What we've learned.

When sales go quiet, the obvious move is to buy more visitors. It's the quickest lever to pull and the easiest to defend: spend goes up, traffic goes up, someone can point at a chart. It's also, in every job we've taken on so far, the wrong thing to do first — because in every one of them the real problem was what happened after the click, not before it.

Three examples, all of them businesses that came to us wanting more customers:

• Two dental practices were already advertising on Google, Yelp and Facebook, paying around $80 for every inquiry. The ads pointed at general pages rather than pages about the treatment being advertised, and their booking system recorded patients without telling the marketing anything.
• A geotechnical and civil engineering firm had no website at all, and nowhere to record an inquiry. Every client came by word of mouth — one to five a month, at about $15k in monthly revenue.
• An online store was losing more than 99 of every 100 people who visited, had no way of collecting email addresses, and was making about $1.1k a month.

In each case, spending more on day one would have bought more visits into something that was already losing most of the ones it had. The extra money would have made the top line look bigger for a while, and made each customer cost more, not less.

What we mean by "the leaks"

It's not a figure of speech about branding. It's the specific, checkable set of steps a stranger goes through between seeing an ad and paying you, and each one either holds or leaks:

• The page they land on. An ad about emergency dentistry should lead to a page about emergency dentistry, not a homepage slideshow.
• The bit where they get in touch. A form, a phone call, a booking — something that puts them somewhere you'll actually see them, rather than an inbox.
• The follow-up. Someone, or something, gets back to them on a set schedule.
• The bit that closes the loop. What happened to that person — booked, bought, wasted your time — makes it back to the ad platform, so the ads get better at finding more of the good ones.

Most businesses we look at have one or two of these. Very few have all four. The last one is the one that compounds: without it, your ads never learn what a good customer looks like, and you're paying to guess the same thing again every month.

Ads multiply whatever you already have. Most businesses are multiplying by a number smaller than one.

The order we do it in

Website and follow-up before ads. Store and email before paid. Then — and only then — spending, going up month by month for as long as the numbers hold, and stopping the moment they don't.

This is slower to start, and it's genuinely awkward in week one, because the client is paying us while nothing appears to be happening in the ad account. What it buys is that every extra dollar later lands on something that turns it into money. At the dental practices, that showed up as the cost of an inquiry falling 31% while the numbers were going up — the opposite of the usual trade-off. At the engineering firm it meant their first ads ran into a working business rather than a void.

How to know if this is you

Two questions. First: if you doubled your ad budget tomorrow, what would break — the phone, the diary, the follow-up, actually doing the work? If you can name it, that's the first job, not the ad account. Second: can you say what your last twenty inquiries cost you, and which of them turned into money? If not, you can't yet tell whether spending more is working, and buying more of it is just a guess with a bigger number on it.

What an inquiry costs is the number every local business owner can quote off the top of their head, and it's the one most often misread. It gets treated as an advertising number — something you haggle down with better keywords, better bidding, a better agency. It's really a whole-business number. It's what you spent, divided by how many people the rest of the business managed to hang on to, and it's that second half where the room to move is.

When we took on two dental practices, they were paying about $80 an inquiry and booking roughly 10 patients a month through Google. Ad budget wasn't the problem; they had it and were spending it. The problem was that the budget was being asked to do work nothing else in the business was doing.

Why this bites hardest locally

In a local business you can't outrun a bad cost per customer by doing more of it. There are only so many people within driving distance who need a dentist this quarter, so you can't just buy your way bigger and fix the margins later. What a patient costs you decides whether they're worth having at all. At $80 an inquiry, with a normal share of people who never book, growing meant paying more money for the same patients.

What an inquiry costs isn't what Google charges you. It's what the business wastes, divided by the people who survived it.

What we changed

• A page per campaign. Every campaign got a page built for exactly what it was advertising, instead of pointing paid visitors at a general page and hoping.
• A booking system that does something. They had one; it recorded patients. We tidied it up so every inquiry was captured, tracked and chased on a set schedule — the ones that used to disappear between the phone call and the appointment stopped disappearing.
• Closing the loop back to the ads. Once the ads could see which inquiries became patients, they stopped chasing cheap form fills and started finding people who actually turn up.
• Then more of it, across both practices. Google, Yelp and Facebook, run as one approach rather than two separate accounts.

What happened

Google bookings went from about 10 a month to 37 in the first three months. Over twelve months, patient visits doubled. The cost of an inquiry fell from $80 to $55 — down 31% — while the numbers were going up, which is the opposite of what usually happens when you spend more.

None of that came from clever bidding. It came from the second half of the sum: more of the people the ads were already reaching made it all the way to a chair. The same money, spent into something that leaked less, simply bought more patients.

The sum worth doing

Take what an inquiry costs you and the share of them that turn into customers, and work out what a customer actually costs. Then hold that against what a customer is worth over the years they stay with you. If that comparison is uncomfortable, the answer is almost never a cheaper click. It's the gap between the two — the people you already paid for and then lost.

Word of mouth is the best source of work in the world, right up until it's the only one. It arrives already trusting you, it costs nothing, and it closes fast. It also can't be scheduled, predicted, or turned up when you need it — which means a business running on it alone is running on something it doesn't control.

That was the position of a geotechnical and civil engineering firm we started working with in 2023, as its acting CMO and CTO. One to five clients a month, all referred, roughly $15k in monthly revenue. The number swung about month to month. Nothing existed to keep track of possible work. To anyone outside the owner's own network, the firm may as well not have existed. Every decision about growing — hiring, equipment, going after bigger jobs — was waiting on revenue that word of mouth couldn't reliably produce.

Why we didn't start with ads

The obvious move is Google Ads. People searching for engineering services are usually ready to buy, the searches are specific, and the jobs are worth real money. But advertising a business with no website and nowhere to put an inquiry doesn't produce clients; it produces an invoice. The click has nowhere to go, the inquiry has nowhere to land, and nobody can tell you afterwards which spending won which job.

Ads only pay off when the people who click have somewhere to land and someone to follow up. So that part has to exist first.

The order we built it in

• Somewhere to put an inquiry, first. Before there was any new demand to handle, there was a place to put it — every inquiry captured and tracked, including the referrals that had been living in someone's phone and someone's memory.
• Then the firm's first website, built around getting people to make contact rather than as a brochure. It needed to look credible, but the real point was that someone could get in touch and be recorded.
• Then how a job actually gets won, mapped out with their own team — how work gets quoted, negotiated and signed — so the marketing fed the real process rather than a theoretical one.
• Then Google Ads, the first advertising the firm had ever run, pointed at something that could handle it.

What it produced

The firm passed $50k in a single month for the first time in April 2024 — the first year we worked together. It held $55–80k a month through 2025, and is averaging around $100k a month in 2026. That's about 6.7 times the monthly revenue it started at, and a yearly run rate that went from roughly $180k to roughly $1.2M.

Just as important: revenue is no longer capped by who happens to know them. Advertising brings in work alongside referrals rather than replacing them, and what's underneath — site, inquiry tracking, sales process — can take more spending as the firm adds people to do the work.

The part that's hard to buy from a supplier

We sat inside the business as its acting CMO and CTO rather than outside it as a vendor, and that's why the pieces joined up: the ad, the form it led to, the record that created, and the conversation that won the job were all designed by the same people. An agency can run a brilliant ad and still deliver nothing, because the four things that happen after the click belong to somebody else.

There are two ways an online store's revenue goes up: more orders, or bigger ones. Growth plans almost always chase the first, because order numbers respond fastest to advertising. The trouble is that order numbers are also the easiest thing to buy at a loss. Push hard enough on discounts and you can grow orders forever while what people spend per order slides underneath you and the business gets less healthy every month.

When we came on as a growth partner to a Shopify store in late 2022, it was averaging about $1.1k a month, wasn't advertising, was losing more than 99 of every 100 visitors, and the line on the chart was flat. We were paid out of a share of the revenue and our own money bought the ads, which changed how the problem looked: we hadn't been hired to run campaigns, we were betting our own cash that a real business could be built here. Buying orders at a loss would have meant buying our own losses.

Groundwork before spending

So the first stretch wasn't advertising. We rebranded the company and rebuilt the store from scratch, then set up automatic emails to win over the visitors it already had. Only after that did the Facebook, Instagram and Google ads go live — paid for by us, and raised month by month as the numbers held up.

Sending paid visitors to a store that doesn't sell burns money. When it's your own money, you find that out quickly.

Both halves moved

By 2025 the store was averaging $51.0k a month — 46 times where it started. Yearly revenue went from $150.7k in 2023 to $612.4k in 2025, up 306%. Orders went from 1,363 a year to 4,940, about 3.6 times as many.

The number we watch just as closely: the average order went from $115.97 in 2024 to $123.96 in 2025, up 6.9%. That matters because it tells you whether the growth is real. Orders going up while the average order falls means you're buying volume with your margin. Both going up together means the store, what's on it, and the emails are doing work advertising can't — and it means each extra order is worth slightly more than the last, which is what makes spending more safe.

December 2025 was the store's first six-figure month at $104.2k, roughly 50 times its December 2022.

What we'd tell someone in the same seat

• Fix the store before you buy visitors. A store where fewer than 1 in 100 people buy doesn't have a traffic problem yet.
• Make the most of the people who already find you. Emails win over people who've already turned up, at a fraction of the cost of finding new ones. It's the cheapest revenue in the business, and it usually gets built last.
• Watch order numbers and average order on the same chart. If they pull apart, the growth is rented.
• Grow in steps you can undo. We raised the budget month by month, tied to the numbers holding — not to a plan set in advance.

Kocah works two ways, and which one fits comes up on nearly every first call. The honest answer depends less on your budget than on how far along the business already is.

Retainer: we join your team

You pay us monthly and we run the marketing as part of your team — in some cases as acting CMO or CTO, with a brief to fix the whole business rather than just run a channel. That's what the engineering firm bought: we set up their inquiry tracking, built their first website, worked out with their people how a job actually gets won, and launched their first advertising. It's what the dental practices bought: a page for every campaign, a booking system that actually helped, and the ads on top, run the same way across both.

This fits when the business has revenue and can handle more work, but has nothing reliably bringing customers in — and when someone needs to own the whole thing rather than a slice of it. It's also the right choice when there's a lot to build: a monthly fee pays for work that won't produce revenue for months, which a share of revenue can't.

Partnership: our money's in it too

We put our own money into the business — paying for the ads, the rebuild, the whole push — in exchange for a share of the revenue. That's how we worked with the online store: our cash paid for the Facebook and Google campaigns while the store was rebuilt and the emails went live.

This fits when the product already works but there's no engine to sell it and not much money to build one. It doesn't fit a business that just wants cheaper marketing — a share of revenue costs more than a retainer if it works, and that's rather the point.

In a partnership, being on the same side isn't a slogan; it's the mechanism. We only earn when you do, so there's no version of this where burning your budget pays us.

What doesn't change

The work. Fix the leaks before spending more, build the thing before advertising it, and watch what a customer costs you alongside how many you're getting. Across three industries and both ways of working, that's what produced the results — 2x to 46x growth over one to three years, with the cost of a customer getting better rather than worse as the numbers went up.

And the selection. We take on a handful of clients at a time either way, because both need us inside the business rather than next to it. That genuinely limits how many we can run at once, and it's why the form is an application rather than an order.

How to choose, in one question

Do you need someone to build and run the engine, or someone to fund and run it? If you've got the money and need the people, take the retainer. If the product's proven and cash is what's stopping you, the partnership is the conversation worth having.