Growth Experimentation

Operationsby Helm

Run a steady cadence of tests that compound.

Overview

The Growth Experimentation skill teaches an agent how to run a disciplined testing program: writing falsifiable hypotheses, prioritizing the backlog, designing clean tests, and reading results honestly. It turns scattered ideas into a repeatable loop.

What's inside

The frameworks it teaches.

The experiment loop

Observe, hypothesize, prioritize, run, and decide, then repeat.

Hypothesis writing

A format that forces a falsifiable claim and a kill condition.

Prioritization

ICE and RICE scoring to rank the backlog by expected return.

Reading results

Significance, guardrail metrics, and avoiding false positives.

Example tasks

Things you can hand it.

  • Turn a rough idea into a testable hypothesis.
  • Score and rank a backlog of growth ideas.
  • Design a single-variable test for a pricing page.
  • Decide whether a flat test result means ship or kill.
How it works

Add it to an agent in seconds.

  1. 1

    Enable it on an agent

    Open any agent, pick this skill from the library, and it loads the frameworks into that agent.

  2. 2

    Tune it for the agent

    Override the prompt context for one agent without changing the library copy. One skill, many variants.

  3. 3

    Stack it with others

    Combine skills on a single agent so it carries every framework the job needs at once.

Put Growth Experimentation to work.

Skills are included on every Helm plan. Enable this one on any agent and it starts applying the frameworks right away.

Back to skills