The Infrastructure
Growth OS is the measurement and operations infrastructure we install inside client brands. It runs from the tracking container on your domain, through every sales channel and ops system, to what a customer costs and what a customer returns. One reconciled system, live in days.
Every channel on the same net-revenue basis
The executive scorecard, as your team sees it. Views shown with sample data; every client instance runs on its own warehouse and domain.
End to End
Server-side tracking on your own domain. Clean, deduped conversions reach Meta, Google, and TikTok.
Store, subscriptions, marketplaces, 3PL, and wholesale land in a warehouse you own. Ops and inventory sit next to revenue.
All paid spend on one blended standard: nCAC and aMER. Platform-reported numbers stay reference lines.
Revenue split new vs returning vs subscription. Cohort LTV measured from actual charges, back to CAC.
An executive scorecard, an AI-drafted weekly business review, and alerts when a number drifts.
The Layers
Most brands run these five jobs across a dozen disconnected tools, each reporting its own version of the truth. Growth OS runs them as one system with one set of definitions.
Server-side tracking on your domain.
sGTM container · first-party cookies · dedup + webhook fallback
A first-party tracking container sends clean, deduplicated conversions to Meta, Google, and TikTok. Ad blockers and Safari stop eating your signal, which means the platforms bid on data that is right. Subscription renewals get routed to their own event so they never masquerade as new purchases.
Ops and inventory in the same warehouse as revenue.
Shopify · Recharge · Amazon SP-API · TikTok Shop · Meta · Google · GA4 · Klaviyo · 3PL
Your store, subscription platform, marketplaces, 3PL, and wholesale feeds all land in a Postgres warehouse you own. Days of cover, reorder points, and exception queues come from the same tables your finance numbers do, so the ops meeting and the growth meeting finally use the same math.
Every ad dollar on one blended standard.
nCAC · aMER · MER · self-anchored zones, re-anchored quarterly
Each platform grades its own homework, so its dashboard renders as a reference line under a blended standard: nCAC, what a new customer costs across all paid spend, and aMER, what new-customer revenue returns per dollar. Spend and revenue share one time window, anchored to the latest complete day.
Revenue split into new, returning, and subscription.
Cohort CM-LTV · payback month · MRR waterfall · churn identity
A new customer is a first lifetime order, computed from all-time history. Cohorts show the contribution margin a customer actually returns, month by month after acquisition, measured from real charges. Your CAC ceiling comes from your own payback curves instead of an industry benchmark.
A weekly review that runs the loop.
Executive scorecard · WBR with watchlist · daily digest · drift alerts
The scorecard gives your whole team one set of numbers. An AI-drafted weekly business review reads them first, flags plan variance, and keeps a watchlist so a problem raised in week 3 still has an owner in week 7. Drift and anomaly alerts land in Slack before they land in your P&L.
Customer Economics
Repeat revenue hides in blended averages, and one-time buyers flatter your acquisition math. Growth OS separates them: acquisition cost for a first lifetime order, cohort curves for the margin that comes back, and a payback month you can plan cash around.
Because the curves come from your own charges, the CAC targets are yours too. Green means your cohorts pay back, measured against your own curves instead of an industry average.
Why Trust It
A dashboard is only useful if you stop double-checking it. Every source in Growth OS is reconciled against an independent witness before it renders, and blocking tests guard every refresh after that.
137,799 rows · 0 mismatches
No source reaches your scorecard until it agrees with an independent witness. When we cut one brand's Amazon reporting over from manual exports to the live API, the two sources matched on every one of 137,799 rows.
EMQ 8.7 / 10
We build server-side tracking to an event match quality floor of 8.0, the score Meta itself rates as great, then keep tuning. Our latest install runs at 8.7. In the same account we removed 215,000 fake conversions a month that Smart Bidding had been learning from.
34 blocking checks
Data tests run before every refresh. If a number drifts from its source of truth, the pipeline halts and tells us, instead of publishing a chart that looks fine and is wrong.
Ownership
The warehouse is yours. The history in it replays back to the source bytes. The scorecard runs on your domain, and your instance shares nothing with any other brand we serve, by architecture and by written commitment. If we ever part ways, the system keeps running without us.
Ways In
Growth OS installs in stages, and the first stage ships fast: manual exports light up a reconciled scorecard in the first week while platform credentials come through.
The server-side tracking stack on its own: audit, deploy, verify. Acceptance criteria you can check yourself, including a live test order and an event match quality floor. The fastest fix if ad performance slid after iOS 14 and never came back.
The warehouse, the connectors, the canonical metrics, the scorecard, and the alerting. Your numbers reconciled and live, with manual exports lighting up the dashboard in week one while API credentials come through.
Everything in Core, plus the weekly business review run with Ben: plan variance, watchlist accountability, and a standing hour where the numbers turn into decisions.
Inventory reorder engine, exception queues, wholesale EDI order processing, affiliate and partner programs. Added per brand, when the brand needs them.
Next Step
Bring your ad accounts and your gut feel about what is working. We'll show you where the two disagree, and what the infrastructure would look like inside your brand.