Library Pricing Join the waitlist
product

CRO & experimentation loop

  • cro
  • experimentation
  • ab-testing
  • conversion

Run conversion as a repeating loop, gather behavioural insight, prioritise hypotheses by evidence, build and run the A/B test, analyse honestly, ship the winner, and feed it back into the next round.

Get access See how it runs
5 Steps in the flow
6 Tools orchestrated
2 Human checkpoints
Weekly Run cadence
2–4 shipped winners / month Validated conversion wins

The objective

What this system is built to do

Most tests are flat, that's expected. The win comes from running the loop fast and honestly: watch real users, prioritise by evidence, ship the winners, and let data overrule opinion every time.

Throughput and honesty beat clever ideas. Gather behavioural insight with Hotjar + GA4, prioritise hypotheses by evidence (ICE/PXL), run the A/B in VWO, call the flat ones flat, and ship only validated winners. Not doing: HiPPO-driven testing, calling winners before significance, or A/B-ing pages without the traffic to settle.

How it runs

The steps your team follows

The workflow, exactly as it runs inside . It plays through step by step, or pick any node on the map.

1Gather insight
2Prioritise hypotheses
3Build & run the test
4Analyse
5Ship winner & bank insight
Start
1

Gather insight

Weekly
  1. Pull funnel drop-off from ; find the step bleeding the most users.
  2. Watch session replays in / on that step, log rage-clicks, dead-clicks, stalls.
  3. Run a one-question on-page survey to colour *why* behind the *what*.
  4. Human approval, confirm the priority funnel step before generating hypotheses.
2

Prioritise hypotheses

Weekly
  1. Turn each friction point into one testable hypothesis: change → expected effect → why.
  2. Score every idea with ICE or PXL; back the Confidence column with replay/analytics evidence.
  3. Cut low-evidence ideas; queue only the top-scoring third.
  4. Human approval, approve the test to build (and whether traffic justifies an A/B vs a direct fix).
3

Build & run the test

Weekly
  1. Build the variant in (or ); for B2B, personalise the variant with .
  2. Pre-commit the primary metric, minimum sample, and window, no peeking-and-stopping.
  3. QA the variant across devices, then launch the 50/50 split.
  4. Let it run to the pre-committed sample; don't call it early.
4

Analyse

Weekly
  1. Read the result at significance: win, loss, or flat, most will be flat.
  2. Write down *why* you expected it to move; a flat result with a clear hypothesis is a learning, not a waste.
  3. Segment the result (device, source, new vs returning) for a hidden win inside a flat average.
  4. escalate borderline or surprising results for a human read before you ship.
5

Ship winner & bank insight

Weekly
  1. Roll the winning variant to 100%; archive the losers.
  2. Log the insight in the hypothesis bank, what won, what was flat, and why.
  3. Feed the learning back into next round's prioritisation so the loop sharpens.
  4. Confirm the live lift holds in after rollout.
1 / 5

Setup

The stack

The tools this system drives, and the context you bring once. Connect your own accounts, or swap in your team's equivalent.

Tools to connect

GA4Free Powers Gather insight Or Plausible, Fathom
HotjarFree – $32 / mo Powers Gather insight Or Clarity, FullStory
ClarityFree Powers Gather insight Or Hotjar, FullStory
VWOFree – custom Powers Build & run the test Or Optimizely, AB Tasty
OptimizelyCustom Powers Build & run the test Or VWO, AB Tasty
MutinyCustom Powers Build & run the test Or Intellimize, Optimizely

Context to bring

Your documents 4 files
  • Conversion goals & funnel The steps that matter and where you lose people.
  • Hypothesis backlog The ideas worth testing, prioritised.
  • Brand & design system The components and rules variants have to respect.
  • Analytics access GA4 / your stack plus the testing tool.

Proven in market

Run by 75 teams, verified on real accounts

CRO & experimentation loop is built and verified by Growth Division, the GTM team behind 150+ companies. Here is what the same play has produced for their clients.

600% lower CPA, 2 channels validated

Housecure · PropTech

Higher conversion across the journey

SeedLegals · Legal Tech

+118% demo bookings

Position Green · Sustainability SaaS

Built & verified by Growth Division

What we've learned running it hundreds of times

The small, hard-won details that decide whether it works, from every run.

~1 in 5 tests beat control
600% lower CPA (Housecure)
+31% checkout conversion (best test)
checkout CVR high confidence
Before 5.8%
After 7.6%

A replay-led checkout redesign lifted conversion 31%

The biggest wins are usually friction you can already see in replay, not clever copy. Watch first, then test the obvious fix, and let it run to significance before you believe it.

Checkout A/Bn≈42,000 sessions5 weeks
win-rate medium confidence
Before 11%
After 21%

Scoring ideas with ICE roughly doubled our test win-rate

Prioritisation doesn't make ideas better, it stops you spending scarce test slots on low-evidence ones. Most tests are still flat; you just waste fewer of them.

ICE scoringn≈48 tests2 quarters
step completion medium confidence
Before 41%
After 55%

A replay-spotted friction fix lifted a funnel step, but it took weeks to read

Low-traffic steps don't suit a clean A/B, make the obvious replay-driven fix and read it on the funnel instead. Expect a lag: the verdict takes weeks of cohorts before the line settles.

Funnel fix (no A/B)n≈1,400 / week6 weeks
The full lab notebook Every experiment charted, with the raw before/after and the decision rules behind each call. Unlock free

Get started

How to run this system

Live in an afternoon. Clone it, connect your tools, set your inputs, and turn it on.

  1. 1

    Add to

    Clone the CRO & experimentation loop into your workspace, wired end to end.

  2. 2

    Connect your tools

    GA4, Hotjar, Clarity +3 more.

  3. 3

    Set your inputs

    Your ICP, targets and brand rules, once.

  4. 4

    Turn it on

    It runs weekly on autopilot, checking in with you only where it matters.

~30 min / week your time on autopilot · vs 8–12 hrs / week by hand
2–4 shipped winners / month Validated conversion wins, produced for you

Early access

Run the CRO & experimentation loop in your workspace.

is in stealth with a first cohort of scaleups. Join the waitlist and we’ll open this system, and the full library, to you first.

All systems in the library