What we build
Every layer of recovery — orchestrated, not overlapping.
Event detection feeds segmentation. Segmentation feeds channel choice. Channel choice respects frequency and consent. Every touch is measured against incremental revenue, not last-touch credit.
Email + SMS + on-site exit sequences
Layered recovery flows across email, SMS, and on-site exit-banner — sequenced with cadence rules, consent handling, and locale-aware copy per market.
WhatsApp recovery for SEA
WhatsApp Business API recovery for SG / MY / ID / TH / VN — template-approved, compliance-gated, and routed only where it's the highest-conversion channel for the segment.
Retargeting + suppression
Meta and Google retargeting coordinated with lifecycle flows — suppressions synced so retargeting doesn't double-hit a customer already in an email or SMS sequence.
Smart frequency capping
Per-customer frequency caps across every channel combined — not counted per flow. Quiet-hours per region, fatigue scoring, and rate-limiting built in.
Revenue tracking per touch
Last-touch and incremental revenue per recovery touch. Holdout tests separate the recovery value from the would-have-converted-anyway baseline.
Segmentation by value + tier + LTV
Recovery pressure sized to cart value, product tier, lifecycle stage, and prior LTV — VIPs get one gentle nudge; thin-margin carts get a more aggressive cadence.
Recovery motions we build
Every shape of abandonment. One orchestration plane.
Cart, browse, wishlist, post-purchase win-back — same orchestration layer, different segment definitions and channel cadences. Recovery is a system, not a campaign.
High-velocity DTC stores
Shopify / Magento / headless DTC brands with thousands of carts abandoned monthly — recovery sequences scaled without manual campaign wrangling.
Cross-channel merchants
Brands selling on their own site plus marketplaces — cross-channel suppression so a Shopify recovery flow doesn't fight a Shopee promo already in-flight.
SEA markets · WhatsApp-heavy
SG / MY / ID / TH customers strongly prefer WhatsApp — recovery flows route WhatsApp-first where consent and template approval allow, email/SMS elsewhere.
High-AOV premium retailers
AOV-heavy carts (luxury, homeware, electronics) where a single recovered cart is worth a multi-touch budget — content-rich, personalised recovery sequences.
Browse-abandon + wishlist
Pre-cart recovery — customers who viewed products repeatedly or wishlisted without buying. High-intent, low-friction recovery before they cool off.
Post-purchase win-back
Lapsed customers past typical repurchase windows — reactivation campaigns scored against predicted win-back probability and latent LTV.
Model families we deploy
No single model runs every recovery decision. So we compose them.
Classifiers score intent, bandits pick channels, send-time optimisers pick windows, attribution agents reconcile credit. Each model evaluated independently and retrained on rolling data.
Gradient-boosted classifier scoring every abandon event by recovery probability — combines cart value, engagement depth, prior buyer signal, and real-time session features.
Bandit model picks the highest-engagement window per customer × channel — learned from historical opens, clicks, and responses across email, SMS, and WhatsApp.
Chooses the next best channel (email, SMS, WhatsApp, retargeting) per customer per touch — balances expected conversion, fatigue risk, and channel cost.
Reconciles recovery revenue against holdout groups so every channel gets credit for what it incrementally pulled back — not what it merely last-touched.
Data sources wired into every recovery touch
Every signal that shapes a recovery decision — integrated.
Streamed in real time, consent-filtered at ingestion, normalised into one customer × touch schema, versioned alongside the model that acts on them.
Explainable recoveries, not black-box sends
A recovered cart isn't enough. Finance needs the trail.
Every recovery touch ships with its drivers — segment, cart value, prior LTV, consent state, channel pick — ranked and logged. Marketing can audit, finance can attribute, legal can defend every send.
- Top drivers per touch with contribution weight
- Consent state captured at decision time
- Holdout-tested incremental attribution per channel
- Opt-out + frequency-cap reason codes logged
Frameworks we align to
Why Axccelerate for cart recovery
Not another email flow.
An orchestration plane.
A standalone email platform recovers carts via email. Our orchestration plane recovers them across email, SMS, WhatsApp, and retargeting together — capped, suppressed, and attributed as one motion.
Pricing
Priced to the channels and flows, not the send volume.
Recovery deployments are custom — we scope against your channel count, list size, and market compliance before quoting.
Glossary
The vocabulary behind every recovery decision.
A quick reference for the language lifecycle, deliverability, and attribution teams all use — and the terminology compliance reaches for when reviewing your consent stack.
- Abandoned cart
- Cart created · not completed
A cart where a customer added SKUs but didn't complete checkout. The primary recovery target — typically recoverable at 8-28% depending on segment and cadence quality.
- Browse abandonment
- PDP view without cart-add
Repeated product-detail-page views or wishlist adds without a cart. Pre-cart recovery targets this segment with lower-pressure nudges and social-proof content.
- Send-time optimisation
- Best-window send selection
Picking the optimal moment to send per customer × channel based on historical engagement — usually 15-25% lift over a static send-time schedule.
- Frequency cap
- Max touches per customer window
The limit on total touches across all channels within a rolling window. Set too high and you fatigue customers; set too low and you leave recoverable revenue on the table.
- Suppression list
- Do-not-send register
The registry of customers who cannot be contacted on a given channel — opt-outs, hard bounces, spam complaints, frequency-cap holds, quiet-hours gates.
- Deliverability
- Inbox-placement rate
The share of sends that actually arrive in the inbox (not spam / not bounced). Primary email health metric; degrades fast with bad list hygiene or over-sending.
- Soft bounce
- Temporary delivery failure
A failed delivery that's likely recoverable (mailbox full, server unreachable). Typically retried a few times before being converted to a hard bounce.
- Deliverability rate
- Sent · received share
Delivered messages as a share of attempted sends. Leading indicator for sender reputation — below 95% email deliverability is an investigation trigger.
- Conversion touch
- Touch preceding purchase
A recovery touch (email, SMS, WhatsApp, ad) that occurred in the session window leading to checkout completion — the attribution unit behind channel credit.
- Last-touch
- Credit to final channel
Attributing all of recovered revenue to the most recent touch before conversion. Simple but biased — heavier channels like retargeting get over-credit.
- Incremental revenue
- Lift above baseline
Revenue attributable specifically to a recovery channel, measured against a matched holdout group that didn't receive that channel — the fair credit signal.
- Opt-out
- Customer-initiated consent withdraw
A customer unsubscribing from a channel. Must be honoured within required windows (CAN-SPAM 10 days, etc.) and propagated to every connected channel immediately.
- Honouring
- Respecting gates + caps
The process of enforcing consent, quiet-hours, and frequency rules before any send goes out — the compliance spine of every flow, audited continuously.
- Re-engagement
- Lapsed-customer reactivation
Flows that target customers past the typical repurchase window — different from cart recovery, scored against win-back probability and latent LTV potential.
Your recovery plane, engineered.
30-minute scoping with a senior engineer and a lifecycle operator. You'll leave with a recovery plan, integration sketch, and realistic timeline — not a sales pitch.