Amanoba Logo
Back to Course

Growth experiments

Day 29 of 30 · Generative AI 2026: Build AI Apps and Agents

One-liner: Run small experiments to find early traction.
Time: 20 to 30 min
Deliverable: Growth Experiment Plan

Learning goal

You will be able to: Design a small growth experiment with a clear success metric.

Success criteria (observable)

  • One experiment is defined with a clear hypothesis.
  • A metric and target are written.
  • A start and end date are set.

Output you will produce

  • Deliverable: Growth Experiment Plan
  • Format: One page plan
  • Where saved: Course folder under /generative-ai-2026-build-ai-apps-and-agents/

Who

Primary persona: Digital nomad testing growth channels Secondary persona(s): Users who might try the product Stakeholders (optional): Collaborators

What

What it is

A small, time boxed experiment to test one growth idea. It focuses on learning, not scale.

What it is not

It is not a long term marketing strategy or a big campaign. It is a quick test to find early signals.

2-minute theory

  • Small experiments reduce wasted marketing spend.
  • Clear metrics make learning objective.
  • One experiment at a time keeps focus.

Key terms

  • Hypothesis: A testable statement about expected results.
  • Metric: The number you track to judge success.

Where

Applies in

  • Growth planning
  • Marketing tests

Does not apply in

  • Core product development

Touchpoints

  • Landing page
  • Email signup
  • Ads or outreach

When

Use it when

  • You want early traction signals
  • You have a working product to share

Frequency

One experiment per week or per month

Late signals

  • Many ideas but no tests
  • No measurable growth data

Why it matters

Practical benefits

  • Faster learning
  • Lower marketing risk
  • Clearer channel decisions

Risks of ignoring

  • Random marketing efforts
  • Slow growth learning

Expectations

  • Improves: learning speed
  • Does not guarantee: growth

How

Step-by-step method

  1. Choose one channel to test.
  2. Write a clear hypothesis.
  3. Define a metric and target.
  4. Set start and end dates.

Do and don't

Do

  • Keep the test small and time boxed
  • Track one primary metric

Don't

  • Run multiple tests at once
  • Skip the hypothesis

Common mistakes and fixes

  • Mistake: No metric. Fix: Choose one number to track.
  • Mistake: No time box. Fix: Set start and end dates.

Done when

  • Hypothesis is written.
  • Metric and target are defined.
  • Dates are set.

Guided exercise (10 to 15 min)

Inputs

  • Product value proposition
  • One growth channel idea

Steps

  1. Write a hypothesis.
  2. Define the metric and target.
  3. Set dates and a budget.

Output format

Field Value
Channel
Hypothesis
Metric and target
Dates and budget

Pro tip: Choose a channel you can execute in one day.

Independent exercise (5 to 10 min)

Task

Cut the experiment scope by 50 percent and keep the metric.

Output

Revised experiment plan.

Self-check (yes/no)

  • Is the hypothesis testable?
  • Is the metric clear?
  • Is the time box defined?
  • Is the scope small?

Baseline metric (recommended)

  • Score: 1 experiment defined
  • Date: 2026-02-06
  • Tool used: Notes app

Bibliography (sources used)

  1. Lean Analytics. Alistair Croll and Benjamin Yoskovitz. 2024-01-01. Read: https://leananalyticsbook.com/

  2. Growth Experiment Guide. Reforge. 2024-01-01. Read: https://www.reforge.com/blog/growth-experiments

Read more (optional)

  1. Experiment Design Why: Basics of testing and metrics. Read: https://cxl.com/blog/ab-testing/