CRO Program — Sun Day Red × Slice of the Pie

Same traffic.
More revenue, in ninety days.

The audit found the leaks. This is the repair schedule: three phases, sequenced by payback speed, ending in a disciplined testing program — because after week 13, growth should come from evidence, not opinions.

12Ranked fixes from the audit Site Audit, Module 02

90Day roadmap, 3 phases

8A/B tests, every hypothesis cited

The grades set the order of operations.

Module 02 graded nine areas of sundayred.com from direct observation. The program below simply works the board: the D-grades first (they leak the most, and most are low-effort), the trust and retention gaps second, and a compounding test engine third. Two areas remain pending — Performance and Site Search — and get graded live in the meeting; if Core Web Vitals fail, performance work is inserted at the top of Phase 1.

Three phases. Payback-ordered.

Phase 1 · Weeks 1–3 Phase 2 · Weeks 4–8 Phase 3 · Weeks 9–13

Phase 1 · Stop the bleeding
Weeks 1–3 · low-effort, high-leak fixes

In-stock-first sorting on collection pages — OOS demoted with waitlist chips (audit fix #1: 6 of first 8 polos OOS at the top of "Best Match" Site crawl, Aug 2026).
PDP Klarna line — "4 interest-free payments of $28.75" under the price (fix #4: Klarna live at checkout, invisible upstream sundayred.com/klarna.html).
Cart rescue — recommendations row, Continue Shopping, trust + payment strip on the dead-end empty cart (fix #7).
Incentivized welcome offer on the existing email/SMS capture (fix #5 — offer economics agreed with your team first).
Escalate the restock / demand-planning flag to buying (fix #2) — the one lever no website change replaces.

Exit criteria: no shopper lands on dead stock first; Klarna visible at the decision point; cart has recovery paths; capture has a reason to say yes.

Phase 2 · Build trust
Weeks 4–8 · proof and reassurance layer

Reviews program launch (fix #3) — platform selection, post-purchase collection flows seeded from the existing waitlist/back-in-stock audience, PDP display module, review schema staged for SEO.
Free-returns merchandising at the ATC — "30-day free returns · prepaid FedEx label", their real policy, currently invisible on PDP (fix #6 Site crawl, Aug 2026).
Free-shipping progress bar toward the $150 threshold in cart and mini-cart (fix #8).
Policy advisory: refund-speed and instant-exchange upgrade path (fix #9); content-ops freshness cadence (fix #10).

Exit criteria: first reviews live on PDPs; every trust asset the brand already owns is visible at the moment of decision.

Phase 3 · Compound
Weeks 9–13 · the test engine takes over

A/B testing program goes live — the 8-test roadmap below, run against the clean baseline captured in weeks 1–2.
Structured-data completion (fix #11) once reviews exist to mark up — Product / Offer / AggregateRating.
Backlog grooming: every shipped fix generates follow-on test candidates; the roadmap is re-ranked every two weeks on observed results.
Performance and site-search work slotted here or earlier, pending the live-meeting runs (fix #12).

Exit criteria: a running experimentation cadence with guardrails — the store improves on evidence from week 13 onward.

Eight tests. Zero guesses.

Every hypothesis is tied to a finding we observed and cited in the site audit. Expected direction is stated qualitatively — we will not promise uplift percentages before your baseline exists. Where a range appears, it is labeled illustrative.

#	Test & hypothesis	Primary metric	Secondary	Min runtime	Expected direction
T1	In-stock-first PLP sort vs. current "Best Match" Because 6 of the first 8 polos were OOS at the top of the default sort (audit: Merchandising, D), surfacing buyable product first will lift grid engagement and downstream orders.	PLP → PDP click-through	PLP-sourced CVR, bounce	2 weeks / full biz cycle	↑ CTR, ↑ CVR
T2	PDP Klarna line ("4 interest-free payments of $28.75") vs. none Klarna is live at checkout but unmerchandised on PDP (audit: PDP, C-). Surfacing pay-in-4 at the price will reduce sticker hesitation on the $115–$145 ask.	Add-to-cart rate	CVR, AOV, Klarna share	3 weeks	↑ ATC
T3	Free-returns badge at the ATC vs. none 30-day free returns with prepaid FedEx label is real but invisible on PDP (audit: Trust, B-). Reassurance at the decision point lowers purchase risk.	PDP → order CVR	ATC rate, return rate (guardrail)	3 weeks	↑ CVR
T4	Welcome offer: first-order incentive vs. early-access-only framing Capture currently offers no incentive at all (audit: Email/SMS, C). Testing offer depth finds the cheapest capture-rate winner. Range shown is a planning band, not a promise: Illustrative offer tiers, e.g. 10% vs. access-only.	Email/SMS capture rate	Welcome-flow revenue per recipient, margin (guardrail)	2 weeks	↑ capture
T5	Free-ship progress bar in cart vs. static banner The $150 threshold sits above every single-polo price with no progress merchandising (audit: Cart, D). A live "you're $35 away" bar nudges the designed 2-item basket.	AOV	% orders ≥ $150, cart abandonment	3 weeks	↑ AOV
T6	PDP reviews module vs. control (runs once the Phase-2 program has collected a displayable base) Zero social proof exists at a premium price (audit: PDP, C-). Verified reviews reduce first-purchase risk on a young brand.	PDP → order CVR	Time on PDP, return rate	4 weeks	↑ CVR
T7	Cart recommendations row vs. none (with items in cart) The cart currently has no cross-sell surface at all (audit: Cart, D). "Pairs well with" logic already exists on PDP — extend it to cart.	AOV	Units per order, RPV	3 weeks	↑ AOV
T8	OOS PDP treatment: waitlist + in-stock alternatives rail vs. waitlist only The flagship Icon Polo II was unbuyable in all 7 colors with waitlist as the only path (audit: Merchandising, D). Offering buyable near-neighbors recovers demand that today just leaves.	Revenue per OOS-PDP session	Waitlist signup rate (guardrail — must not drop), exit rate	3 weeks	↑ RPV

Test-design notes: sequencing, sample size, guardrails

Sequential vs. parallel: tests on different funnel surfaces run in parallel (T1 on PLP alongside T2/T3 on PDP); tests sharing a surface run sequentially or in a multivariate slot only if traffic supports it. No overlapping treatments on the same element, ever.
Runtime discipline: minimum runtimes above are floors, not targets — every test runs whole weeks (full business cycles) and continues until the pre-registered sample size is reached. No peeking-based early stops.
Sample-size reality check: final test order depends on actual session volume per surface from GA4 — verify live in the meeting. Lower-traffic surfaces (cart) may need longer windows or larger minimum detectable effects.
Guardrail metrics: return rate (T3, T6), margin per order (T4 offer depth), waitlist capture (T8), and sitewide CVR monitored on every test — a primary-metric win that damages a guardrail is a loss.
Pre-registration: hypothesis, primary metric, minimum runtime, and stopping rule written down before launch. Results reported win, lose, or flat — flat results still retire a question.

Agree on the scoreboard before kickoff.

CVR orders ÷ sessions

Conversion rate — the program's headline metric, read per-surface (PLP-sourced, PDP-sourced) as well as sitewide, so wins are attributable.

AOV revenue ÷ orders

Average order value — the target of the free-ship bar, cart recommendations, and Klarna tests. Watched against the $150 threshold attach rate.

RPV revenue ÷ visitors

Revenue per visitor — the tie-breaker metric. CVR and AOV can trade against each other; RPV tells us whether the store actually got better.

Capture rate signups ÷ unique visitors

Email/SMS capture rate — the top of the retention engine. Read with downstream welcome-flow revenue per recipient, never alone.

Baseline plan: weeks 1–2 capture a clean, annotated 2-week baseline for all four KPIs before any Phase-1 change ships — so every fix is measured against a known floor. Some baselines (surface-level CVR splits, capture rate, checkout funnel steps) require GA4 and platform access — verify live in the meeting. Where history is dirty or missing, we baseline forward and say so.

Built on the audit, not on vibes.

Sources

Module 02 site audit — all findings cited there from sundayred.com rendered-page fetches, Aug 1, 2026 (PLP, PDP, cart, checkout, returns, klarna.html, mobile-messaging-program.html, robots.txt, sitemaps, giftcards, help.sundayred.com)
Klarna — official Sun Day Red store page (Pay in 4 terms, $35–$1,000 carts)
Semrush, US database, July 2026 data month (traffic context for surface prioritization)
Test-design conventions: pre-registered hypotheses, full-business-cycle runtimes, guardrail metrics — Slice of the Pie testing playbook

Verify live in the meeting

GA4 / analytics access — required for baseline splits, sample-size math, and final test ordering
Core Web Vitals runs (pending from Module 02) — a failing CWV result re-ranks performance into Phase 1
On-site search walkthrough (robots-blocked to crawlers) — may add a search test to the backlog
Welcome popup presence, guest checkout, and express pays — confirms the exact scope of T4 and the cart work
Testing-tool and consent setup on Salesforce Commerce Cloud (existing tooling, if any)

Same traffic.
More revenue, in ninety days.

The grades set the order of operations.

Three phases. Payback-ordered.

Eight tests. Zero guesses.

Agree on the scoreboard before kickoff.

Out think. Out work.
Out test.

Built on the audit, not on vibes.

Sources

Verify live in the meeting

Same traffic.More revenue, in ninety days.

The grades set the order of operations.

Three phases. Payback-ordered.

Eight tests. Zero guesses.

Agree on the scoreboard before kickoff.

Out think. Out work.Out test.

Built on the audit, not on vibes.

Sources

Verify live in the meeting

Same traffic.
More revenue, in ninety days.

Out think. Out work.
Out test.