Where you are
You can now read the journey map (a1), judge the doors (a2), tell the search road from the browse road (a3), see what the product page must do (a4), and name what kills checkouts (a5). This session is the payoff: turning all of it into a funding decision. The analyst track's flow simulator (b4) prices every fix; your job is the ranking.
Reach x severity x confidence 10 min live
Every journey fix claims to matter. Three lenses separate the ones that do:
LiveThe lenses, one by one6 min▶
- Reach - how many sessions touch the gate. The first gate (PDP penetration) touches all 360,000 sessions; the last (completion) touches 16,200. A small lift on a huge base can beat a big lift on a tiny one - the b4 simulator makes this concrete.
- Severity - distance to benchmark. Mango Lane's completion (60%) already beats the ~45% typical Shopify store (Littledata, 2025) - little headroom. A gate lagging its benchmark has room to run. Severity is where a5's honest benchmark tables earn their keep.
- Confidence - the grade of the evidence. Your own measured data beats fetched-primary research (Baymard, Spiegel), which beats vendor benchmarks (Airship), which beats a hunch. A fix backed by a hunch is not dead - it just needs a cheap test before a big build.
Self-studyThe two vetoes2 min read▶
Two proposals get sent back automatically, whatever their score:
- No gate delta. A fix that cannot be stated as "gate X from A% to B% by mechanism M" cannot be priced by the simulator or tested by an experiment. "Redesign the homepage" goes back until it becomes a delta.
- Vendor stat, no test plan. A fix whose only evidence is a vendor benchmark ships with an A/B plan or not at all. Vendor numbers locate opportunities; they do not confirm them.
Every fix gets a price tag 8 min live
The analyst track's flow simulator (b4) turns any gate delta into monthly GMV on current traffic. A fully priced roadmap looks like this - four fixes, compounding to +47%:
| Fix | Gate delta | Orders | GMV | vs baseline |
|---|---|---|---|---|
| baseline | - | 9,720 | $660,960 | - |
| Smarter search & nav | PDP penetration 40% → 43% | 10,449 | $710,532 | +7.5% |
| + Richer PDP | add-to-cart 25% → 28% | 11,703 | $795,796 | +20.4% |
| + Persistent cart | checkout start 45% → 50% | 13,003 | $884,218 | +33.8% |
| + One-page checkout | completion 60% → 66% | 14,304 | $972,639 | +47.2% |
LiveReading a priced roadmap as a leader4 min▶
Three habits when the table lands on your desk:
- Ask for single-fix prices, not just the stacked ladder. The ladder shows the dream; the per-fix delta shows what each team's ask is worth alone.
- Check the deltas against a5's benchmarks. A proposal moving an already-strong gate (60% → 66% completion) needs stronger evidence than one lifting a lagging gate.
- Remember the honesty rail. The arithmetic is real; the effect sizes are estimates until tested. Fund the fix AND its experiment.
Why gate-delta roadmaps win funding meetings. "Checkout redesign, Q3, 2 engineers" invites debate. "Completion 60% to 66% = +$88K GMV per month on current traffic, confirmed by a 4-week A/B" invites a yes. Same work - the second sentence carries its own business case.
What goes on the weekly review 6 min live
The metrics course built the weekly business review (its a4); this course adds the journey rows: the four gates WoW, the door mix (share per entry), search share and the zero-result rate, and the per-step checkout completion. Four rows, five minutes, and a journey leak never survives two weeks unseen.
LiveThe experiment handoff4 min▶
The ranking ritual ends where experimentation begins: every funded fix ships with an A/B plan - the hypothesis is the gate delta, the success metric is the gate itself, the guardrails come from the metrics course (margin, return rate). The craft of running that test honestly lives in learn-experimentation-with-phoebe. Until the test reads out, the simulator's numbers are the estimate, not the result.
Self-studyThe ritual's fine print3 min read▶
- Reach and severity are computable; confidence is a judgment. Write the evidence grade next to each score so the judgment is at least visible.
- Re-rank quarterly, not weekly. Priorities that flap with every WBR exhaust teams; the WBR watches, the quarter decides.
- The ritual is craft, not law. No standards body certifies reach x severity x confidence - it is a distilled practice. Its value is that everyone argues inside the same three questions.
Rank three competing fixes ★ 12 min · everyone judges
Three teams, one budget. Walk the ritual out loud:
Fix A - one-page checkout. Reach: 16,200 checkout sessions. Severity: LOW - 60% completion already beats the ~45% benchmark. Confidence: HIGH - Baymard's field research is fetched-primary. Verdict: solid but not urgent.
Fix B - zero-result search rescue. Reach: HIGH - all 36,000 search landings plus every in-app searcher on the highest-converting path (12% vs 4.5%). Severity: HIGH - zero results is a dead end on the best road. Confidence: HIGH - your own search logs show the terms. Verdict: fund first.
Fix C - homepage redesign. Reach: 180,000 sessions - huge. Severity: unknown. Confidence: LOW - no gate delta stated, no benchmark named. Verdict: veto one applies - send it back to be restated as "Home reach-PDP from 20% to X% by mechanism M".
Final ranking: B, then A, then C-after-restatement. Say the reasoning in the three-lens vocabulary and the meeting ends without a fight.
Try it yourself - this week ◐ 15-20 min total
- Take your team's current journey roadmap. Score every item on reach x severity x confidence; write the evidence grade next to each.
- Send back every item with no gate delta - kindly, with the template: gate, from%, to%, mechanism.
- Add the four journey rows to your weekly review: gates WoW, door mix, search share + zero-result rate, per-step checkout.
- Ask "what is the test plan?" at your next funding decision. Watch what happens.
Sources this session draws on
The ranking ritual is distilled craft; the numbers inside it come from measured canon and flagged research. This page draws on:
Three questions before you go 🎯 ◐ 90 seconds
1 · A fix on the first gate touches 360,000 sessions; one on the last touches 16,200. What lens is this?
Reach counts the sessions that touch the gate. It is why gate-1 fixes so often out-value glamorous last-mile work - the base is 22x larger.
2 · "Redesign the homepage" arrives on the roadmap. What happens under the ritual?
A fix without a gate delta cannot be priced or tested. The veto is not a rejection of the idea - it is a demand that the idea become measurable.
3 · A funded fix's only evidence is a vendor benchmark. What must ship with it?
The second veto: vendor stat, no test plan, no ship. The simulator and the benchmarks estimate; only the experiment measures your store.