Growth is a system

Context

Once the strategy team existed, a large part of its job was growth: getting more students to find IGotIn, sign up, and stay. While I was there, the app expanded to more than 50 campuses.

The problem

It’s tempting in a startup to look for the one idea that changes everything: the campaign that goes viral, the feature everyone shares. I came in half expecting growth to work that way. What the company needed more was a reliable way to tell which ideas were actually working.

What I did

Pick one small change
Show it to some users, and the old version to others
Measure the difference
Keep it or drop it

Then the next change

Fig. 1 The loop behind every growth experiment.

We ran A/B tests¹ on the landing page and the calls to action. Each change was small and measured against the version before it. Over time, sign-up conversion rose by about 22%.

Moving the spend

Acquisition had leaned on paid Instagram. We shifted toward organic content on TikTok and other social platforms. The cost of acquiring each user fell by about 35%. Our campaigns passed a million impressions, and brand awareness rose by about 35%.

sign-up conversion: +22%
cost to acquire each user: −35%
campaign impressions: 1M+

Fig. 2 The numbers that held up.

Running many small bets

In total I worked on more than ten growth initiatives across the company’s products. Most were small, and the point was to find out quickly which ones to keep.

Connecting growth to the pivot

The surveys from my first weeks had pointed toward letting students earn money through the app. That turned growth and monetization into the same question: get students in, and give them a reason to come back after the first weeks of the semester.

What I learned

Growth isn’t a viral idea. It’s a system of small experiments, measured honestly, with the ones that don’t work dropped quickly.

It was also the first time the failure of my tutoring platform felt useful. I knew what it looked like when a product had no plan for distribution. At Bondit, distribution was the job.

An A/B test shows two versions of something to different users at the same time and measures which one performs better. ↩