ICE Growth-Testing Framework
Turn marketing from guessing into a repeatable testing engine.
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 3
- Confidence
- 78%
Fechter frames growth marketing as the disciplined application of a testing framework borrowed from consulting: ICE (investment, cost, expected result). Each experiment carries a cost, an expected result, a timeframe, and an owner, plus an explicit hypothesis about the outcome. You then measure the actual result against the hypothesis. Because modern channels make everything measurable, marketers can finally close the loop that was previously open. This converts marketing from art into a repeatable engineering discipline.
Origin
Fechter observed that consultants used a structured ICE-style framework while marketers were traditionally left free to do whatever without measuring results; as channels became measurable, he applied the same discipline to growth work across roughly 50 companies.
Core principles
- 01Marketing is not guessing — there is a framework that says test A before B.
- 02Every experiment has an investment cost, an expected result, a timeframe, and an owner.
- 03State a hypothesis (expected outcome) before you run, then compare it to actual results.
- 04Because everything is measurable now, the result half of the equation can finally be tracked.
How to run it
- 1
Frame the experiment with ICE
Lay out the investment cost, the expected result, and the timeframe for each marketing test before running it.
Pro tip Borrow the discipline consultants already use — this is the same structure they apply to investment decisions.
- 2
Assign an owner and a hypothesis
Name who owns the experiment and write down what you expect the outcome to be, so the test is accountable and falsifiable.
Watch out Without a written hypothesis you can't tell whether the result confirmed or refuted your bet.
- 3
Measure actual results and categorize
Record the real outcome, compare it to the hypothesis, and classify the result — for example whether it compounds over time or is one-off.
Pro tip Prefer tests that build compounding assets even if their immediate result is smaller.
In the wild
Fechter worked with around 50 Series A and Series B companies and repeatedly found CMOs who hired agencies but couldn't measure or reproduce results themselves, illustrating what happens without a testing framework.
→ He concluded teams must understand growth frameworks and be able to execute a test before outsourcing it.
Common mistakes
Running on hunches
Treating marketing as unmeasurable freedom leads to unaccountable spend; without a hypothesis and measured result you never learn which tests actually worked.
Outsourcing before you can execute
Hiring an agency to run a channel you can't run yourself means you can't judge whether their results are any good.
Is it for you?
Best for
Startup founders and growth marketers who need predictable, repeatable, scalable growth rather than intuition-driven campaigns.
Not ideal for
Older-style CMOs who prefer high-level strategy and hunches over hands-on measurable testing.
From the transcript
“marketing is not guessing there's actually a framework here that says you should test a first before you test B”
“growth marki is just implementing a framework for testing so testing different marketing strategies”
From the episode
Josh Fechter: Growth Hacker Marketing Secrets
Josh Fechter