Template

CRO Roadmap Template (12 weeks)

The framework we use to plan a CRO program for $1M+ Shopify and DTC brands. Copy any section into Notion, Linear, or Google Docs.

Sprint structure

A 12-week CRO program has four phases. Skip phase 1 only if you already have rigorous research + a working measurement stack — most brands at $1M+ ARR don't, and skipping it produces tests that test the wrong things.

Weeks 1–2

Foundation — research & instrumentation

  • Audit GA4, Shopify analytics, Klaviyo, and ad platform data — confirm event coverage (add_to_cart, begin_checkout, purchase, refund).
  • Set up a single source-of-truth dashboard. Metrics to track: sessions, CVR, AOV, RPV, CAC, MER, LTV by channel.
  • Conduct heuristic walkthrough on top 5 templates (PDP, PLP, cart, checkout, homepage) using 48-heuristic framework.
  • Run 3 user interviews with recent buyers (5 questions: discovery, hesitation, comparison set, final trigger, regret).
  • Install Microsoft Clarity (free) for heatmaps + session recordings if not already deployed.
Weeks 3–4

Hypothesis backlog + prioritization

  • Synthesize research into 40–60 hypotheses. Format: "If we [change], we expect [metric] to [direction] because [behavioral reason]."
  • Score every hypothesis using ICE (Impact 1–10, Confidence 1–10, Ease 1–10) or PIE.
  • Cluster hypotheses by page template — surface the top 12 to schedule for tests in Weeks 5–12.
  • Define guardrail metrics (refund rate, support tickets, page speed) so a "winning" test that breaks them gets caught.
Weeks 5–8

Sprint 1 — 4 tests

  • Week 5: Launch tests 1 + 2 on highest-traffic templates (typically PDP + cart drawer).
  • Week 6: Launch tests 3 + 4 after sample-size calculation confirms power.
  • Run minimum 2 full business cycles per test. No peeking — set a fixed end date based on sample size.
  • Document each test with: hypothesis, screenshots, segmentation breakdown, decision, follow-up.
Weeks 9–12

Sprint 2 — 4 tests + program review

  • Week 9: Launch tests 5 + 6 informed by Sprint 1 learnings (winners get a follow-up, losers spawn iterations).
  • Week 10: Launch tests 7 + 8.
  • Week 11: Mid-sprint sync — review guardrails, retention metrics, and 30-day cohorts of Sprint 1 winners.
  • Week 12: Quarterly retro. Compute win rate, average lift, revenue attributed, learnings doc.

Weekly cadence

Discipline beats velocity. Run this rhythm consistently for a quarter and you'll outperform teams running 3x more tests with no cadence.

  • Monday — Sprint planning — confirm tests live this week, sample size, end date.
  • Wednesday — Mid-week check — guardrails OK, traffic split balanced, no data quality issues.
  • Friday — Reading — review running tests against sample size; archive completed.
  • End of month — Program review — win rate, learnings, hypothesis backlog refresh.

Hypothesis template (copy this)

Every test should be documented with this exact structure before it goes live. If you can't fill in all fields, the hypothesis isn't ready.

Hypothesis #X — [test name]

Page / element: [PDP / cart / checkout / homepage]
Traffic source: [paid / organic / direct / all]
Device: [desktop / mobile / both]

Observation:
[what data, research, or session recordings revealed]

Behavioral reason:
[which heuristic — anchoring, social proof, framing, default effect, loss aversion, etc.]

If we change:
[exact change to ship as the variant]

We expect:
[metric] to [direction] by [estimated lift %].

Because:
[the behavioral mechanism connecting cause and effect]

Sample size needed: [X visitors per variant — calculated via /tools/sample-size-calculator]
Minimum duration: [Y business days]
Primary metric: [CVR / AOV / RPV]
Guardrail metrics: [refund rate, page speed, support tickets]
Segmentation cuts: [new vs returning, traffic source, device]

Tools we use for this

Want us to run this for you?

The 90-Day CRO Sprint is this template, run end-to-end by our team — research, hypothesis backlog, 6–8 A/B tests live, +15% CVR guaranteed.

See the 90-Day Sprint →