Built from what Matt and Lorraine have asked for in writing since March. Five minutes to read; the working mock sits behind it.
Matt opens one export each week and pastes the discovery numbers straight into his report. Lorraine sends Paul a monthly pack that reads like a story: found, chosen, ready, per estate, this year against last. Anyone at the brewery can pick it up cold and understand what NeameGraph is doing and why it matters. Nobody argues with a wall of dashboards again.
Build the Digital Discovery Report as its own reporting app, fed by NeameGraph. Three chapters, matching the three measurement lenses Lorraine already put in front of the exec board: Found (are we visible where people now look), Chosen (do they act: directions, calls, bookings) and Ready (is our data clean, current and machine-readable). Every period is selectable, every number carries its source, and the year-ago column is built in from day one.
Do not start from a blank page. Half the report already runs every night inside NeameGraph, and the app already has an outside door for feeds. The build is a reporting surface plus two new data feeds, not a new platform.
1. Discovery has already moved, and the current reports cannot see it
Over the last reported quarter the managed pubs' Google profiles were viewed two million times while their websites took four hundred thousand sessions. Customers are finding pubs on Maps and in AI answers, then walking in. A sessions and pageviews report measures the wrong door.
2,017,253 profile views against 405,323 sessions, Dec 2025 to Feb 2026. Source: Lorraine's Q1 exec deck, slide 3.
2. Half the report is already built and proven nightly
Every night NeameGraph walks the whole estate: it checks the pages, spots what changed, regenerates the structured data machines read, pushes it live and logs the fixes the web team should make. The Ready chapter is that engine reporting on itself; the period selector, comparison lines and narrative already exist in the app's Gazette screen. Matt's Scan, Compare, Report, History shape is already the skeleton in the code.
In one week: 19 schema updates deployed across 13 pubs, plus 26 corporate updates. Source: the weekly NeameGraph update, 7 May.
3. The missing half plugs in, it is not a research project
The Found and Chosen chapters need the view-and-tap numbers, how many people looked at a listing and asked for directions. Those live in Yext. NeameGraph already reads Yext every night, so they plug straight into the report. No new supplier, nothing to wait for. The site's own visit data is the one genuinely outside feed, and Shepherd Neame is already testing the route for it. Until each one is switched on, the report shows an example and marks it, rather than writing around the gap.
BigQuery is already "being tested to give us better reports" in SN's own words. Source: Q1 exec deck, slide 3.
The report never invents a number: every gap below is named on the page until its feed is wired. Each one has a known route in.
Already ours and live: the nightly change watch that catches every page change and what the engine did about it, schema coverage, listings accuracy through Yext, page health, and the fix list for the web team. We found the change watch had quietly stopped writing its record back in May, and we put it right, so it is logging again.
The mock is ready to show today and the engine half is already running. Shall we put it in front of Matt and Lorraine this week?