Skip to main content

Build Platform

Where does the data live, what does the monitoring gap cost, and what unblocks AI?

Crackerjack runs on 8 systems. None of them talk to each other. Every week someone manually stitches the numbers together in a spreadsheet — and the Monday decision lands on Wednesday.

The Warehouse Group is 18–24 months ahead. The window is closing. This is not a technology problem — it is a data flow problem that technology can solve in 90 days.

7

Hand-offs per weekly report

2–4

Days from Monday question to answer

6 / 6

Critical data types with no SSOT

§1

Tech Stack Inventory

The data is in seven systems and the buyer judgement is in a head. No active API connections between any tools observable in the public surface. POS will not export structured data on demand. Manual export is the lived reality. There is no single source of truth across the merchandising loop.

Why this matters: any system you build deeply on top of a tool rated "Replace" is wasted investment. Build on what will survive the next 24 months.

POS System

POS / Transaction

INFERRED LOWStable

Build

E-commerce / CMS

Web / Commerce

WEB HIGHMonitor

Build Lightly

Catalogue Platform

Marketing / Catalogue

WEB HIGHReplace

Do Not Build

ERP / Merchandising

Backend / ERP

INFERRED LOWStable

Build

Email Service Provider

Marketing / Comms

INFERREDStable

Build

Google reCAPTCHA

Security

WEB HIGHStable

Stable

Google Maps

Location

WEB HIGHStable

Stable

Spreadsheet BI Layer

Manual Reporting

INFERRED LOWReplace

Do Not Build

Tool count: 8 [INFERRED ±2]. Sprawl signal: YES — major workflows touch ≥3 platforms with no observable API connections; reconciliation is manual.

→ Pattern →
§2

Business Analysis Pipeline

The Business Analysis Weekly Report has the highest hop count — 7 hops, 5 artifact. This is the primary flow constraint. POS and E-com data reconciled manually by Finance through spreadsheet — 7–13 hours active time, 2–4 days clock time. The Monday number lands Wednesday.

Why hand-offs matter: every time data moves between systems without an API, a human carries it. Each carry is a hand-off where errors enter, time disappears, and Monday's number becomes Wednesday's number.

Business Analysis Weekly Report — Primary Flow Constraint

7–13 hrs active time · 2–4 days clock time · number lands Wednesday

CSVexportqueryCSVexportshareverbalsavePOS Export1–2 hrsBOTTLENECKFinance Spreadsheet3–6 hrsE-com Export0.5–1 hrMerged View1–2 hrsSummary Email0.5 hrCFO Decision1–2 hrsSSOT Driftlow disciplinePOSERPE-comCatalogueBI SubstrateMonday 8am Report

Three workflows mapped Create / Manipulate / Share / Delete. Hops removed is the metric that matters per workflow.

Workflow 1 — Catalogue saleId production + publication

  • Step 1 Create — saleId planning sheet (Marketing + Buying human, 4-8 hrs)
  • Step 2 Manipulate — ERP/Merchandising landed-cost confirm (Buying human, 2-4 hrs)
  • Step 3 Share — Spreadsheet → designer brief (Marketing human, 2-3 hrs)
  • Step 4 Manipulate — SKUs loaded into Catalogue Platform with saleId metadata (Marketing human, 4-8 hrs)
  • Step 5 Manipulate — Price changes pushed to POS + e-commerce (manual export/load, 3-6 hrs)
  • Step 6 Share — Catalogue announcement to email club + social (Marketing, 1-2 hrs)
  • Step 7 Delete — saleId expires (auto)
  • Hop count: 7 | Artifact hops: 4 | Real hops: 3 | Clock time: 3 working days | Active person-time: 16-31 hours

Workflow 2 — Business Analysis Weekly Report (PRIMARY)

  • Step 1 Create — POS export (CSV) by Operations/IT (manual export, 1-2 hrs)
  • Step 2 Manipulate — Spreadsheet pivot/reconcile/annotate by Finance (3-6 hrs)
  • Step 3 Create — E-commerce export (CSV) by E-com/IT (0.5-1 hr)
  • Step 4 Manipulate — Online + in-store merged into one view by Finance (1-2 hrs)
  • Step 5 Share — Spreadsheet → email to CFO + Owner + Buying Lead (Finance, 0.5 hr)
  • Step 6 Manipulate — Decisions taken / next-week priorities set (verbal, 1-2 hrs)
  • Step 7 Delete — Spreadsheet saved to share (low discipline; SSOT drift)
  • Hop count: 7 | Artifact hops: 5 | Real hops: 2 | Clock time: 2-4 working days | Active person-time: 7-13 hours

Workflow 3 — Store-Level Inventory Replenishment

  • Step 1 Create — POS sales accumulating (system, continuous)
  • Step 2 Manipulate — Store Manager reads sell-through and identifies stockouts (1-2 hrs/day per store)
  • Step 3 Share — Phone/email/spreadsheet to Buying Lead (0.5 hr/day)
  • Step 4 Delete — Often absent; discontinued SKUs persist in ERP (drift risk)
  • Hop count: 4 | Artifact hops: 2 | Real hops: 2 | Clock time: continuous daily | Active person-time: 1.5-2.5 hrs/day × 15 stores
→ Pattern →
§3

Hop Count Summary

Catalogue Production

7 hops

4 artifact · 3 real · 3 working days

Business Analysis ← PRIMARY

7 hops

5 artifact · 2 real · 2–4 days

Replenishment

4 hops

2 artifact · 2 real · daily

Highest hop count workflow: Business Analysis Weekly Report (7 total, 5 artifact).
Primary flow constraint: the spreadsheet reconciliation layer between POS, e-commerce, catalogue, and supplier cost — the named Data Siloes gap.

Why this matters: every workflow improvement becomes easier once the highest-count constraint is gone. Fixing the wrong one first buys you nothing.

→ Pattern →
§4

Longevity Risk Assessment

Stable / Monitor / Replace — and build decision

  • POS System: Stable — modern AI-native challengers exist but switching costs are high → Build (deepen integration if API access available)
  • E-commerce / CMS: Monitor — AI-native commerce stacks accelerating → Build Lightly
  • Catalogue Platform: Replace — Catalogue-as-static-PDF is being disrupted → Do Not Build (treat as transitional)
  • ERP / Merchandising: Stable — mid-market retail ERP is slow-moving category → Build (assume long-lived)
  • Email Service Provider: Stable — AI features layered on; deepen segmentation → Build
  • Spreadsheet (BI layer): Replace — the named flow constraint; replaced by unified BI substrate → Do Not Build

Replace-rated tools: Catalogue Platform (category direction), Spreadsheet BI layer (the named constraint). Neither should receive deep integration investment.

→ Pattern →
§5

Flow State Assessment

Speed × Openness

  • Speed (real-time vs batch sync): Slow — weekly Business Analysis cycle named as Data-Siloes problem; daily store replenishment is verbal-channel-based
  • Openness (API access / data portability): Locked — no public API surface observable; POS export likely manual or batch; CSV-driven reconciliation
  • Position: Walled Garden quadrant — the opportunity is to move toward Flow State via the unified BI substrate that bridges existing siloed tools without forcing full replacement
Real-time ←
→ Walled Garden
→ Pattern →
§6

Single Source of Truth Audit

6 of 6 critical data types lack SSOT today

  • Customer/loyalty records — ESP (email club) + POS (transactional) — NO SSOT — email club has no transactional join key
  • Revenue / invoicing — POS + E-commerce + Finance/Spreadsheet — NO SSOT — three sources, manual weekly reconciliation
  • Inventory / stock — POS + ERP + E-commerce — NO SSOT — timing-of-truth differs across sources
  • Supplier / landed cost — ERP + Spreadsheet (buyer deal sheets) — NO SSOT — likely in buyer-owned spreadsheets, not normalized in ERP
  • Catalogue performance attribution — Catalogue Platform + POS + E-commerce — NO SSOT — requires manual cross-reference of saleId → SKU sales → store + online
  • FX exposure / hedge cover — Spreadsheet (CFO) — NO SSOT — open orders × supplier currency × forward delivery dates spreadsheet-managed

SSOT gap count: 6. Most expensive gap: Catalogue Performance Attribution — highest decision-loop velocity (weekly/fortnightly).

Why this matters to you: every decision made on conflicting data is a decision made blind. You are making six types of decisions this way every week.

→ Pattern →
§7

Primary Flow Constraint

Business Analysis Weekly Report — the spreadsheet reconciliation layer between POS, e-commerce, catalogue, and supplier cost is the primary flow constraint. Every Monday-decision the CFO + Owner + Buying Lead make depends on this 2–4 day reconciliation cycle that the input file names as "Data Siloes preventing Accurate Timely Insights."

Hop count: 7 | Artifact hops: 5 | Clock time: 2–4 days → estimated active time: 7–13 hours per cycle. This is the first flow constraint to address. Every other workflow becomes possible once this is replaced.

First action — this week

Request POS API access — in writing.

Without an API connection, the Finance spreadsheet stays and so does the 7-hand-off cycle. With it, the weekly report compresses to a dashboard that refreshes automatically. The CFO gets Monday's number on Monday.

Kill switch: if the POS vendor refuses API access in writing within 30 days — that is the kill-switch signal. Name it now, before any investment is made.

7→1

hand-offs removed

~10 hrs

person-time saved weekly

2–4 days

decision lag eliminated

~520 hrs

freed per year

Then, in order

  1. Connect E-commerce and ERP for unified stock view. Eliminates daily store phone calls; reduces stockout drift across 15 locations.
  2. Link the email club to transaction history. Turns one-way marketing into a two-way loyalty loop. Required for any AI personalisation.
  3. Replace the Catalogue Platform with AI-native tooling. The market is moving. Build nothing on the current platform — migrate when AI-native alternatives land.
→ Pattern →
§8

Dig Deeper — Blueprint Instruments

Four blueprint instruments anchor the Tech & Data lens. Each is a deeper pattern the Tech Lead or Owner-as-builder can read to navigate the data substrate decision:

→ Pattern →