Archived Engagement Document

Platform Roadmap & Data Strategy

How the platform is built. Data quality audit, data capture strategy across three phases, and the two hosting architecture options. For findings see the PoC Analysis; for scope, timeline, and commercials see the Scope of Work.

Prepared for Liam Vasey, Co-Founder WeBuyVintage April 2026 Archived — superseded by current scope

1. Where We Are Now

We've completed a full audit of your Google Sheets data. The raw material is rich; the structure needs work. Here's the shape of it.

59
Google Sheets tabs ingested
289
Columns in main events sheet
2,312
Events loaded (Apr '23 – Apr '26)
14,725
Individual SKU items tracked
22,107
Downstream sale records
51
Dealers across 17 regions
1,470
Unique venues with postcodes
3 yrs
Event history
Data quality assessment: rich but fragmented. Information is duplicated across multiple tabs, dealer names and initials are inconsistent between sheets, there's no relational structure linking events to items to sales, and many tabs serve as presentation dashboards rather than clean data tables. All fixable, and we've already normalised it into a structured database for analysis.
The PoC confirmed that your data supports all five models. Full prototype outputs with charts, dealer rankings, and white-space opportunities are in the PoC Analysis. This document focuses on the technical how: data capture strategy and hosting architecture.

2. Better Data In, Better Decisions Out

The models above were built on your existing data. Small changes to how data is captured will dramatically improve their accuracy and unlock new capabilities.

Phase 1 — Weeks 1–4

Quick Wins

  • Standardise Jotform submissions. Enforce required fields, add postcode validation, replace free-text dealer names with a dropdown. Prevents initials collisions and ensures clean linking between events, dealers, and sales.
  • Capture customer postcodes. Even the outward code is enough. Unlocks demand heatmapping: see where your sellers are actually coming from, not just where the venue is.
  • Standardise item categories. Add a category dropdown to the SKU jotform (Gold, Silver, Watches, Cameras, Medals, Coins, Jewellery, Other). Currently inferred from free-text titles; a dropdown gives clean data instantly.
  • Add footfall counter. Simple tally of total walk-ins vs those who sell something. Gives a conversion rate per venue, a powerful signal for the pattern analysis model.
Phase 2 — Weeks 5–10

Process Integration

  • Structured metal tracking. Replace manual gold/silver tracking with a structured form linked to event number. Eliminates the reconciliation problem between the events sheet and gold_sales/silver_sales tabs.
  • Dealer self-reporting. End-of-day form: items bought, estimated values, categories. Gives per-dealer item-level data, currently only available at event level.
  • Sale outcome linking. When items sell on eBay, Fellows, or trade, link the sale back to the original SKU. Closes the loop from purchase to resale, the missing piece for true margin analysis.
  • Venue feedback loop. Automated post-event venue rating form. Builds the venue scoring database that drives Models 1 and 5.
Phase 3 — Weeks 11–14

Platform Integration

  • Google Sheets sync. Automated daily pull from existing sheets into the platform database. No manual exports. Team keeps working in Sheets as they do now.
  • Webapp pipeline handover. The rota tool currently reads from Google Sheets. When the internal webapp (replacing Sheets) is ready, coordinate with the webapp team to switch the ingestion layer: map our input fields (rota, dealers, drivetime, holidays, tracker, May events) to the new API endpoints, agree field names and types, and decide who owns the translation layer. Sign-off needed both sides before cutover to avoid silent data gaps.
  • Jotform webhooks. New submissions flow directly into the platform database in real time. Models update overnight on fresh data.
  • Dashboard access. Ops team gets web-based access to all model outputs: event calendar, dealer leaderboard, area explorer, capacity dashboard.
  • Automated alerting. Weekly email digest: areas ready for return visits, dealer performance flags, emerging white-space opportunities.

3. How It Works

Two hosting options. Both are fully cloud-hosted — WBV needs nothing installed locally. Access is via any web browser, mobile-friendly for dealers in the field.

Recommended

Option A: Simple MVP

Your team keeps using Sheets exactly as they do now. Data flows automatically into the platform.

Google Sheets
↓
Automated sync (daily)
↓
PostgreSQL database
↓
Analytics models (nightly)
↓
Web dashboard

£50–100 / month hosting.

Option B: Full Platform

Custom data entry replaces some Jotform usage. Real-time updates, not daily batch.

Jotform + custom forms
↓
Webhooks + API layer
↓
PostgreSQL database
↓
Analytics models (real-time)
↓
Web dashboard + alerts

£150–300 / month hosting.

WBV needs nothing installed locally. Everything is cloud-hosted. Access via web browser on any device. Mobile-friendly for dealers in the field.

This roadmap covers the technical delivery approach only. Timeline, commercials, and next steps are in the Proposed Scope. Proof-of-concept findings are in the PoC Analysis.