What do you optimise the account for?
The account is optimised for the lead that goes on to buy, not the cheapest lead the platform can find, because the CRM tells Meta which leads booked, showed and paid and Meta learns from that.
Cheap leads and good leads are often different people. An account that chases the first one fills the pipeline with names that never answer, and the cost per lead on the report looks like a win the whole time.
- The event Meta optimises on is chosen for how often it happens and how closely it tracks money
- Deeper outcomes stay as the scale gate, read by a person
- Lead quality is reported next to lead cost, never after it
Acquisition Network: paid acquisition on Meta and Google, managed daily, $2.9M from zero
How do you decide which ad lives?
Each ad is judged on the cheapest metric that can be read at the spend it has received so far, and it moves up to a deeper metric only once it has earned the budget to produce one.
At a small budget the readable signals are hook rate, click rate and cost per lead. Further up the ladder it is the cost of a completed application or a qualified lead. Only at full maturity does a booked call or a sale decide whether an ad is cut or scaled.
- A written ladder of metrics, one per spend level
- A written spend floor before any rare event can cut an ad
- The same ladder on every account, so decisions do not depend on the mood of the week
A documented creative-testing method with its rules written down and reused on every account
Where does the creative come from?
New creative is built from what has already converted in your account and in your market, so each batch changes the hook and keeps the body that produced customers.
Scripts are written as variations off winning patterns, each with a stated hypothesis about why it should work. The brief names the audience, the voice and the angle before anyone films, so a result can be traced back to a decision rather than a guess.
- A research pass on what is working in your category
- Briefs that name the hypothesis before the shoot
- Proven bodies reused with fresh hooks
Smart Sellers Academy: funnel architecture and qualification behind the ads, $1.5M
How do you keep the platform honest?
The CRM is the source of truth, and the ad account is checked against it every week, because the platform reports its own credit and your bank account reports what actually happened.
Every ad carries a code that survives into the CRM, so a booked call, a show and a sale can be tied back to the ad that began the journey. Where a lead arrives with no code, we fix the tracking rather than guess at the source.
- Ad codes carried from the click into the CRM record
- Results read in the CRM first and the platform second
- Conversion events sent back to Meta from your own system
Smart Sellers Academy decides against a live KPI warehouse every week
How much do you spend on testing versus winners?
The split follows how many proven winners the account has: a larger share goes to testing while winners are scarce, and a smaller share once several are carrying the account.
A protected testing line stops scaling from consuming the budget that finds the next winner. When a winner scales, blended cost falls, which supports a larger total budget and a larger testing pool, so each cycle funds the next.
- Testing is a separate line item that scaling cannot raid
- The share shrinks as proven winners accumulate
- Winners are reviewed for promotion on a schedule, not when someone remembers
Applied on every account from the documented method
How many ads do you launch at once?
Fewer than the budget can fund to a readable level, because launching forty ads at once starves all of them and nothing reaches a result anyone can trust.
Batch size sets how long a read takes, not how many reads the budget can buy. We launch in waves sized to the spend, screen cheaply at the start so weak ads leave early, and backfill the freed budget with the next ads from the backlog.
- Waves sized to what the budget can read
- A cheap early screen so duds exit fast
- A creative backlog that decouples filming from launching
Documented method, reused on every account
How do you know an ad has worn out?
An ad is retired on a rule, not a feeling: when its readable metric drifts past the floor set for that spend level, or when fresher creative from the same concept beats it.
Winners are concepts, not single ads, so when one tires the next move is a new execution of the same idea. That keeps the account from running one lucky video until it collapses, and from restarting from zero each time.
- A written retirement rule per metric
- Concept families tracked, not only individual ads
- Replacements built from the same winning angle
Part of the written creative-testing method
How do you tag the creative so patterns show up?
Every ad is tagged by audience, actor and copywriter, plus one dominant value for hook, angle, proof and offer, so results read as patterns instead of a pile of single ads.
With the audience and the person on camera tagged separately, a gap between two ads can be traced to the right cause. The tags sit in one library next to each ad's results, and the next round of scripts is written with that data open.
- Audience, actor and copywriter recorded for every ad
- One dominant value per tag, so patterns are readable
- Patterns trusted only once enough qualified outcomes exist
A creative library wired to live results, used to brief every new script
What do you refuse to do?
We refuse to cut an ad on a rare event before the spend exists to read it, to call a lead a win because it was cheap, or to report a number we cannot trace to your CRM.
We also refuse to keep the method private from your team. The rules are written, the ad account is yours, and anyone you hire later can pick up where we left off without starting from a blank page.
- No cuts on a rare event at tester budgets
- No reporting that cannot be reconciled with the CRM
- No account that only we can read
Zero ads cut on a rare event at tester budgets, by rule
