Go-to-market
Growth Marketing for Startups: Your First Five Experiments
Growth marketing is not a bag of tricks. It is running small, measurable experiments on the part of your funnel that is weakest, and keeping only what works. Here is how to set up your first five.
Growth marketing for startups is the practice of running small, measurable experiments across the funnel, from first visit to paying customer to referral, and keeping only what moves a chosen metric. Start after you have a product some users love, pick the weakest step in your funnel, write each idea as a testable hypothesis, prioritise with a simple score, and run five experiments.
Growth marketing works best once there is something worth growing. If you are not sure users value the product yet, the product-market fit guide comes first.
Key takeaways
- Start after early users love the product, not before.
- Find the weakest step in your funnel and experiment there first.
- Write every idea as a hypothesis with a metric and a threshold.
- Prioritise with ICE: impact, confidence and ease, each scored 1 to 10.
- Keep an experiment log, including the ones that failed.
When should a startup start growth marketing?
Once a group of users gets clear value and would miss the product. Growth experiments on a product people do not value mostly measure how fast you can lose new users.
A widely used test comes from Sean Ellis: ask active users how they would feel if they could no longer use the product. Superhuman used it and found that companies where more than 40% answered "very disappointed" tended to grow well (First Round Review). If you are well below that, work on the product for the users who love it before scaling acquisition.
Where should growth experiments focus?
On the weakest step of your funnel. Measure the conversion rate between each step, and start where the biggest drop happens relative to what you would expect.
| Step | Example metric |
|---|---|
| Acquisition | Qualified visitors or sign-ups per week |
| Activation | Share of sign-ups who reach the first real result |
| Retention | Share of activated users still active after four weeks |
| Revenue | Share of active users who pay |
| Referral | Share of customers who bring in another |
A common finding: acquisition is not the bottleneck. If most sign-ups never reach a first result, more sign-ups will not help. Fix activation first.
How do you write a growth experiment?
As a hypothesis with a change, a metric, an expected result and a deadline. If you cannot say what result would count as failure, it is not an experiment yet.
A useful template:
If we [change], then [metric] will move from [current] to [target] within [time], because [reason from evidence].
For example: "If we replace the empty dashboard with a sample project, then the share of sign-ups who create their first report will rise from 30% to 40% within two weeks, because interviews showed new users did not know where to start." The numbers are yours to set; the point is setting them before you begin.
How do you prioritise growth experiments?
Score each idea for impact, confidence and ease, from 1 to 10, and run the highest scores first. This is the ICE method, created by Sean Ellis.
Growth Method describes ICE as "three factors used to score and rank experiment ideas on a scale of 1 to 10": impact ("How much will this move your target metric?"), confidence ("How sure are you it will work?") and ease ("How quickly can you ship it?") (Growth Method). Teams combine the three scores in different ways, most often by averaging them.
The score is rough on purpose. Its job is to stop the team arguing about which idea to try first, not to predict results.
What should your first five growth experiments be?
One per funnel step, aimed at the weakest steps first. Here are five common starting points; replace them with ideas from your own users.
| Step | Experiment | Metric |
|---|---|---|
| Activation | A guided first task or sample data for new users | Share reaching a first result |
| Activation | A personal email or call to every new sign-up in week one | Share reaching a first result |
| Retention | A weekly summary email showing the value the user got | Share active after four weeks |
| Revenue | Test a different plan structure or trial length | Share of active users who pay |
| Referral | Ask users at their moment of success to invite a colleague | Invites sent and accepted |
Run one experiment per step at a time, so you can tell what caused a change.
How do you learn from growth experiments?
Keep a log of every experiment: the hypothesis, the result, and what you decided. Failed experiments are as valuable as successful ones, because they rule things out.
A one-line log entry per experiment is enough: date, hypothesis, ICE score, result, decision (keep, drop, iterate). Review the log monthly. Patterns across experiments, such as "anything that adds a human touch in week one works", often matter more than any single result.
For choosing which acquisition channel to focus on in the first place, see startup marketing strategy.
Frequently asked questions
What is growth marketing?
Growth marketing is running small, measurable experiments across the whole customer journey, from first visit to referral, and keeping only the changes that improve a chosen metric.
What is the ICE score?
A prioritisation method created by Sean Ellis. Each experiment idea is scored from 1 to 10 for impact, confidence and ease, and the scores are combined, usually by averaging, to rank ideas.
How long should a growth experiment run?
Long enough to see a meaningful change in your metric at your current volume. Set the duration before starting, and avoid stopping early just because the first days look good.
Is growth marketing different from growth hacking?
The terms overlap. Growth hacking often refers to quick, creative tactics; growth marketing usually means a steadier process of experiments across the whole funnel.
What should a startup's first growth experiment be?
One aimed at the weakest step of your funnel. For many early products that is activation: helping new users reach their first real result.
Start where you are
Experiments work best on a funnel you can see and a buyer you can name.
Sources
- Growth Method, ICE Framework
- First Round Review, How Superhuman Built an Engine to Find Product/Market Fit (2018)
The example experiments and targets are illustrations. Choose metrics and thresholds that fit your product and volume.