Guides versus app reviews
App reviews score products. Guides score decisions. A Klaviyo cart flow and a Sequenzy cart flow both fail when timing is instant for every AOV, SMS fires parallel to email one, and discounts stack with welcome offers. Guides document the operating system — when to wait, when to text, when to offer nothing, how to measure incrementality — that makes any competent platform recover profit.
Timing playbook — the highest-leverage guide
First-touch delay should vary by cart value and vertical consideration cycle. Impulse consumables: 45–90 minutes. Enthusiast hobby catalogs: 2–4 hours to avoid surveillance tone. High-AOV trust purchases: 4–12 hours with financing or guarantee copy on email one. Second email: 24 hours standard; 48–72 for fashion when size deliberation is norm. SMS branch: 4–6 hours after email one if unopened and TCPA consent exists — not simultaneous blast.
Quiet hours matter for SMS and international email — respect recipient local evening. BFCM compresses delays only for engaged segments; cold profiles stay suppressed regardless of merchant urgency.
Margin-safe incentives — stop the discount treadmill
Email one: product reminder, shipping clarity, trust signals — no code. Email two: social proof, scarcity only when inventory confirms. Email three: graduated offer capped by cart value and prior discount history. VIP full-price loyalists may never see cart discount — correct behavior. Document margin floor with finance; recovery rate without margin is a vanity metric.
Channel mix without fatigue
Email owns narrative and proof. SMS owns urgency when opt-in and economics justify fee. Push owns Android-heavy flash sales only when attribution positive — iOS limits apply. One intent window, one primary offer. Cart SMS while cart email one is unopened is escalation; cart SMS while email one is unread in inbox is spam training.
Checkout friction diagnosis
Recovery cannot fix broken checkout. If abandons cluster on shipping quote step, recovery emails should address shipping transparency — not deeper discounts. Payment failures need different trigger than browse-and-leave. Export Shopify checkout funnel monthly; if recovery rises while checkout completion falls, fix checkout before adding email three.
Attribution and holdouts
Hold out 5% of eligible abandons monthly. Compare purchase rate sent versus excluded cohorts. Report incremental recovery revenue and contribution margin after discounts and channel costs. ESP attributed cart revenue overstates lift when loyal return buyers would have purchased anyway. Finance spreadsheet beats dashboard screenshots in board meetings.
Consent and compliance separation
Email cart recovery requires marketing consent. SMS cart recovery requires TCPA-documented opt-in — checkout checkbox, keyword, or compliant popup. Email consent does not transfer to SMS. Quiet hours, opt-out handling, and message frequency caps are legal and deliverability requirements, not nice-to-haves.
Flow collision prevention
Priority stack: transactional and shipping always win; cart beats browse; cart suppresses welcome offer for 48-hour new subscribers; winback pauses when cart active; loyalty points offers defer to cart completion. Document in wiki; audit quarterly. Collisions cause unsubscribes that look like creative problems but are architecture bugs.
Guide ownership for lean teams
Assign one owner per guide family: timing and channel mix (ops lead), incentives and discount tests (retention plus finance), attribution (retention plus finance monthly), consent (legal or ops with TCPA checklist). Owners publish changes to shared collision doc before peak season. Guides without owners become stale — timing ladders drift, SMS fees spike, holdouts stop running.
BFCM recovery prep
October: freeze new discount experiments on cart; define engaged segment for recovery. November 1–15: engaged-heavy recovery only; suppress 90-day inactive from cart promos. November 16–30: no migration, no new SMS provider, no re-permission blasts. December: measure holdout incrementality before declaring victory — peak recovery rate without margin math misleads.
Discount testing governance
Run incentive tests only after baseline no-discount email one performs for 30 days minimum. A/B subject lines before A/B discount depth — creative fatigue masquerades as incentive problem. Hold out 5% on any test including discount; report margin not conversion alone. Finance signs off on maximum discount depth per cart band before tests go live. Failed tests revert in 48 hours — do not leave 20% cart code running because test "won" on revenue while margin collapsed.
Inventory-aware recovery deep dive
Sync inventory at send time via Shopify variant API — not flow enrollment snapshot. Partial cart OOS: recover in-stock lines, acknowledge unavailable item with waitlist link. Full cart OOS: suppress entire send, enroll waitlist if opted in. Fashion size runs and limited drops live this daily — sending "last chance" for sold-out size trains customers to ignore future urgency. Klaviyo, Sequenzy, and Omnisend support dynamic blocks when catalog sync healthy; operator must configure check not assume default.
SMS quiet hours and regional timing
TCPA quiet hours typically 8am–9pm recipient local time — cart SMS at 10pm trains opt-outs. Postscript and Attentive enforce configurable windows; verify timezone on subscriber record not store timezone. International SMS adds cost and compliance layers — US-only list should not receive UK sender experiments without legal review.
Email quiet hours softer but exist for brand tone — luxury jewelry recovery at 3am feels predatory. Schedule sends by recipient timezone when platform supports; otherwise bias toward late morning and early evening local. BFCM exception: engaged segments only; never cold list 2am blasts.
Checkout friction taxonomy
Classify abandons before tweaking recovery: shipping sticker shock, account creation wall, payment method missing, trust uncertainty, comparison shopping, technical error. Each class gets different recovery copy — shipping shock needs policy transparency; comparison shopping needs proof and differentiation not discount. Monthly funnel export from Shopify Analytics drives classification; guessing produces wrong incentive.