Case study
CAC fell 20% the year they stopped buying tools.
A good team with a number moving the wrong way. Paid media kept getting more expensive: Meta CPMs at all-time highs, Google Shopping CPCs up a third in a year. The known way out is to lean less on paid: owned channels, faster creative iteration. This team couldn’t. Campaigns took three weeks to ship and the best creative died in six days.
So they did what buried teams do: bought more tools, hired more agencies. The stack got heavier, the production got faster, and the thinking stayed stuck behind the making.
The antagonist · CAC, indexed, 24 months
Every quarter, the same conversation: spend more, or grow slower. Hold this chart in mind. It comes back at chapter 06.
A Practical Intelligence AI strategist sat inside the org for four weeks and traced campaigns from brief to launch to reporting. Four findings made the file.
Nobody trusts the numbers.
Every platform claims the same conversions, over-counting revenue by 30 to 100%. Reporting is assembled by hand, and budget gets allocated on whichever number survives the meeting.
Paying for tools nobody uses.
Ten tools, roughly half the licenses idle, about a third of the capability surface in use. The personalization add-on: last opened five months ago. Still billing.
Creative can't keep up.
Brief to launch: 15 working days. Their best ad fatigued in six. The team ships a handful of variants a month; winning at today’s CPMs takes ten times that.
More demanded, less given.
Each marketer loses ~10 hours a week to versioning, reformatting, and reporting, while ~20% of the budget leaks to agencies for production the team could own. Leadership wants more output. The team is flat.
They didn’t need an eleventh tool.
They needed the operating layer the other ten were promised to be.
A content factory, an alignment layer,
and their marketing agent on top.
THEIR EXISTING STACK · CONNECTED, NOT REPLACED
Spend and revenue pulled from every platform, deduped against Shopify orders, maintained as the one blended view the board can trust. The weekly report writes itself and lands in Slack before Monday standup. Two ad sets running below blended breakeven get flagged, not defended.
Budget now moves on truth. The ops manager's report-assembly week is gone.
The alignment layer gave the keepers shared context, and made the rest redundant. The idle and overlapping tools were retired, one by one, as the agent absorbed their jobs. Most AI vendors add to the stack. This subtracted from it.
~$1.2M a year off the software bill. Itemized in chapter 06.
An approved brief becomes channel-ready everything in a morning: ad variants, email flows, product copy, on-brand from Figma, paused for approval. The watch never stops: fatigue gets caught in days and fresh variants are queued before the winner dies.
Brief to launch in 4 days, not 3 weeks. ~3x the output, agency bench in-housed, variants rotated before fatigue.
Monday starts with a ranked shortlist of campaign concepts, grounded in the brand’s own blended numbers, briefs attached. A marketer sharpens the angle and hits approve. Judgment stays with the team; everything downstream is the agent’s job.
Same headcount. The ~10 hours a week each marketer lost to production now goes to strategy and testing.
Smarter every week
Every approval, every edit, every result feeds back in. The agent learns their brand voice, their winners, their margins.
~$5.3M a year at run-rate. $2.3M you can point to on an invoice. ~$3M back from the media budget as CAC fell.
That’s the return on the saving: every point of CAC on a media budget this size is real money, every year, and it scales with spend. The board picks which way to take it: margin, or growth.
Cumulative value · first 24 months
Orange: invoice-backed savings (retired software + in-housed production). Periwinkle: media efficiency from CAC. The slope steepens as the system reaches full coverage and keeps learning. It paid for itself inside year one.
The antagonist, ended · CAC indexed
Chapter 01's line, continued. Two years of climbing, ended.
Campaign cycle
from ~15 business days
MER
Email share of DTC
Content output
Hours back / marketer
How we count it. Conservative where it counts. The invoice line includes only retired software and reduced agency spend you can point to. The media line applies the CAC gain to last year’s acquisition volume on a ~$15M paid budget; it scales with spend, and we discount year one for ramp. Redeployed hours and lifecycle revenue margin are left out of the number entirely. We show the full model and expect you to stress-test it with your own media math.
Close of file Nº 002
Open a file on your marketing org.
The Audit is fixed-scope and fixed-fee. We come to you, sit with your operators, and leave you a plan to put AI into production against the number your board already watches.