Recap of the first 90 days of GrowthBar

Recap of the first 90 days of GrowthBar
Photo by Anne Nygård / Unsplash

We've documented GrowthBar extensively both internally and externally. It is our largest acquisition to-date. There is a lot on the line to make this business one of our wins.

You can watch/listen to my full walk-through explanation in this loom video:

Generally, the first 30 days after acquisition is lowkey. We might break production once or twice, but overall, from a growth perspective, we are building a baseline and evaluating any quick wins to implement. For the other PLG companies, we've historically identified 1-2 experiments to run on the funnel (ie. tweak pricing, change the duration of the free trial, or change the limitations of the free version, etc.).

What happened?

I want to just recap the last 60 days of GrowthBar today:

  • June 23-July 23: We left everything as-is
  • July 23-August 23: We made the following changes: removed CC requirement, added back a handful of restricted IP addresses, and extended the free trial period from 5 days to 14 days
  • August 23-September 23: We added BACK the CC requirement, shortened the free trial period from 14 days to 7 days, and defaulted the sign-ups to a monthly subscription plan (instead of annual)

The changes on July 23rd netted the following results:

  • 1639 total new free trial sign-ups (according to our Metabase account)
  • -$12,000 MRR in churn (according to Stripe)
  • -127 active subscribers (according to Stripe)
  • 22 new customers added (according to Stripe)
Data from Stripe

Something was wrong... very very wrong. The data was frightening in comparison to our control time period, and it didn't seem like it was going to trend back up without us making some changes (or reverting some of our changes back).

To be honest, we expected:

a) a decrease the quality of the sign-ups: 51.25% of sign-ups during the July-Aug time period had NO usage. This was significantly up from 39.5% in the Jun-July period and 25.5% in the Aug-Sep period.

Usage Data from MetaBase

b) an increase in the total number of sign-ups: 1639 sign-ups generated in the July-Aug time period. 4x-5x more than average.

MRR/Subscriber Data from Stripe

c) same, if not, a slight increase in new customers/revenue: 50% decrease in new customers and 50% decrease in new MRR added.

New MRR/customer data from ProfitWell
New MRR dropped to $1473 ($1,000 MRR difference from the previous month)

Lessons Learned

  • Quality is NOT the only consequence when prioritizing more quantity. The friction we attempted to remove not only affected the quality of the sign-up, but it obviously did something to the commitment of the user.
  • Even in product-led companies, you HAVE to ask for the sale intently and repeatedly, especially when the activation experience is NOT optimized.
  • The product-led funnel is extremely complex and unique. We often focus a lot on optimizing the sign-up experience and the user experience once they are IN; however, in this case, we couldn't help but wonder about the quality of the traffic we were driving.

So what now?

  • We implemented a data gathering step in our onboarding flow so we can start identifying any potential patterns in who is signing up and who is paying.
  • We are also building a couple features to help us be a bit more prescriptive on who we want as customers (and who we believe is our ideal ICP)

Watch the segment of the Loom video where I talk about this here:


Any feedback/thoughts... please email me.