[Attached: Screen recording: Demo account with made-up data] Built in Claude Code with a running feedback log. Still a work in progress: we keep building and monitoring as new MCP and template updates land. WHO WE ARE - A coaching business with three revenue lines: high-ticket coaching sold on calls (HighLevel), digital courses (SamCart) and a physical store (Shopify). - Traffic is mostly organic (YouTube, email, Instagram), plus Facebook ads. - The questions we needed answered: which content makes money, is coaching converting, what is a customer worth, where should we spend, and can we trust the numbers? - This entry has two parts: the dashboard I built on HYROS Vibe, and the MCP process that decided what got built. PART 1: THE VIBE DASHBOARD WHAT WE BUILT (on top of the Vibe template) - 9 tabs, plus a data health strip that’s always visible. - One shared control bar: Dates (presets or a custom range), Compare, Attribution (first/last touch), Channel (paid/organic/no source), Business line (coaching/digital/physical) and saved views. - Overview: the whole business by day, rebuilt from CRM records and ad spend. It shows wins (new coaching clients, large sales), top movers and spike days, with notes you can pin to dates. - Ads: spend, revenue, ROAS, cost per lead and cost per call by ad account and ad set. Any date range is pulled on demand. - Records: leads, sales and calls, filterable, and beyond the 30-day window when you need it. - Content: every organic source (each video, email, post, ManyChat flow) ranked by the revenue it’s credited with, split by business line, with leads and coaching calls. - Coaching: - calls booked → shown → new coaching clients - close rate per person and time to close - what books the calls - HYROS lead stages - new this week: stats per HighLevel calendar, covering calls, cancellations, no-shows, show rate, close rate and revenue - Customer value: value per customer by first source at 7/30/60 days, repeat rate, digital → coaching upgrades, what they buy next, and first-purchase cohorts. It’s computed from sale records, because the LTV fields need forecasting enabled on the account and ours doesn’t have it. - Customer patterns: touches before buying, time from first click, the most common channel and content paths, which sources assist vs close, and where leads stall (from conversion paths). - Scale Advisor: Grow / Fix / Leave alone / Not enough data yet, for every ad account, top ad set and organic source.