Recommendations from the current engagement engine review, split into product workflow, marketing action, and measurement priorities.
Keep the app anchored around a weekly merchant review: refresh data, score customers, review segments and inventory, approve draft actions, then export.
Highest-value next build
Guided weekly run workflow
Marketing focus
Segment-specific campaigns, not one broad win-back
Risk control
Draft review before codes or exports
The engine has the pieces, but merchants still need to infer the operating sequence from separate screens and worker logs.
Add a guided run page that shows sync freshness, blocks scoring while sync is active, starts scoring, follows segmentation and tag-sync progress, then opens the review/export queue.
Customer scoring now uses primary category and recent purchase signals, but inventory fit is still a broad category/timing proxy.
Compute matched products per customer using category, price band, current stock depth, size availability, and campaign bucket; store the top matches against each analysis run.
The Phase 1 promise is merchant review and export, while the current offer worker can create Shopify discount codes directly from active rules.
Create draft offer records with copy, product picks, incentive posture, and governance flags first; only create Shopify codes after explicit merchant approval.
The inventory review is useful now, but campaign confidence improves when stock age and recent sell-through are based on snapshots rather than current quantity alone.
Add inventory age, sell-through bands, and stale-stock reasons to the existing inventory recommendations and CSV exports.
A consolidated checklist version of the recommendations, designed for ownership, notes, and status review.
Highest-value product and workflow improvements to make the app easier to operate and safer to use.
Turn sync, scoring, segmentation, and offer review into one weekly run
The engine has the pieces, but merchants still need to infer the operating sequence from separate screens and worker logs.
Add a guided run page that shows sync freshness, blocks scoring while sync is active, starts scoring, follows segmentation and tag-sync progress, then opens the review/export queue.
Owner:
Status:
Notes:
Replace remaining heuristic gaps with product-affinity and inventory-fit evidence
Customer scoring now uses primary category and recent purchase signals, but inventory fit is still a broad category/timing proxy.
Compute matched products per customer using category, price band, current stock depth, size availability, and campaign bucket; store the top matches against each analysis run.
Owner:
Status:
Notes:
Keep offer generation in draft-review mode before discount creation
The Phase 1 promise is merchant review and export, while the current offer worker can create Shopify discount codes directly from active rules.
Create draft offer records with copy, product picks, incentive posture, and governance flags first; only create Shopify codes after explicit merchant approval.
Owner:
Status:
Notes:
Use inventory snapshots for aging and sell-through confidence
The inventory review is useful now, but campaign confidence improves when stock age and recent sell-through are based on snapshots rather than current quantity alone.
Add inventory age, sell-through bands, and stale-stock reasons to the existing inventory recommendations and CSV exports.
Owner:
Status:
Notes:
Changes that turn separate screens into a repeatable weekly merchant review process.
Add a Weekly Review command center
Dashboard actions are split between manual sync and scoring, while segments and inventory live elsewhere.
Show one status rail: last product/order/customer sync, active jobs, latest analysis run, segment count, tag-sync status, draft offers, and export readiness.
Owner:
Status:
Notes:
Add approve, skip, hold, and note states for recommendations
Merchants need a review workflow, not just recommendations, so the app can remember judgment across runs.
Persist merchant decisions per customer/product recommendation and let future runs surface repeats, changed recommendations, and unresolved holds.
Owner:
Status:
Notes:
Surface consent, exchange detection, and product data warnings before campaign planning
The segmentation history shows consent and tag sync can materially affect segment counts and confidence.
Add pre-flight checks for marketing eligibility freshness, Loop exchange detection rate, products without productType, products without images, and stale sync jobs.
Owner:
Status:
Notes:
Create role-specific exports
Campaign-safe product lists and social-opportunity lists serve different teams and should stay separate.
Add export presets for email campaign planning, socials, suppression/risk review, and merchandising restock review.
Owner:
Status:
Notes:
Campaign and audience ideas that can be worked through with merch, email, and social teams.
Run a curated comeback edit
These customers have proven spend and should not be treated like broad discount shoppers.
Pair safe-to-promote products with light urgency, VIP framing, and a moderate incentive only where needed.
Owner:
Status:
Notes:
Send category-specific newness and back-in-stock campaigns
Their category signal is the value; generic win-back messaging dilutes the reason to return.
Match primaryCategory to stocked products, avoid cross-category recommendations, and lead with relevance before discount.
Owner:
Status:
Notes:
Protect margin with exclusivity instead of discounts
The playbooks correctly identify this segment as high-margin and at risk of unnecessary discount conditioning.
Use early access, limited arrivals, personal styling notes, or gift-with-purchase rather than a percentage code.
Owner:
Status:
Notes:
Build demand for visual stocked products before they become stale
The inventory review already separates social opportunity from campaign safety, which is exactly the right team split.
Give the social team a weekly list with product image, hook, stock reason, and whether the item needs demand creation or clearance support.
Owner:
Status:
Notes:
Suppress risky customers from broad offers
High exchange or return behavior can make reactivation look better than it is on gross spend.
Keep these customers in analysis, but route them to service-led review, fit guidance, or suppression rather than discount campaigns.
Owner:
Status:
Notes:
Reporting and risk controls that make each future run more useful than the last.
Add run-over-run comparison
The app stores analysis runs but the UI does not yet turn them into a habit loop for weekly review.
Show segment movement, newly lapsed customers, recovered customers, offer approvals, and inventory bucket changes between the latest two completed runs.
Owner:
Status:
Notes:
Track manual campaign results back into the next run
Phase 1 avoids sending, but it still needs feedback from exported/manual campaigns to improve confidence.
Let the merchant import or mark campaign cohorts, then report repeat purchase, revenue, exchange rate, and discount usage after outreach.
Owner:
Status:
Notes:
Show why a customer or product should not be targeted
Suppression recommendations are as valuable as opportunity recommendations when exchange risk and poor size spread are involved.
Expose reasons like non-opt-in, high exchange rate, open return, weak size spread, low stock, or no product image directly in review rows.
Owner:
Status:
Notes: