Multi-store retail — 6 outlets, one owner-operator

Every Monday, the owner spends four hours finding out what the weekend actually did.

Six stores generate a weekend's worth of POS exports, payment-gateway settlements, delivery-app statements, WhatsApp complaints, and stock notes. Monday morning is spent stitching them together to answer three questions: does the money reconcile, which customers are angry, and what's about to run out. The answers exist by noon — after the freshest problems have aged three days.

The situation

The weekend is a black box until Monday noon

Each store closes Saturday and Sunday with its own POS export. Card settlements arrive from the gateway on their own schedule, delivery-app orders settle weekly with fees deducted, and cash counts come in as photos of handwritten sheets in the managers' WhatsApp group. Reconciliation means matching all of it — and the mismatches that matter (a settlement short by a few thousand, a store whose cash count doesn't match its POS) hide among dozens of routine timing differences.

Customer issues take the same path: complaints arrive via store WhatsApp numbers, Google reviews, and the delivery apps. A refund promised by a weekend staffer at one store is invisible to the owner unless someone remembers to forward the thread. Stock is worse — managers note shortages verbally, and the reorder happens when someone remembers, which is how the best-selling SKU goes out of stock at two stores while sitting overstocked at a third.

The owner doesn't want an autonomous system touching money or messaging customers. He wants to open one document Monday at 8am and know where to point his attention.

The work

What Taz builds for this chain

Taz designs a weekly pulse workflow: ingest the weekend's exports, apply written matching and triage rules, and produce one digest that leads with exceptions — reviewed by the owner, acted on by humans.

  • Ingestion and normalization: POS exports, gateway settlement files, delivery-app statements, and the cash-count sheet format are mapped into one normalized structure per store per day, using the exports the systems already produce — no new hardware, no POS replacement.
  • Reconciliation rules: Written matching logic — what counts as a routine timing difference versus an exception worth attention, with thresholds the owner sets. Every exception in the digest shows the underlying numbers so it can be verified in one glance.
  • Issue triage rubric: Complaints and reviews from all channels are categorized — refund promised, quality complaint, staff issue, delivery failure — with anything involving money or a public reply flagged for the owner personally.
  • Stock signal extraction: Weekend sales velocity against last-known stock produces a "likely to stock out this week" list per store, plus cross-store imbalances where one outlet's overstock covers another's shortage.
  • The Monday digest: One document, exceptions first: unreconciled amounts with evidence, customer issues needing the owner, stock actions suggested — each item linking back to its source data.

The deliverable

What the owner receives

  • Working pulse pipeline — ingestion through digest, running on the chain's real exports, with setup and operating instructions a store accountant can follow.
  • Reconciliation rulebook — the matching logic and exception thresholds as a written, owner-amendable document, not code nobody can read.
  • Triage rubric — the categorization and escalation rules for customer issues, including what always goes to the owner.
  • Four live digests — produced from real weekends during the engagement, so the format is proven on the chain's own data before handover.
  • Operating guide — who runs it, the weekly time cost, how to adjust thresholds, and what the workflow deliberately never does (send money, message customers, place orders).

What changes after

Monday starts with answers

The four-hour reconstruction becomes a twenty-minute review of exceptions. Settlement shortfalls and cash mismatches surface Monday morning while the weekend is still fresh enough to investigate.

Refund promises stop disappearing into store WhatsApp threads, and the stock-out list arrives before the stock-out.

What to measure

Evidence after a real baseline

  • Owner time from opening the digest to decisions made
  • Reconciliation exceptions caught, and days-to-detection
  • Customer issues reaching the owner within 24 hours of the weekend
  • Stock-outs on top-20 SKUs per store per month

Boundary

What stays human-owned

The pulse reads exports and drafts a digest — it never moves money, adjusts ledgers, replies to customers, or places purchase orders. Accounting judgments and disputes with payment providers remain with the owner and accountant. This is an operating digest, not bookkeeping, and it does not replace the chain's accounting.