Building a high growth product. 10 founder mistakes & their quick fixes.

Based on 8 years of consulting early-stage startups

I have consulted a dozen of early-stage startups across various industries and business models. All of them had around 10–35 employees, already had some sort of traction and had problems with growth.

These are the 10 patterns I have seen with all of them:

1. Selling to everyone instead to a clear customer segment

Startups need to find a niche, where they can outperform incumbents that already have a 10 year head start.

Take Tesla, they first built a “sexy sports car” for people who care about stuff like 0 to 100 acceleration — something where electric cars have an edge over traditional cars. After they had enough traction within this customer segment, they went further down towards mass-market.

Getting this right is everything but trivial, however it has by far the highest leverage for success.

Next step: Talk to your best performing customers and look for patterns. Why are some users struggling and others happy? Why are some using only feature A+B and others only B+C? Find the patterns and you will find the segments.

Note: Customer segmentation is never about demographics (e.g. industry), but around customer-problems and jobs to get done.

2. Talking to customers, but not asking the right questions

A proper customer interview takes 30–45 minutes and it should be 90% them doing the talking.

If you want to get true insights, keep your pitch deck at home. You need to focus on them: their context & initial starting point.. what they are trying to achieve & why.. what triggered them to look for a solution.. what solutions they were already trying out.. why it was not working out.. why they were trying out your solution.. what their initial doubts were.. how it was for them to switch from X to your product, etc..

Next step: Schedule 30+ minute calls with 7 recent signups who already show strong product usage. Make the call about their journey that led them to you. Don’t talk about the product.

3. Optimizing for a few loud power users instead for new customers

I hope you are not shocked when I tell you that ~50% of your signups will churn in their first session and usually after trying out 1–2 features for 3–4 minutes. They usually do not bother to reach out to customer support.

Unless your power users are part of your growth flywheel, your growth will ultimately come from new users. But guess who bloats your roadmaps with feature requests? Power users. Listen less to them and focus more on improving the new user experience.

Next step: Watch 15 people using your product for their first time (e.g. through UX tests or Hotjar). Bonus step: Segment by their jtbd (see pattern #1).

4. Using revenue as north star, instead of unique business-drivers

Revenue and # of signups are bad predictors for the future, they are lagging metrics. Instead figure out what uniquely drives your business, your leading metrics.

E.g. Uber might look at # of rides as overall north star (a lagging metric), but to drive the north star, they will look at stuff such as average pick-up time for riders and how much money drivers make per week:

• If supply-density is low (= high pick-up time), then riders will churn.
• If rider-density is low (= low car-utilization = low driver-revenue/week), then drivers will churn.

Next step: Think about what drives your business, what drives user value and what would measure good user experience? If you come up with metrics that could work for IBM and Facebook as well, then you need to dig deeper.

5. Adding more users instead of retaining existing ones

Adding more water to a bucket with holes will result in an expensive water bill 💸. This is obvious, but still, every startup I have consulted did exactly this.

Let’s look at a non-trivial example to understand the reason: Berlin-based Wundertax helps you to file your taxes online. Back in 2017 they focused on lagging metrics (see pattern #4) such as revenue, # of tax-returns and CACs and were quite blind on retention due to the infrequent nature of taxes — it takes a year to know if a customer retains. By increasing marketing spend, they were able to increase revenue and the number of tax returns. Investors were happy, CEO was happy. Until bad user reviews started to accumulate and internal worries about retention increased 🪣💧.

After all, revenue is an extremely lagging metric for retention. The solution was to introduce a leading metric called tax-gap (the diff. between the tax refund promised and the tax refund received). It gave them a way to predict 1y-retention with 90% accuracy and also a tool to improve retention systematically (see pattern #6).

Next step: Check out your users that signed up 6 months ago. How many were active within the last 30 days? Not enough? Fix this before anything else. Bonus step: Send an automatic email to every user that stopped coming back for some time and ask them why.

To avoid a Clubhouse fate you need to work backwards: Fix retention first, then activation and acquisition last.

6. Working on lots of stuff instead of laser focus

You are a small team, you have to create a laser-focus on what’s truly important now. Force-rank Growth, Engagement and Monetization and allocate 70% of the company resources to the first.

Stop starting, start finishing
Founders are impatient, I get it. Instead of starting initiative after initiative, become obsessed with finishing them. Track the time of initiatives from end-to-end: from commitment to seeing the impact. So instead of making sure that everyone is busy, make sure high-impact initiatives get done in rocket-speed 🚀.

Next step: Get everyone that counts into a room and don’t leave before you all agree on the same ranking for Growth, Engagement and Monetization.

7. Caring about features instead of impact

If you regularly ask about the effort to build feature A or B, then this chapter is for you. Fixing this will also magically improve employee retention, because product people hate to work in “feature factories” ;P

Who hasn’t been there: You have a MVP, you add features regularly, but your numbers don’t really improve. You need to change the order — Start with the numbers and work from there:

1. What metric do you wanna move? Hopefully its a business driver (see pattern #4).
2. What customer problems could improve the metric? Talk to customers or watch them using your product to validate them (see pattern #2, #9).
3. Pick the biggest problem (= huge pain, for lots of your users, regularly).
4. Now think of the leanest & smallest way to fix it. Now and only now you should think of the effort.

Next step: Start simple: How many % of users used your latest feature? Bonus step: Add a column “Measuring impact” to your product board (in Jira/Linear). It didn’t move? See pattern #10.

Note: Attaching a metric to an existing idea rarely works. Don’t fool yourself.

8. Handing over a final design instead of involving engineers from the beginning

The main preconception about software engineers is that they only care about technology and don’t really care about the business and the users. Even though some engineers fit this stereotype, most engineers want and should contribute early on.

It is your job to identify a problem worth solving and your designers & engineers job to solve it. They will come up with solutions that are much faster to implement.

Next step: Include engineers in user tests & customer interviews and witness the magic.

9. Iterating in code rather than in Figma

When building new stuff, a lot can go wrong: You might end up “solving” a non-existent problem, fail to fully solve the problem, or even create new problems in the process. Product development is hard. 🤪

Always de-risk features by:

1. Validate if the problem exists, with interviews, data & user tests
2. Validate if the feature solves it, with mockups & prototypes in Figma
3. Measure impact after the release, so you establish a feedback loop

I already hear the impatient founder saying “I have no time for this bloated process”, so many skip some or all 3 steps to save time 😓. It’s like skipping the route-planning in order to arrive faster. You will not arrive faster, you will just drive around more. 🌀

Remember: Shipping features no one wants is the biggest time-waste. In the end it’s much faster to iterate in Figma than in code.

For your next feature: Before writing any code, talk to 7 customers and ask them how they tried to solve the problem so far. Did they even try?

10. Prioritizing “quick wins” over big bets

Truth is, unless you are a very early-stage company, all the obvious quick-wins have already been exploited. So you end up with low impact/low effort vs high impact/high effort topics. Impatient founders tend to chose the first, also because shipping lots of small things will feel good & look like progress. The numbers however will usually stay flat.

The key is to take on big bets and think of ways to break them down into low effort experiments.

For example: Should you build a mobile app to reduce cost-per-signups? Maybe you can just wrap your web-app in a mobile container and drive 1000 visitors on it?

Next step: Take on a big topic you have been procrastinating about and really understand what it is all about. Almost anything can be tested with a small experiment.