Projects / Mosaic nav IA

Mosaic nav IA

Logged in: Northside Grill · path Covers · platform Google · win metric covers / visits

Not a toggle. Set once at onboarding. A plumber sees Leads here; a Shopify shop sees Shopify. Same app, different chip — not three sidebar versions.

Sidebar below is what this one account sees. No Leads playbook / Covers playbook / Shopify playbook rows — just Follow-up, which means covers for this restaurant.

Actuation

Alpha

How the four sections split

Where the path shows: the orange chip under the customer name. Not a toggle, not three sidebar items. This account is Covers, so Follow-up means visits and Forecast means covers. Leads and Shopify are other customers, not other menus.

Optimizing is actuation — spend, pacing, tracking, follow-up, write-back. The weekly task they pay to remove. Dead creative does not get invented here; it opens a creative brief.

Forecasting is pacing’s sibling, not a fourth dashboard. Pacing asks whether this month’s spend is on track. Forecasting asks how many jobs, covers, or orders the next 7 and 30 days will produce — and whether that exceeds owner-set capacity. Trailing-rate math with a high/low band. Hide the line if tracking is junk. Never auto-scale spend because the number looks good.

Planning / Creative is brief-driven. Kickoff locks the pitch brief and the marketing calendar (prix fixe weeks, hail, drops). The weekly brief then fires from that calendar plus live pacing/forecast — it does not invent the week. Creative briefs still open when a flight hits or Optimizing flags dead creative. Studio executes an approved brief.

Publish is GitHub first, then skyesmith.dev. The site is a deploy target, not a CMS. No GitHub SHA, no live page. Forecast math ships on the methodology page at launch — no predicted-ROAS claims.

Object	When it exists	Who writes it / how it ships
Account path	Onboarding, once	Niche router: leads, covers, or Shopify. Stored on the account, shown under the customer name. Not a UI toggle. Wrong path is a settings fix.
Pitch brief	Once per program / when the offer changes	Human or AI draft → human approve. Holds budget band, owner-set capacity, and the marketing calendar.
Marketing calendar	Kickoff, with the pitch brief	Flights decided before the first weekly brief. Prix fixe, hail, tune-up, drop. Feeds planned spend into the forecast. Mid-flight edits are exceptions, not a weekly planning meeting.
Weekly brief	Fires on schedule from the calendar	Pacing + 7/30-day forecast vs capacity + this week’s calendar row + recs to approve. Does not decide the program. Hide forecast if tracking is junk.
Forecast line	Inside the weekly brief	Trailing 28-day jobs/covers/orders rate × remaining/planned spend (including calendar bursts), high/low band, last week’s miss, one sentence of why.
Capacity	Pitch brief, closed beta	Owner sets N jobs / covers / orders per week. Forecast > cap → recommend pause or lower budget, approve-then-apply. Never infer from Jobber/Housecall Pro.
Creative brief — calendar	Flight already on the marketing calendar	AI can draft from pitch + calendar; human can write it. Calendar flights feed planned spend into the forecast.
Creative brief — system	Optimizing flags dead RSA, fatigued Meta, dead Spark	AI drafts from performance + pitch; human can override
Public post (skyesmith.dev)	Approved case or program writeup	Commit/PR to GitHub → deploy. No GitHub, no site. Methodology page explains forecast math; no predicted ROAS.