), you want to prove yourself wrong
and avoid wasting time on a bad idea.
Remember: most startup ideas suck (it's true - I've seen
hundreds, and most will never amount to real business - be very,
very critical about your idea).
9. REACH FOR
Make many interviews.
Many = as long as you notice there are no more new
insights. In research, this is called saturation. You want to
reach saturation and make sure you've identified the major
10. WAIT FOR IT…
Only in the very end introduce your solution. Then ask openly
what he or she thinks about it: "What do you see problematic
about it?" Also ask if they know someone who would like this
“Avoid talking about your product or your ideas until the end - but
then DO give the person the opportunity to ask you some
questions. This is NOT a chance for you to sell your idea, it's just
an equalizer. You've been asking questions the whole time, now
it's their turn.” (Cindy Alvarez)
FIVE BARRIERS TO
1. ”Build it and they will come” (you don’t need research)
2. Internalizing problem (you already know the potential
3. Confirmation bias (you want to be right)
4. Social desirability bias (they want to please you)
5. Recall bias (they remember wrong)
The point is that you want people to tell you honestly
what they think, and you want to interpret it in an
objective way, not being too fixed on your initial
assumption (i.e., hypothesis). Be ready to change your
opinion, like Gandhi advised.
SUNK CODE FALLACY
Tendency of people sticking to bad choices because they have
costed time and effort.
AVOID THIS: ”I have gone through so much effort to create this
feature, so we’ll keep it.”
Rather than writing perfect code, hack things to work at minimum
• If the product has genuine demand* (small chance), you can
restart and do it properly
• If the product doesn’t work business-wise (big chance), you
anyway have to forego your earlier work.
*Genuine demand: People are willing to PAY for the product. (In most
cases, forget about indirect monetization – it is much harder.)
IS IT ENOUGH TO MAKE WHAT
Mr. Paul Graham: “Make things people want.”
Mr. Joni: “Make things people want… to pay for.”
The first one targets users by providing a useful
service with the assumption that money can be
generated later on in some way. The second one
goes straight after business. Both can work, but the
second one is easier!
Gate 1: A list of hypotheses placed on a BMC ("educated
guesses") [picture / text]
Gate 2: A description of customer development methods applied
in your case (e.g., data collection method, a list of people to
interview, interview questions) [text]
Gate 3: Presentation of your results [Powerpoint]
• List of hypotheses:
• H1: Customer group X experiences problem Y
• H2: Customer group X is willing to pay for a solution
• H3: Our solution for problem Y is feasible for customer group X
• H4: Our product’s core benefit for customer group X is Z
• H5: Potential marketing channels were we can profitably reach customer group X are A, B, and C,
and we can apply tactics A1, B2, and C3
• H…n: […]
• Method: Interviews, goal: 12 interviews
• List of interviewees:
• I1: A member of customer group X
• I2: A member of customer group Y
• I…n: […]
• List of interview questions
• IQ1: In your daily life, how are you doing [a thing relating to Y]?
• IQ2: In your daily life, do you experience problem Y? On a scale 1-10, how big of a problem is it?
• IQ3: How do you solve problem Y?
• IQ4: What would be the ideal, imaginative solution for problem Y?
• IQ5: How valuable would this ideal solution for problem Y be to you? (in words, in euros)
• IQ…n: […]
EXAMPLE QUESTIONS FOR
(CF. CINDY ALVAREZ)
• How do you currently deal with this task/problem? (What solutions are they
• What do you like about your current solution/process?
• Is there some other solution/process you’ve tried in the past that was better or
• What do you wish you could do that currently isn’t possible or practical?
• If you could do [answer to the above question], how would that make your life
• Who is involved with this solution/process? How long does it take? Where
does it take place?
• What is your state of mind when doing this task? How busy /hurried /stressed
/bored /frustrated? [learn this by watching their facial expressions and listening
to their voice]
• What are you doing immediately before and after your current
• How much time or money would you be willing to invest in a solution that made
your life easier?
FIVE QUESTIONS I’LL
1. What were your hypotheses?
2. How did you decide to validate/disprove them?
3. How many informants/respondents did you have?
4. What were the results? (What did you learn? How did your
initial plans change?)
5. Based on your results, what are the next steps?
1. If you have to read one book about this topic, read this
2. If you want to read another book, then it's this
If you need to read a third book, then you should stop doing
a startup and become a researcher (welcome :)
(Reach me at LinkedIn)