Ecom pdp improvement list

PDP Improvement List

Operating question: Which product pages should be improved first?

Run this as an evidence-first operating play. Gather the evidence below, separate exact / estimated / partial / unavailable data, apply the decision rules and vetoes, and never invent missing numbers. Be adversarial toward platform self-reported figures — trust commerce truth.

Evidence to gather

• GA4, per PDP — sessions (or view_item events), add-to-cart events, and purchases, at landing-page and product-detail granularity, last 28 days plus the prior equal period. You need both the rate and the volume; a rate without sessions can't be ranked.
• Commerce catalog, per product/variant — selling price and contribution margin (or COGS), live stock by variant, review count and average rating, and image count per PDP.
• A benchmark CVR per page — not a single site average. Derive it from the page's own category and traffic mix (a sale-collection PDP and a cold-paid PDP do not share a benchmark).
• AOV for the product — to convert recovered conversions into recovered revenue.

Optional, if available:

• PDP load time / Core Web Vitals (LCP, INP) per template — slow pages bleed mobile add-to-cart.
• Traffic-source split per PDP (organic / paid / email / direct) — separates an intent problem from a page problem.
• Recent merchandising or price changes, and launch date — a 9-day-old PDP has no stable CVR yet.
• On-site search and PDP exit/bounce rate — corroborates where attention dies.
• Returns rate by product — a fit/expectation gap that no PDP edit alone will close.

How to decide — in order

1. Gate on tracking. If GA4 purchases for the window diverge from store orders by more than ~20%, stop — set the whole list to FIX and repair events first. A wrong CVR denominator mis-ranks everything.
2. Filter to a rankable sample. Drop any PDP with fewer than ~300 sessions in 28 days (or under ~5 transactions) to a holding WATCH — low CVR on low traffic is noise, not signal, and will dominate a naive "worst CVR" sort.
3. Compute the gap, then the dollars. For each surviving PDP: gap = benchmark CVR − actual CVR. If gap ≤ 0, it's at or above benchmark → KEEP. If gap > 0, recoverable revenue = sessions × gap × AOV × margin. Rank descending by recoverable revenue. This is the queue.
4. Apply the veto pass (see Veto Rules) before diagnosing — out-of-stock variants, pre-order/coming-soon, and single-source-intent pages get pulled or relabeled here so you don't "fix" a page that isn't broken.
5. Diagnose the root cause on the top recoverable PDPs by reading the micro-funnel. Add-to-cart rate is the fork:
   • Low session → add-to-cart = the page failed to convince. Look at images (count/quality), reviews (count < ~5 or rating < ~4.0), missing size/fit/spec info, unclear price/shipping. → REFRESH.
   • Healthy add-to-cart but low purchase = the leak is downstream (cart/shipping/checkout), not the PDP. → relabel; don't queue it as a PDP fix.
   • Slow LCP on mobile + add-to-cart gap concentrated on mobile = performance. → FIX (template/perf), route to web.
6. Assign status + owner + recheck date. Each top PDP gets one named driver, one owner, and a date to re-measure after the change.

The prompt to run

You are my ecommerce merchandising analyst running the "PDP Improvement List" play.

GOAL: rank my product detail pages by RECOVERABLE REVENUE and assign each top page a
diagnosed root cause and a fix owner. Recoverable revenue = sessions x (benchmark CVR -
actual CVR) x AOV x margin. Rank by dollars recoverable, NOT by worst conversion rate.

I will paste a per-PDP table: page, sessions (28d), add-to-cart rate, actual CVR,
benchmark CVR, AOV, margin %, review count, avg rating, image count, stock status by
variant, and mobile LCP where available. Some fields may be missing.

PRE-FLIGHT: First list which required inputs I provided vs. missing. If per-PDP sessions
(traffic) + CVR (and margin) is missing, STOP and return only (a) what's missing and
(b) how to get it -- never estimate it or proceed. Without per-PDP traffic you cannot
compute recoverable revenue, and without margin you cannot rank by contribution.

RULES:
- Tracking gate first: if I tell you GA4 purchases and store orders diverge >20%, mark
  the whole list FIX and stop ranking.
- Drop any PDP under ~300 sessions (or <5 transactions) in 28 days to WATCH. Low CVR on
  low traffic is noise, never a top-of-queue fix.
- For rankable pages: gap = benchmark CVR - actual CVR. If gap <= 0, KEEP. Else
  recoverable = sessions x gap x AOV x margin. Rank descending by recoverable revenue.
- VETO before diagnosing: if a PDP has out-of-stock or pre-order variants, do NOT call it
  a page problem -- the stock state explains low CVR. Flag it, don't queue it as REFRESH.
- VETO on intent: if a low-CVR page's traffic is mostly one cold source (e.g. broad paid),
  say so -- segment before judging; intent differs by source.
- Diagnose via the micro-funnel: low add-to-cart = page failed to convince (images,
  reviews, fit/spec info, price clarity) -> REFRESH. Healthy add-to-cart but low purchase
  = downstream cart/checkout leak, NOT a PDP fix -> relabel. Mobile-only add-to-cart gap
  with slow LCP -> FIX (performance).
- Every row must carry a number, source, time window, and confidence level. Separate
  exact / estimated / partial / unavailable evidence. Do not invent missing data.

RETURN:
1. A 3-sentence executive read naming the single highest-recoverable PDP.
2. A table ranked by recoverable revenue, using EXACTLY this header row:
   | PDP | Sessions (28d) | Actual CVR | Benchmark CVR | Recoverable $ | Driver | Status | Owner | Recheck |
   Use "--" for any cell you cannot fill. Do not add or drop columns, and do not replace
   the table with prose.
3. Vetoes/caveats that pulled or downgraded any page.
4. What evidence is blocked and what you'd need to upgrade a WATCH/FIX to a fix decision.

Decision rules

• REFRESH — rankable sample (≥300 sessions / 28d), actual CVR below benchmark by a gap that yields material recoverable revenue, and the add-to-cart drop traces to a fixable on-page cause (thin/low reviews, ≤1 or poor image, missing size/fit/spec, unclear price or shipping). The page has a credible reason to improve.
• FIX — tracking drift (GA4 vs. store orders >20%), a broken GA4↔catalog join, or a mobile performance problem (slow LCP) that blocks a clean PDP read or is itself the cause.
• WATCH — directional only: under ~300 sessions or ~5 transactions in 28 days, a PDP launched in the last ~14 days, or a window polluted by a promo or stockout.
• KEEP — actual CVR at or above its page-specific benchmark (gap ≤ 0); nothing material to recover.
• KILL — reserved: a chronically dead SKU (high traffic, near-zero CVR over multiple clean windows, no stock/intent veto) that merchandising recommends delisting rather than improving. Rare for a PDP play; most pages are REFRESH, not KILL.
• Every recommendation must include a number, source, time window, and confidence level.

Vetoes — stop if any apply

• Do not label a PDP a page problem when it has out-of-stock or pre-order variants — the stock state explains the low CVR, not the layout. Pull it from the REFRESH queue; route to planning.
• Do not judge a low-CVR page before segmenting by traffic source — a page fed mostly by broad cold paid traffic carries different intent than one fed by branded search; the "low" CVR may be a mix artifact.
• Do not rank by worst CVR — low CVR on low traffic is noise; it must clear the session floor to enter the queue.
• Do not queue a PDP fix when the add-to-cart rate is healthy but purchase is low — the leak is in cart/checkout, downstream of the page.
• Do not make a recoverable-revenue claim without margin and AOV; revenue at risk is not contribution at risk.
• Do not push template, copy, image, or price changes live without an explicit human approval step.

Output

A list of PDPs ranked by recoverable revenue, each with a diagnosed driver, owner, and recheck date.

PDP	Sessions (28d)	Actual CVR	Benchmark CVR	Recoverable $	Driver	Status	Owner	Recheck
/products/example	3,400	1.1%	2.6%	$4,180	2 images, no reviews	REFRESH	Merch	14 days

---

Run this on your live store data in one prompt — start a free 14-day trial of ShopMCP (no credit card). It connects your store, ads, analytics, email/SMS, and finance and runs this play read-only by default, applying changes only on your approval.