Almost every piece of marketing math you have ever been handed quietly assumes the customer comes back. Lifetime value assumes it. Payback period assumes it. "Break even on the first order, make your money on the second" assumes it so completely that it has become a slogan.
Then there are the businesses where the customer never comes back, because there is nothing to come back for. You replaced the roof. It is done. You will not hear from them for twenty years, and by then they may have moved.
If that is your business, most of the advice aimed at you was written for someone else.
Every Number You Have Been Given Assumes a Second Sale
Think about who actually lives in this category. Roofing. Solar. Window replacement. HVAC system changeouts. Foundation repair. Capital equipment that runs for fifteen years. A custom tank fabricated for one plant. A pool table that gets handed down to the grandchildren. A wedding venue. A funeral home, where the second purchase is a conversation nobody wants to have.
These are not small businesses in the sense of small money. They are often the opposite: high ticket, high consideration, long sales cycles, real craftsmanship. But they share one structural fact that changes everything downstream. There is exactly one transaction, and then the relationship with that customer is essentially over.
Compare that to a store where your first sale barely breaks even and the second one is the profit. That business can deliberately lose money acquiring a customer, because it knows the customer will be back in six weeks, and again in five months, and the third order is where the margin lives. Its entire acquisition strategy is a bet on the future. That bet has real problems of its own, not least that lifetime value does not pay this month's invoice. But at least the bet is available.
You do not get to make that bet. Your future does not contain another order from this person.
What Actually Breaks When There Is No Second Sale
Three specific things break, and it is worth naming them separately, because owners tend to feel all three at once as a vague sense that the numbers are not working.
Your ceiling has nothing to borrow against. In a repeat business, what you can afford to pay for a customer is calculated across everything they will ever buy. In yours, it is calculated across one job. That is not a philosophical difference. It is a number that comes out several times smaller, and it means an acquisition cost that a competitor in a repeat category would shrug at can quietly put you underwater.
The platform's value signal has very little to learn from. Smart bidding gets better by watching outcomes. A store feeding it hundreds of purchases a month, with repeat buyers and rising order values, is handing it a rich picture. You might close eleven jobs a month, each one six weeks after the click that started it. The system is learning from a thin, slow trickle, and it will be confidently wrong for longer than you expect. This is worth setting expectations around before you start, not after, and it is a large part of how long it takes before ads actually work.
A bad month stays bad. Repeat businesses have a kind of shock absorber. Spend too much in March and the cohort you bought in March keeps paying you back through September. You can be wrong for a while and recover quietly. With one transaction per customer, March is closed when March closes. There is no back end coming to rescue it.
Work the Ceiling Backwards From One Job
The good news is that this math is knowable before you spend anything, and it is simpler than the repeat version because you are not forecasting a future. You are looking at one job.
Start with what a job is actually worth to you, which is not the invoice. It is gross profit: the invoice minus materials, minus labor, minus the subs, minus anything else that only exists because you took that job.
Say the roof is fourteen thousand dollars and your gross margin is thirty percent. That is forty two hundred dollars of gross profit. Now decide, deliberately, what you intend to keep from that job as actual profit. Say twenty two hundred. What is left, two thousand dollars, is the most you can pay to win that job and still make what you decided to make.
That two thousand is your ceiling per sold job, and it is the number almost nobody writes down.
Then work outward. If one in four estimates turns into a signed job, you need four booked estimates to produce one sale, so your ceiling per booked estimate is five hundred dollars. If roughly half the leads you get turn into an estimate that actually happens, you need eight leads, and your ceiling per lead is two hundred and fifty.
Those are illustrative numbers, not benchmarks. Your margin, your close rate and your show rate are yours, and the whole point is that you can only do this with your own. But the shape holds for everyone in this category: one job, worked backwards, gives you a hard number, and that number is a ceiling rather than a target. The full mechanics of turning that into a bid target are in why you cannot set a target CPA you have never calculated.
The Lead Is Not the Product. The Booked Estimate Is.
Here is where most accounts in this category go wrong, and it is an honest mistake.
You tell the platform that a form submission is a conversion. The platform, being obedient and literal, goes and finds you the cheapest form submissions available. Your cost per lead drops. Everyone is pleased. And your calendar fills with people comparing six quotes, people who wanted a number for an insurance claim, and people who will not be doing anything until next spring.
In a repeat business, a weak lead is an annoyance. In yours, it is the whole month, because you only had eleven chances and you spent four of them on estimates that were never going to close.
The fix is not to write cleverer ad copy. That lever is real, and we have covered it in what most Google Ads advice gets wrong about a local business. The deeper fix is to change what you report back to the platform as a conversion in the first place.
If a booked estimate is the thing that actually predicts revenue, then a booked estimate is what should be counted, not a form fill. That usually means sending the outcome back from your CRM or your scheduling system rather than letting the website be the last thing anyone measures. It is more setup than the default, and it is the difference between an algorithm optimizing toward people who fill in forms and one optimizing toward people who buy roofs.
Feed it the wrong definition of success and it will pursue that definition relentlessly. That is not the platform being difficult. That is it doing exactly what you asked.
Your Repeat Customer Is the Referral Chain
Since the homeowner is not coming back, the recurring relationship in your business has to live somewhere else. It lives with whoever touches that customer before you do.
The question worth sitting with is this: what do people buy immediately before they buy from you? For a roofer it might be a home inspection, a property purchase, a storm damage claim, or a conversation with a realtor trying to close a sale. For a fabricator it is an engineering firm specifying the job. For a commercial equipment seller it is the contractor already on site.
Each of those people meets your future customers over and over, on a schedule, forever. The inspector will walk two hundred roofs this year. Your customer will buy one.
That is the closest thing to lifetime value your business model offers, and it is usually under-built because it does not look like marketing. It looks like having lunch with three inspectors. But a referral relationship that sends you four jobs a year, year after year, is doing precisely the job that repeat purchase does in a store: it turns a one-time acquisition cost into a stream.
None of that means turning the ads off. Paid search is where the person who already has a leak goes, and that demand is worth capturing at a controlled price. It means the ads are one leg of the thing rather than the whole thing, and the other legs are cheaper and slower to build.
Reputation Is the Asset That Compounds, Not the Customer
A purchase somebody makes once a decade gets researched in a way a repeat purchase never does. Nobody reads nine reviews before reordering dog food. Everybody reads them before letting a crew take the roof off their house.
So the thing that accumulates value in this business is not a customer list. It is the public record: reviews under your name in more than one place, photographs of finished work that look like your work, warranty follow-through that people mention by name, a crew that shows up when it said it would.
That record lowers the cost of every job you win afterward, because it shortens the part of the sale where a stranger decides whether to trust you. In a repeat business that trust gets built by the product arriving and being fine. In yours it has to be built before anyone hands you a deposit.
What This Means for the Budget
Put all of it together and you get a budget that behaves differently from the one in the case studies.
It has to be self-funding, month by month. You cannot spend into a hole in the expectation of climbing out of it on the back of a cohort maturing, because there is no cohort and nothing matures. What you spent in March gets settled by what March produced.
It has to hold a ceiling rather than chase a volume target. When a platform offers you more jobs at a higher cost per job, a repeat business can sometimes say yes and mean it. You have to check the number you calculated earlier and be willing to say no, even when saying no feels like leaving growth on the table.
And it should concentrate. Splitting a modest budget across three channels to see what sticks is expensive everywhere, and it is worse here, because your conversion volume was already thin and you have now divided it into three piles too small for any system to learn from. One channel, funded properly, with a ceiling you actually enforce.
That is also why the comparison worth making is not your ads against your competitor's ads. It is your ad budget against everything else you could buy with it: another truck, another crew, a better showroom, the inspector lunches. In a business without a second sale, that comparison is the strategy.
The businesses that do well in this category are not the ones that found a cheaper click. They are the ones that worked out what a job could carry, refused to go past it, and spent the rest of their energy on the two things that do repeat: the people who send them work, and the reputation that makes saying yes easy.
We run paid search for businesses like this all day long, and the first thing we do on a new account is the arithmetic in this article, because everything downstream depends on it. If you would rather work it out yourself, everything above is what we would check first, and the numbers are all sitting in your own books.
