A better CRO process starts before the experiment.
Most testing programs jump straight to building. We research first, diagnose the real friction, and turn evidence into a prioritized roadmap before a single experiment goes live.
Seven steps. One loop that never stops running.
Every engagement runs the same seven-step system, from the first Audit to the tenth cycle of Ongoing Experimentation.
Research
Understand what users are doing and why. Qualitative interviews, session recordings, and a review of your current analytics, before we assume anything.
- Stakeholder and customer interviews
- Session recordings and funnel analysis
- Tracking and data-layer sanity check
Diagnose
Find friction and conversion opportunities. The research gets turned into a ranked list of where the real problems, and the real opportunities, actually sit.
- Funnel and page-level friction points identified
- Behavioral data reviewed against the qualitative research
- Opportunities ranked by where the impact actually sits
Hypothesize
Turn evidence into testable ideas. Every hypothesis is framed as if/then, with a clearly predicted outcome, and traces back to a specific finding.
- Framed as if/then, with a predicted outcome
- Tied directly to a research finding, never a hunch
- Written so it can actually be proven wrong
Prioritize
Use the PROOF Score. Every hypothesis is scored on potential impact, reach, evidence, confidence and effort, so the roadmap reflects what actually moves revenue.
- Every idea scored on the same weighted framework
- Roadmap agreed with you before a single test is built
- Biggest, most evidence-backed bets rise to the top
Experiment
Run statistically disciplined experiments where traffic supports them, and evidence-based rollouts where it doesn't. Never faked statistics or a test called early to hit a deadline.
- Proper sample-size and duration planning
- QA across your key browsers, devices and segments
- No peeking, no early calls, no exceptions
Analyze
Understand both outcomes and learnings. Results are read for statistical validity, not just a green metric, and reported honestly, wins, losses and flat results alike. (For stores where the sale doesn't close entirely on-screen, we can extend measurement to the actual outcome, delivery, disbursement, in-store visit, as one example of more advanced measurement; it's not what defines the engagement.)
- Read for statistical validity, not just a green metric
- Wins, losses and flat results, reported the same honest way
- Learnings extracted whether the test wins or not
Scale
Implement winners and feed learnings into the next research cycle. Nothing here is a one-off project. Scale closes the loop straight back into Research.
- Winning variants rolled out fully
- Learnings folded back into the next round of research
- The loop starts again, not a one-time redesign
The PROOF Score: your experiment decision engine.
Score any idea the way we do. Five inputs, one weighted number, so the biggest, most evidence-backed bets rise to the top of the roadmap, not whatever's easiest to ship.
Potential impact, evidence and confidence carry the most weight, so the biggest, best-evidenced bets should always rise to the top.
Effort is scored in reverse — less effort earns more points, not fewer.
This is exactly how we prioritise your roadmap →Why most CRO programs fail.
Not because testing doesn't work. Because the process feeding it is broken in one of six predictable ways.
Random testing
Ideas picked by gut feel or whatever's trending. Nothing compounds, because nothing was ever connected.
Redesign-first thinking
Rebuilding the page before understanding why it's actually underperforming.
Insufficient research
Jumping straight to hypotheses with no research behind them to test.
Weak measurement
Tracking that can't reliably tell a real win from statistical noise.
Low-impact tests
Testing a button color while the real friction sits untested, unmeasured.
Win/loss-only thinking
Every test graded pass or fail, so the losses never get to teach anything.
Every engagement produces the same seven outputs.
Whether it's a single Audit or a running Ongoing Experimentation program, the system guarantees the same things land in your hands.
Research findings
What we learned about your shoppers and where they actually hesitate.
Hypotheses
Specific, evidence-backed ideas for what to test, and why it should work.
Prioritized roadmap
Every hypothesis scored and ranked on the PROOF Score, agreed with you upfront.
Experiment designs
Fully specified tests: variants, sample size, duration and the primary metric.
Test results
Statistically read outcomes, reported honestly: wins, losses and flat results alike.
Learning library
Every result captured in one place, so nothing gets re-tested by accident.
Next experiments
Learnings folded straight into the next round of research. The loop, not a one-off.
Honest questions about the method itself.
Ready to put a real roadmap through the system?
Start with a discovery call, or the CRO Audit if you're ready to go straight to a prioritized, PROOF-scored roadmap.
Get CRO Audit →