BOPIS + curbside · Per-store abandonment classification · Multi-location retail

Your BOPIS completion is 86% at the brand level. Seven stores at 71% are pulling it down. Find the per-step gap.

You run BOPIS or curbside across 50-1,500 retail locations. The customer-experience dashboard shows one number: 86% completion. It looks acceptable. The per-store distribution hides the seven stores running at 71%. Worse — those seven stores are abandoning for different reasons. One has reliable lockers and slow staff. Another has fast staff and a broken notification integration. A third has chronic pick-pack errors. A brand-level improvement initiative can’t fix three different problems with one solution. Per-location per-step classification can.

Audit your BOPIS abandonment (Tier 1)Explore Fractional CMO operation

Published May 30, 2026

The 12 steps where abandonment hides

Store #47 has reliable lockers and slow staff handoff — customer-receive step. Store #112 has fast staff and a broken ready-notify integration — customer-arrive step (the customer arrived before they were notified). Store #229 has chronic pick-pack errors — the abandonment is downstream of an unfulfillable order.

A brand-level completion rate of 86% averages this away. Three stores at 96% and seven stores at 71% produce 86%. The seven stores are where 60-plus percent of the abandonment volume lives. Per-location per-step classification surfaces which step is failing at which store.

We’ve built per-location per-step classification for retail operators. Here’s what we know.

You probably already use an order-management system (Manhattan, Salesforce, NewStore, Aptos), an in-store workflow system, an ecommerce front-end (Adobe Commerce, Shopify, BigCommerce), and an inventory system (Cin7, FlowHub). Each system emits events. The gap is the classification layer that joins their per-step events per location and routes the resulting per-pattern diagnosis to the right team (staff rotation goes to ops; broken integration goes to systems; pick-pack errors go to the warehouse).

We have built this for multi-location retail operators. We know which canonical patterns recur (ready-notify lag, locker saturation, staff-handoff bottleneck, ID-verify friction, pick-pack error). We know how to route each pattern to the right remediation team without creating a cross-team blame loop. We bring the canonical pattern catalog and the per-pattern routing playbook.

How we get from one brand-level number to per-store per-step routing

Step 1 — Tier 1 AI Readiness Assessment ($10k, 2-3 weeks). We audit your current per-step event coverage across the BOPIS flow. We classify which steps emit events and which are dark. We sample 30-60 days of abandonments and run the canonical-pattern classifier against available signals. Output: the per-location per-step instrumentation gap map, the canonical-pattern baseline by location, and the per-pattern remediation routing spec.

Step 2 — Tier 2 AI Swarm Setup Sprint ($25-50k, 4-8 weeks). We build the friction-detection layer end-to-end: event-ingestion connectors for each upstream system, per-location per-step join pipeline, canonical-pattern classifier, per-pattern remediation routing to customer-recovery + staffing-rotation + attribution pipelines. Your engineering team receives the running system, all source code, all credentials.

Step 3 — Tier 3 Fractional CMO with AI Swarm ($15-25k/month, 6-month minimum, 1-2 days/wk). We operate the layer in production. Triage per-pattern classifications daily. Extend the canonical-pattern catalog as new patterns emerge in your data. Coordinate per-location service-recovery cadence with your operations + marketing leadership. Roll up a monthly per-pattern outcome report (which interventions produced measurable improvement).

What changes for you

You stop running brand-level improvement initiatives that miss the actual problem. The per-store per-step diagnosis surfaces which intervention each store needs.

You stop blaming store managers for abandonment they can’t fix without systems-team help. The per-pattern classifier names whether it is a staff issue, an integration issue, an inventory issue, or a customer- experience issue.

You can answer the question your VP of Operations asks every quarterly review: which patterns are driving network abandonment, and which interventions have produced measurable improvement. The per-pattern outcome rollup is the answer.

You can onboard a new format (a new store concept, a new curbside-only location) with the per-step instrumentation built in from day one rather than discovering the abandonment surface later.

Frequently asked

Why does the brand-level BOPIS completion rate hide where the actual abandonment is happening?

The brand-level rate is an average across all stores. A network with three stores at 96% completion and seven stores at 71% averages to 86%. The brand-level number looks acceptable; the actual abandonment concentrates in the bottom-quartile stores. Worse, the bottom-quartile stores are abandoning for different reasons. Store #47 has reliable lockers and slow staff handoff — the abandonment cluster is at the customer-receive step. Store #112 has fast staff and a broken ready-notify integration — the abandonment cluster is at the customer-arrive step (customer arrived before they were notified). Store #229 has chronic pick-pack errors — the abandonment is downstream of an unfulfillable order. The brand-level number cannot distinguish these. Per-location per-step classification can.

What are the 12 steps in the BOPIS flow and how does abandonment cluster across them?

Order-place, order-confirm, fulfillment-route, pick-pack, ready-notify, customer-arrive, locker-or-counter, ID-verify, staff-handoff, customer-receive, receipt-emit, satisfaction-survey. Each step can fail. Each store fails differently. Five canonical friction patterns recur across multi-location operations: ready-notify lag (notification fires more than 15 minutes late; customer arrives before notified); locker saturation (locker bank fills, overflow routes to counter, staff-handoff queue builds, customer leaves); staff-handoff bottleneck (single staff member during peak hours, queue forms); ID-verify friction (ID-check takes too long or fails repeatedly for legitimate customers); pick-pack error (wrong item picked, customer arrives, order is unfulfillable). Each pattern fires a different fix.

How does per-location per-step classification produce different remediations than a brand-level improvement initiative?

A brand-level initiative on staff training will help stores with staff-handoff bottlenecks. It will not help stores with broken ready-notify integrations or saturated lockers. A brand-level initiative on locker capacity will help stores with locker saturation. It will not help stores with chronic pick-pack errors or ID-verify friction. Per-location per-step classification surfaces which intervention each store needs. The staff-handoff store gets staff-rotation changes. The locker-saturation store gets locker-capacity changes. The notification-timing store gets ready-notify integration fixes. The pick-pack store gets pick-staff training. Three different stores, three different fixes, instead of one brand-level program.

What does Completions commit to on Tier 3 if we run this layer in production?

Tier 3 process commitments include: daily per-location per-step abandonment classification cycle; per-pattern remediation routing within 24 hours to the responsible per-location field-team or systems-team; weekly per-location per-step trend report routed to your operations leadership; monthly per-location per-pattern outcome rollup (which interventions produced measurable improvement, which did not); quarterly review of the canonical-pattern catalog as new abandonment shapes emerge in your data. We commit to the operating discipline. Per-pattern precision is tuned per stack and recorded as engagement KPIs.

Who owns the per-step event store, the canonical-pattern definitions, and the credentials post-engagement?

Your team owns the per-location per-step event store, the canonical-pattern definitions, the remediation playbooks, the per-location field-team roster, the systems-team escalation paths, and the credentials. Completions owns the orchestration knowledge: the per-pattern classifier-tuning history, the per-vertical pattern catalog, the per-location staffing-pattern remediation playbook. At engagement end we transition operational ownership back to your team over 30-60 days with documented handover.

How does the friction-detection layer connect to the rest of the operations + marketing stack?

The detection layer subscribes upstream to your order-management system (Manhattan, Salesforce, NewStore, Aptos), your in-store workflow system (NewStore or Aptos or platform-native), your ecommerce front-end (Adobe Commerce, Shopify, BigCommerce), and your inventory system (Cin7, FlowHub, platform-native). It joins these event streams per location per step and classifies each abandonment against the canonical patterns. Downstream it emits classified-abandonment events to the customer-recovery pipeline (service-recovery offers per customer per pattern), the staffing-rotation pipeline (per-location staffing adjustments against pattern frequency), and the attribution pipeline (per-location revenue impact + customer-lifetime-value erosion per location). Four upstream feeds, three downstream consumers, one classification contract.

Start with the audit

Tier 1 AI Readiness Assessment ($10k, 2-3 weeks): we audit your per-step event coverage, sample 30-60 days of abandonments, run the canonical-pattern classifier against available signals, and produce the instrumentation gap map + per-location per-pattern baseline + per-pattern routing spec. If you decide to build, Tier 2 ships the detection layer. If you decide to operate it with us, Tier 3 runs the daily classification cycle in production. You choose the next step at each gate.