Strategy & Growth
Growth Engineering
We build the technical foundation for testing growth ideas, then run experiments with your team across activation, onboarding and retention.
The problem
Growth ideas sit in the engineering backlog for weeks, and when one finally ships, nobody can tell if it worked.
Product roadmaps are planned around features. Removing a signup step, adding a referral mechanism or changing an email trigger keeps losing to bigger items. Meanwhile marketing keeps buying traffic for a product experience it can't touch.
Measurement makes it worse. Event names drift, user identity breaks across devices, and there is no control group. A before-and-after chart can credit a release for what was really seasonality or a campaign running the same week.
Teams often conclude they need better ideas. What they usually lack is a low-risk way to ship a change and a trustworthy way to read the result.
You may recognize this
- A minor onboarding tweak has to wait for a full release cycle.
- The same event returns different numbers in different tools.
- Nobody can say what a new user must do to count as activated.
How we work
Tracking plan and identity
We write an event dictionary with naming rules, stitch anonymous and logged-in users together, and validate critical events server-side.
Find the activation moment
Retained and churned cohorts are compared to isolate early behaviors linked to retention, and the candidate definition is then tested directly.
Flags and experiment plumbing
Using tools like GrowthBook, PostHog or LaunchDarkly, we separate deployment from release, confirm clean assignment with A/A tests, and size each test from a minimum detectable effect.
Build compounding loops
Referral mechanics, shareable outputs, data enrichment and programmatic pages get built in your codebase, so each new user can create the next acquisition opportunity.
What we deliver
- Event dictionary and tracking plan
- Feature flag and A/B testing setup
- ICE-scored experiment backlog
- Shipped onboarding or lifecycle flows
- Experiment log and decision template
How we measure
- Activation rate: tells you whether new users reach the product's core value.
- Valid experiments completed per period: your real learning capacity.
- Net lift against a long-running holdout: confirms that individual wins add up in aggregate.
First use in an invoicing product
Illustrative scenario: An invoicing SaaS for small businesses sees plenty of free-trial signups, but many users leave before creating a first invoice, which may be the activation moment. The team ships a prefilled starter workspace behind a feature flag, keeps part of new signups in a control group, and reads the effect in retention data.
Common questions
How does growth engineering differ from CRO?
CRO researches and tests conversion on existing pages and flows. Growth engineering writes product code: new onboarding, referral systems, automations and the instrumentation behind them.
How do you work with our developers?
In your repository, under your code review and release rules. Experiment code ships with a cleanup plan so flags do not pile up.
What if we don't have much traffic?
A sample size calculation will show that early. At low volume we test only large changes and rely on user interviews and session evidence for the rest.
Before we work together, let's think together.
Requests are answered by the founder.