The First Month Isn't the Campaign

Most agencies start spending on day one, because spending is the thing that looks like work. The problem is that nobody has established what a lead is worth to you yet — so when the first month's numbers come in, there is no honest way to say whether they were good.

A $60 lead is a bargain for one business and a slow bleed for another. Same number, opposite meaning. Until you know which one you are, every decision after that is a guess you are paying for.

So the first month is not the campaign. It is finding out what your numbers actually are, with instruments built for your business specifically. Then the campaign is the thing that scales what we already proved.

Step one: what a job actually costs you

Not what you charge. What it costs. Almost nobody has this number to the dollar, and the ones who think they do are usually wrong in the same specific way.

Not because the owner was careless — no accounting package computes this, and every operator in the trade runs on the same rule of thumb until somebody builds the thing that doesn't.

For a junk removal client I built a job estimator that takes load size and density, drive distance and average speed, crew size and hourly wage, disposal tonnage and facility minimums, stairs and flights, difficulty multipliers, and the overhead the business needs to carry. It builds the cost up from those, then sets price from a target margin.

The estimator also carries a 1.30× labor burden by default, because a $25/hour employee does not cost $25 an hour. Payroll taxes, insurance, and unbilled time are real and they land on every job whether or not anyone accounts for them. Every default in the tool is an industry figure that gets replaced with the client's real number the moment we have it — the defaults exist so the thing is useful on day one, not so anyone relies on them.

The output is the number the whole engagement rests on: the profit left over after a job is actually done.

Step two: what a lead is worth

Once we know what a job leaves behind, a lead's value is arithmetic. Job profit multiplied by the share of leads that turn into jobs. That is it.

It sounds obvious written down, and yet the number of businesses spending money on ads without it is close to all of them. Work yours out — it takes about thirty seconds and you probably know all three inputs off the top of your head.

The number that comes out is a ceiling, not a target. Pay less than it and you make money on every lead. Pay more and you are buying revenue at a loss, which is a thing you can do deliberately and a thing you should never do accidentally.

Step three: what you can afford to spend

Now the ad question becomes answerable instead of philosophical. There is a number you cannot exceed per lead, and a minimum monthly budget below which the platform never gathers enough data to optimize at all.

Both matter and they are different constraints. Plenty of businesses fail on the second one — the cost per lead was fine, there just were not enough of them for the system to learn anything, so it never got better and the account was judged a failure at week three.

There is a budget and learning-phase check on the Meta page that works out whether the numbers you have can support a campaign at all. Sometimes the honest output is no, and that is a useful answer that costs nothing.

Step four: what is worth scaling

By this point we have a ceiling, a floor, and a live account. Now the job is finding the one thing that beats the ceiling — a service, an audience, a specific offer, a particular search term — and putting the budget there instead of spreading it thin across everything.

This is the part people imagine is the whole job. It is maybe the last third of it, and it only works because the first three steps produced a standard to measure against. Scaling without that standard is just spending faster.

Why the tools get built in an afternoon

The reason I can hand a client a costing tool built around their actual business, rather than a spreadsheet template, is that the system I work in already knows them.

Their pricing, their service area, their trucks, their disposal facilities and what those charge, what they have already tried, what the last campaign did. That context is written down and searchable, so building the tool is mostly a matter of deciding what it should compute. I am not reconstructing the business from memory while also trying to get the arithmetic right — which is exactly the point at which these things normally get abandoned half-finished.

A concrete one, from a different client. Their homepage was shipping 23 megabytes to every visitor — the hero video streaming in roughly 430 separate requests and never stopping — on a store where most of the ad spend is mobile. I found it by pointing my own tooling at the live site, re-encoded the video that afternoon, and it shipped the same day with the owner signing off on the quality tradeoff first. 23,134 KB down to 5,418 KB. About 430 requests down to one.

I can still quote those numbers because the audit that found them is a document in a system, not something I half-remember. That is the whole advantage. It is not that I am fast — it is that I never start from nothing, and the second tool for a client is always quicker than the first.

And then it all lands in one place

None of the above is worth much if the numbers live in four dashboards and a spreadsheet nobody opens. Every figure — leads, source, cost per lead, close rate, the job value they turned into — gets plumbed into your CRM, so the reporting is just your own system telling you the truth in the place you already look.

No monthly PDF. No screenshot of a graph with the axis cropped. You can check it on a Tuesday at 6am without asking me anything.

If you already know your cost per job, your close rate, and what you can afford to pay for a lead, you do not need month one and I will say so. Most people do not know all three, and the ones who find out usually change something about how they price before we ever run an ad.