The cheapest growth isn't new revenue. It's the profit and the customers you already paid for.

TL;DR: Founders chase growth outward: new leads, new markets, new products. Meanwhile two piles of money they already earned sit uncollected inside the business. The first is margin: the blended number on the P&L hides which products and customers actually make money and which quietly lose it. The second is dormant relationships, the customers and leads you already paid full price to acquire and then stopped talking to. Both are cheaper than anything in your growth budget, because you paid for them once already. Most founders walk past both on their way to spend more on a stranger.

Every founder I talk to is trying to grow the same way. Outward. More leads, more customers, a new market, a new line. And the cost of all of it keeps climbing. Ad prices up, response down, sales cycles longer than they were two years ago.

Meanwhile the cheapest money in the business is already inside it, in two piles almost nobody counts. Not because the piles are hidden. Because collecting them is less exciting than chasing something new, and excitement is what tends to set the agenda.

Ask a founder what their margin is and you get one number. A blended gross or EBITDA (earnings before interest, taxes, depreciation, and amortization) margin, averaged across everything the company sells.

That number is the most comfortable lie in the business, because no customer buys the blend. They buy a specific product, on specific terms, through a specific channel, and every one of those carries its own margin. Some are excellent. Some are underwater. The average tells you the company is healthy while one line quietly bleeds and another quietly carries it.

There are three cuts that expose it, and most operators have never run any of them. By product: which lines actually earn, which leak, which just look busy. By customer: your largest account is often your worst-priced one, locked in years ago and never revisited. By channel: the same product sold three ways rarely keeps the same number of dollars.

And the line founders are least willing to examine is usually the one they are proudest of. The flagship. The product you are known for and would never discount, except you do, constantly, because it is the one customers negotiate hardest on. It is often the most support-heavy and the most expensive to deliver.

I have sat with founders whose signature offering was running at a contribution margin near zero, in some cases priced so low it could never have made money on its own, quietly subsidized by a boring line they barely marketed. And most of them kept it on purpose, because they believed they needed it to feed the more profitable products and services behind it.

They were not always wrong about that, which is why killing it is rarely the answer. Sometimes you reprice it. Sometimes you deliver it for less. And sometimes you leave the headline price alone and bolt an instant upsell onto it, an add-on with a take rate high enough to drag the whole transaction back over the profitability line, so the product that wins you the customer finally stops losing you money. But you cannot make any of those calls from a blended number.

The second pile is sitting in your own database. You already paid, in full, to acquire a long list of people who bought from you, trusted you, and then drifted off your radar. They usually sit in four piles of their own.

You watch CAC (customer acquisition cost) like a hawk. There is no line on the P&L for "revenue we already earned the right to and walked away from," so it never gets watched.

A new customer costs hundreds to thousands of dollars in marketing and sales effort to land. A dormant one costs an email, a call, and an offer built for someone who already knows you. Same revenue, a fraction of the cost, and it closes faster, because the trust was already bought. Once.

And the first move is almost never a campaign. It is a single, plain, personal email, the kind that looks like it came from your phone and not your marketing department. Dean Jackson's old nine-word version still beats anything fancier: you ask, in plain text, whether they are still trying to solve the exact problem they first came to you for, and nothing more.

You are not pitching. You are starting a conversation, and the people who reply have just raised their hand and told you they are still in pain. That is who your real offer goes to next. The heavier machinery can come later. The cash is in simply finding out who is still listening.

Neither pile announces itself. New acquisition has a budget line, a dashboard, and a team whose job is to grow it. The margin you are leaking and the customer you stopped calling have none of those things, and what nobody owns and nobody watches does not get worked.

There is also an ego tax. Landing a new logo feels like winning. Re-signing someone you let slip, or admitting your favorite product loses money, feels like cleanup. So the founder chasing growth never turns around to collect what is already sitting behind them.

Owners collect before they chase. Knowing exactly which line earns, which leaks, and which customers you have already paid to win is a form of clarity most operators trade away for the thrill of the new.

The most expensive dollar in your business is the one you spend chasing a stranger. The cheapest is the one you already earned and left sitting on the table because collecting it felt too quiet to bother with.

— Roland

The best AI advice might be destroying your exit value.

Right now, everyone is being told to become irreplaceable. Keep the relationships. Keep the judgment. Keep the magic that AI can't replicate.

Buyers don't pay premiums for businesses that depend on one person. The more essential you become, the harder it is for someone else to own what you've built.

There's a fundamental difference between building a business that performs well today and building one that someone will pay a premium to acquire tomorrow.

Want to see why being irreplaceable can lower your valuation? Watch the video below.