Benchmarks & Guardrails (Directional)

This page is a sanity-check tool, not a promise.

Walmart Connect performance varies dramatically by: - category (CPG vs hardlines vs seasonal) - item price / AOV - retail readiness (in-stock, buy box, reviews) - placement mix (Search vs PDP vs Display) - promotion/calendar (events change auction pressure)

Use this page to answer: - "Is this obviously broken?" - "What should we fix first?" - "What do we tell stakeholders about what’s normal?"

Confidence labels

High (internal observed): validated with your account history (see internal tables below).
Medium (external directional): published by reputable vendors/tools; not your account-specific reality.
Low (vendor/marketing): may be cherry-picked; only use as a loose reference.

Metric definitions (what to look at first)

Metric	What it measures	Use it for	Biggest gotchas
CTR	click appeal (creative + relevancy)	diagnosing targeting/creative mismatch	highly placement-dependent
CVR	PDP/offer conversion after click	diagnosing retail readiness + intent	stock/buy box can tank it
CPC	cost per click	diagnosing auction competitiveness	reserve prices + competition day-to-day
CPM	cost per 1k impressions	diagnosing display inventory cost	value depends on viewability + frequency
ROAS	efficiency (sales / spend)	stakeholder reporting + scaling	attribution windows + halo muddy the story
Spend / pacing	delivery health	diagnosing underdelivery	budgets can cap before peak hours

Guardrails (if X happens, do Y)

1) Underspend / not delivering

Treat underspend as a structured checklist, not a guess.

When to worry: - Spend is <25% of daily budget by mid-day (local time), and you expected steady delivery. - New campaign spends $0–$10 after 12–24 hours.

First actions: - Check approvals/policy (creative + product eligibility). - Check in-stock + buy box for advertised items. - Broaden targeting (especially for Display audiences). - Raise bids deliberately (small steps) and isolate changes.

Go deeper: - Performance Triage

2) High CPC / high CPM

When to worry: - CPC spikes and efficiency drops without a change in conversion rate. - CPM rises and reach collapses (frequency rising but unique reach not).

First actions: - Tighten to higher-intent targeting (phrase/exact; stronger audience signals). - Split campaigns by placement intent (Search vs PDP conquest) so you can change bids cleanly. - Improve retail readiness (PDP quality and offer) instead of buying your way out.

3) Low CTR

When to worry: - CTR is trending down as spend increases (you’re scaling into weaker inventory).

First actions: - Refresh first-frame video hooks / headline; clarify offer. - Narrow to fewer messages (one promise per asset). - Switch targeting type (persona prospecting → behavioral retargeting; or vice versa).

4) Good CTR but weak CVR

When to worry: - People click, but they don’t buy, and CVR is materially worse than your comparable items/campaigns.

First actions: - Check buy box and in-stock first (do not waste spend if broken). - Evaluate pricing vs main competitors. - Fix PDP basics: title clarity, imagery, review count, variants. - Ensure landing page matches the ad promise (shelf merchandising matters).

KPI hierarchy by objective (what to optimize first)

Objective	Primary KPI	Secondary KPI	What not to overreact to
New launch	CTR + learning (search terms)	CVR trend by week 2–4	early ROAS
Defend	branded impression share + ROAS	CVR	external benchmark CTR
Seasonal	pacing + conversion KPIs in peak window	CTR in pre-season	ROAS differences pre vs peak
Challenger	CVR (proof of differentiation)	CPC in controlled tests	top-of-search vanity

Internal observed benchmarks (from exports)

These are computed from CSV exports in backend/wmc-docs/reports/.

Refresh command: - python3 backend/wmc-docs/scripts/generate_benchmarks.py

High (internal observed) derived from backend/wmc-docs/reports/ exports.

Method (so we don't fool ourselves): - Aggregation level: campaign + placement totals across time. - Filter: ≥ 50 clicks per unit (reduces noise). - Bands: p25 / p50 / p75 across units. - CVR uses total orders / total clicks per unit (click-weighted across days).

Coverage caveats: - These are from a single advertiser set + time window; do not generalize. - Small slices are omitted to avoid making decisions on noise.

Date coverage (from placement export): 2025-10-20 to 2025-12-18

Sponsored Products / Brands / Video — sliced benchmarks

Interpretation: use these as a healthy vs suspicious range check, not KPI targets.

External benchmarks (optional)

External benchmarks change and are often biased. If we add them, keep them labeled (Medium/Low) and dated.

Slice	Placement	Units	CTR (p25/p50/p75)	CPC (p25/p50/p75)	CVR (p25/p50/p75)	ROAS (p25/p50/p75)
TaiPei · EVG · SP	Carousel	5	0.42% / 0.51% / 0.76%	$0.55 / $0.56 / $0.71	58.39% / 65.88% / 75.45%	7.41 / 11.04 / 13.77
TaiPei · EVG · SP	Search In-grid	5	1.78% / 2.11% / 3.38%	$0.61 / $0.61 / $0.81	56.78% / 60.22% / 60.67%	9.13 / 9.47 / 11.01
TaiPei · OTHER · SP	Browse In-grid	6	0.72% / 0.83% / 0.83%	$0.44 / $0.62 / $0.72	40.28% / 41.68% / 48.63%	6.30 / 7.34 / 7.51
TaiPei · OTHER · SP	Buybox	8	0.32% / 0.58% / 0.96%	$0.54 / $0.65 / $0.82	24.29% / 37.29% / 45.43%	4.63 / 6.49 / 8.47
TaiPei · OTHER · SP	Carousel	8	0.37% / 0.54% / 0.72%	$0.39 / $0.49 / $0.54	59.58% / 76.26% / 89.04%	11.65 / 14.29 / 22.67
TaiPei · OTHER · SP	Homepage	6	0.28% / 0.32% / 0.36%	$0.59 / $0.61 / $0.62	77.16% / 90.14% / 98.34%	12.70 / 16.20 / 18.91
TaiPei · OTHER · SP	Search In-grid	8	1.16% / 1.82% / 3.36%	$0.54 / $0.65 / $0.78	48.96% / 59.71% / 70.04%	4.15 / 13.07 / 14.81
TaiPei · OTHER · SP	Stock Up	6	1.47% / 1.67% / 1.90%	$0.54 / $0.56 / $0.59	108.59% / 109.15% / 116.23%	13.29 / 19.52 / 26.28