C01
45.0%
at its norm
Diagnosis 2026-08-30
- Site
- Example Outdoor Co. (fictional)
- Window
- 2026-06-01 to 2026-08-29, United States
- Access
- full
- Impact scoring
- LTV-weighted
- Findings
- 7: 3 Quick Win, 4 Implement + Measure
- Rule check
- Check passed
One-pager
$55k
at stake in the worst cell, this windowRecommendation: show the shipping cost before checkout on mobile. New visitors on mobile finish 46% of the checkouts they start, against 70% on desktop. That gap is about 466 purchases a quarter, $55k (modelled). The mobile cart hides the shipping cost until checkout step 2; desktop shows it in the cart.
Both gates passed: analytics sees 95% of platform orders, steady by month, and overseas bot traffic is filtered out of every rate.
| Finding | Sources | Lane | Owner | |
|---|---|---|---|---|
| 1 | Mobile checkout completion for new visitors is 46% against 70% on desktop (F01) | Analytics, recordings | Implement + Measure | Ecommerce |
| 2 | Shipping cost first appears at checkout step 2 on mobile; 31 of 50 watched exits left within 20 seconds of seeing it (F02) | Walk, recordings, support | Implement + Measure | Ecommerce |
| 3 | Both competitors show the free-shipping threshold on mobile product pages (F03) | Competitor walk | Quick Win | Merchandising |
| 4 | Mobile listing click-through 34% against 40%; filters reset on back (F05) | Analytics, walk | Implement + Measure | Ecommerce |
Quick wins this week: the threshold line on mobile product pages (F03), search synonyms (F06), the bot filter on saved reports (F07).
Not checked: customer calls, site speed, retention. Measure the checkout cell weekly against 46.0% after the change.
Gates
No funnel conclusion until both gates pass, or the block is written down.
Gate 1: Bots and overseas trafficThe market filter removes 22,400 sessions from Vietnam with 0 purchases (F07)
CleanGate 1: Tracking changes in the windowNo analytics annotation and no tag release in the window
CleanGate 2: Metric definedSession rate per transition, United States, sessions that fired each event
CleanGate 2: Orders vs platformAnalytics sees 4,191 purchases against 4,410 orders, 95.0%; 94.9% to 95.2% by month
CleanGate 2: Bots excludedMarket filter on every query
CleanFunnel
Every cell against its norm, ranked by purchases at stake.
Desktop
Mobile
New
Returning
New
Returning
Landing to product view
C02
50.0%
at its norm
C03
42.0%
about 24 at stake
C04
47.0%
about 24 at stake
Product view to add to cart
C05
9.0%
at its norm
C06
14.0%
at its norm
C07
7.5%
about 178 at stake
C08
12.5%
about 131 at stake
Cart to checkout start
C09
50.0%
at its norm
C10
58.0%
at its norm
C11
47.0%
about 57 at stake
C12
55.0%
about 60 at stake
Checkout start to purchase
C13
70.0%
about 14 at stake
C14
80.0%
at its norm
C15worst
46.0%
about 466 at stake
C16
72.0%
about 121 at stake
Listing view to item click
L01
40.0%
at its norm
L02
34.0%
about 108 at stake
Purchases at stake in this window
C15 (Checkout start to purchase, New, Mobile): 892 of 1,939 progressed, 46.0% against a norm of 70.0% (Desktop). 1,047 lost, about 466 purchases at stake ($55k at AOV $118.00).
Runner-up: C07 (Product view to add to cart, New, Mobile), about 178 purchases at stake ($21k at AOV $118.00). Checkout: 65.5% (Desktop 75.0%, Mobile 57.4%). Split trusted: yes.
Two older rules would have missed this. Ranked by sessions lost, C07 wins (50,875 lost). A fixed 65% checkout pass mark clears checkout, because the overall rate is 65.5%. Built by derive/funnel.py (data/derived/funnel/how.md).
Norm: the same visitors on the other device, or last periodBelow its normAt or above its norm
Desktop against mobile, new visitors
DesktopMobile
Landing to product view
+3.0 points
Product view to add to cart
+1.5 points
Cart to checkout start
+3.0 points
Checkout start to purchase
70.0%46.0%
+24.0 points0%25%50%75%100%
Show the numbers
| Step | Desktop | Mobile | Points |
|---|---|---|---|
| Landing to product view | 45.0% | 42.0% | +3.0 |
| Product view to add to cart | 9.0% | 7.5% | +1.5 |
| Cart to checkout start | 50.0% | 47.0% | +3.0 |
| Checkout start to purchase | 70.0% | 46.0% | +24.0 |
Checkout steps, all devices
Is it the page?
Before blaming the page, rule out what else moves this cell: the traffic it gets, stock, and dated changes to prices and promotions.
Traffic mix: 12% of the gapMix explains 3.0 of the 24.0 points. The rest is the same channels converting less here.
Ruled outStockApplies at product view to add to cart. Stock errors in checkout show in the checkout steps and support tickets.
Not applicablePrice and promo changes: 0% of the gap3 dated changes in the window; none moved this cell beyond the 6.7-point swing more than desktop.
Ruled outLeft for the page and what's on it: about 88% of the gap. Built by derive/rule_out.py (data/derived/rule-out/how.md).
Sources at the worst cell
F01Analytics
New visitors on mobile finish 46.0% of the checkouts they start, against 70.0% on desktop.
F02Heuristic
On mobile the cart hides the order summary, so the shipping cost first appears at checkout step 2; the desktop cart shows an estimate and the free-shipping threshold.
F03Competitor
Both competitors show the free-shipping threshold on the product page and a shipping estimate in the mobile cart; this store shows the threshold only in the footer on mobile.
Points at
C15
Checkout start to purchase, new, mobile46.0% against 70.0% (Desktop)3 findings from 3 sources point at C15: analytics, competitor, heuristic. One source alone caps Confidence at 2.
Register
Every finding has evidence with a receipt, a result that would prove it wrong, a score, a lane and an owner.
Impact against ease, sized by confidence
Quick WinImplement + MeasureBigger dot: higher confidence
Lane
Type
Status
| ID | Finding | Type | Lane | I / C / E | ICE | Status |
|---|
Click or press Enter on a row for evidence, cross-checks and the handoff.
Last run
None recorded.
Not assessed
What this run could not check. /cro:next works through them one at a time, worst cell first.
| Check | Why | How to close |
|---|---|---|
| Customer calls | None done yet | /cro:calls: a call guide from this run; you make 5 to 8 calls, the agent tallies the answers |
| Survey at mobile checkout | No survey tool in the demo | /cro:survey: a poll for new mobile visitors at checkout and one question after purchase; the agent tallies the export |
| Usability tests | None run yet | /cro:usability: five new visitors on their own phones, tasks through checkout; the agent tallies what they hit |
| Site speed | Not pulled in the demo | /cro:speed: needs a free PageSpeed API key |
| Retention and lifetime value | No order emails in the demo data | /cro:retention: needs order access or an export with emails |
| Competitor prices beyond one product | One product compared | /cro:pricing: the worst cell's top products against both competitors |
Rule check
0 errors, 0 warnings