Why Metrics Will Save Your Innovation Projects

Innovation managers increasingly look at data to steer their projects. However, too many metrics can be confusing and even a distraction. Focus on just a small number of metrics, but choose them wisely.

Kennen Sie Ihr wahres Unternehmenspotenzial?
Zu tief im Tagesgeschäft begraben um Innovation umzusetzen?
Sie möchten mehr Experimentieren und weniger Lamentieren?
Sie benötigen strategische Grundlagen um die Zukunft Ihres Unternehmens nachhaltig gestalten zu können?

Lassen Sie uns gemeinsam eine Reise antreten

You may have heard the statement “data is the new oil”. As data has become a more valuable resource, the companies which exploit it, have grown to become the world’s most valuable businesses. Inspired by the meteoric rise of data-driven companies, many innovation managers now measure a long list of metrics. They want to show traction in the market, something that investors and project sponsors are eager to see. However, with so much data available, it can be difficult to focus on the right metrics. I know this from firsthand experience, as I have spent months analyzing the wrong KPIs. This article will help you avoid the most common pitfalls.


Consider the fictional story of Vincent. He is the owner of a successful pizza restaurant in Zurich, Switzerland. Recently, he launched “Pizza Click”, a minimalistic food delivery app, which allows users to order pizza with just one click. Vincent hopes to double pizza sales within a year. The restaurant sells on average 500 pizzas per month. So, if Pizza Click could generate another 500 pizza orders per month, that would be a great success. After 12 months, things look good:

Pizza orders kept coming in during the first year and by December, Vincent delivered almost 600 pizzas per month to Pizza Click users. In addition, Monthly Active Users have also steadily increased. Vincent is very happy!


Inspired by growth hackers, Vincent also created a conversion funnel and carefully tracked every step of the customer journey: downloading the app; completing the account setup; placing the first pizza order; and ordering for a second time.

The metrics look impressive at first sight. The app was downloaded by more than 7500 people; they have almost 5000 accounts with credit cards ready to place orders; and more than 2500 people have ordered pizza at least once using the Pizza Click app.

Only the number of returning customers is a bit low. But even this number is slowly increasing.

Vincent starts wondering… All the metrics are going up, why do returning users remain so low? Did he miss something? Vincent decides to run a “cohort analysis”. This means instead of counting the total number of downloads, active users etc., he looks at the new users in each month as a cohort. For example, in December 1502 people downloaded the app. What happened with them? After some numbers crunching, Vincent sees the results.

Out of the 1502 people who downloaded Pizza Click in December, 60% completed the account set up (down 10 percent points since January). 31% then continued and ordered their first pizza (down 18 percent points since January). And almost none of the 1502 people ordered pizza for a second time. This is a disaster!! It is already late at night and Vincent should get some rest, but he wants to get to the bottom of this now.

Finally, Vincent looks at the conversion funnel in a different way: namely, as transitions from one stage to the next. He quickly sees where he loses most users: only 2 percent of the users who ordered their first pizza in December, order pizza again. What went wrong here?

A quick analysis reveals:

1.     The marketing campaign went really well. Pizza Click downloads increased every month.

2.     The onboarding process is easy enough that 60% of people create an account with a credit card.

3.     After that, the UI of the app is so simple to use that more than 50% of the users place a first order for pizza. Pretty good!

4.     However, something must be wrong with the pizza itself! 98% of all people who received their first pizza in December, never order again with Pizza Click.

Alarmed by the numbers, Vincent picks up the phone early the next day and calls some of his recent customers to ask them what went wrong. He quickly finds the problem: when the pizzas arrive, they are cold. While friends and family who joined back in January were forgiving, most other customers were not. You only get one shot to make a first impression. So many of the customers from this year are probably lost forever.

What is even more upsetting is the fact, that Vincent could have identified the problem as early as February. If he had only looked at the right metrics. The conversion from Active Users to Returning Users was already low in January, but the number steeply dropped in February. At this point at the latest, Vincent should have hit the breaks. Instead, his marketing efforts kept pouring in new users, thereby eroding the customer base.

 In innovation lingo, this phenomenon is called “premature scaling” and it is one of the deadly sins, if you are managing an innovation project. Of course,everybody wants to see their product in the hands of hundreds of people as soon as possible. However, following a systematic innovation process saves you a lot of trouble in the long run. Only start to scale once you have a small, but loyal customer base. Don’t fall into the same trap as Vincent, look at your data in cohorts!

Check out our upcoming blog post about how to run a metrics sprint and define the right metrics for your innovation project or startup.

Check out other posts