How to treat Meta ads as an investment and measure the return with Customer Lifetime Value

Give two founders the same budget and the same Meta account, and you'll often find they're doing something completely different with it, even when the campaigns look identical from the outside.

The first founder sees the money as an expense. It buys sales this month. Every pound is judged on what comes back within the window, and if the return isn't there quickly, the spend was wasted. The logic is tidy and it feels responsible, money in, sales out, measured to two decimal places.

The second founder sees the same money as capital. It's buying something that will still be paying out long after the campaign ends: recognition, trust, an audience that comes back. The return isn't only this month's sales. It's the value of every customer the spend brought in, across everything they'll go on to buy.

This isn't a disagreement about tactics. Both founders might run similar ads to similar audiences. It's a disagreement about what paid media fundamentally is, and that single difference changes how everything downstream gets decided: which creative gets made, which customers get pursued, and crucially, how success gets measured.

Two camps

These two founders represent two camps, and most of the paid media industry sits firmly in the first.

Call it the sales camp. In this view, paid media is a machine. Get the account structure right, find the winning audience, optimise the funnel, pull the right levers, and growth comes out the other end. There's a whole industry of advice built on this premise, the account structure that unlocks scale, the targeting trick, the campaign setup that supposedly prints money. It's seductive for good reasons. It's measurable. It feels controllable. There's always something to adjust, always a number moving. But activity is easy to mistake for progress, and a busy account is not the same thing as a growing business.

The other camp sees paid media differently. Not as the engine of growth, but as a multiplier of it. In this view the brand creates demand, through what it makes, what it stands for, and how it shows up, and paid media's job is to capture and accelerate that demand, not to manufacture it from nothing. Spend amplifies whatever is already there. Where there's a brand, it compounds. Where there isn't, it just buys expensive, forgettable clicks.

The distinction sounds subtle. In practice it's the difference between a business that has to keep paying for every customer and one that increasingly earns them. And the reason the sales camp is so crowded isn't that it's right. It's that it offers the thing every founder under pressure wants: the promise that growth is an engineering problem with a solution you can buy.

An asset, not an expense

Here is the distinction underneath everything else. An expense is something you pay to get a result now. You spend it, you get the thing, the transaction is closed. Working capital behaves this way, you pay to move stock, and when the stock is sold the money has done its job and you start again. There is nothing wrong with expenses. Every business runs on them. But an expense doesn't build anything. It buys an outcome and then it's gone.

Capital is different. Capital is spent to build something that keeps returning value long after the initial outlay. It sits on the business rather than passing through it. A brand is capital in exactly this sense. The recognition, the trust, the audience that comes back without being paid for again, these don't expire at the end of the campaign. They accumulate. Money spent building them is not consumed, it's converted into something that keeps working.

The sales camp treats paid media as the first thing, an expense, judged on the sale it returns this month. The brand camp treats it as the second, capital, judged on the asset it builds over time. And this is where the category error lives. The sales approach isn't just measuring over a shorter horizon. It's trying to manufacture the output without building the input. It wants the returns of an asset while treating the spend as an expense, and those two things don't reconcile.

This is also why hacked accounts don't compound. There's nothing underneath the spend for it to multiply. You can optimise an account structure to perfection, but if there's no brand generating demand, all that precision is being applied to thin air. Paid media doesn't create wanting. It captures and accelerates wanting that already exists. Pour spend onto a brand people care about and it compounds. Pour the same spend onto one they don't, and you're renting attention you'll have to rent again tomorrow at a higher price.

The measurement trap

If the asset view is so obviously right, why does the expense view dominate? The answer isn't that founders are short-sighted. It's that the tools they were handed could only ever see expenses.

For most of the last decade, digital marketing ran on click-based attribution. A customer clicks an ad, buys within some window, usually 28 days, and the platform records the sale against the spend. It felt like a breakthrough. For the first time, marketing seemed perfectly measurable, every pound traceable to a result. But look at what that system can and cannot see. It captures the person who clicks and buys quickly. It is blind to the person who sees your ad, thinks nothing of it consciously, and buys three months later because the name felt familiar. It cannot see the customer who becomes worth ten times the average over two years. Everything that makes paid media an asset happens outside the window, and a system that can only see inside the window will therefore only ever value the expense.

This is the trap. The measurement model didn't just fail to capture brand value. It quietly trained a generation of marketers to stop building it. You optimise for what you can see, and the window decided what could be seen. Anything slow, compounding, or psychological, which is to say anything that builds a brand, was rendered invisible and treated as if it didn't exist.

It's worth being clear about who built this worldview and why. The push toward total trackability came largely from the more technical, engineering-minded end of marketing, where the instinct is reasonable: if you can measure something precisely, you can optimise it. The blind spot is that a system built to be precise will systematically undervalue whatever it can't track cleanly, and the things it can't track, memory, trust, association, the slow accumulation of wanting something, are exactly where brands are made. Precision isn't the same as completeness. A measurement model can be accurate about the wrong things.

The irony is that this supposedly precise system has been quietly breaking. As privacy regulation tightened and third-party cookies were deprecated, the identity tracking that click attribution depends on degraded sharply, with usable coverage falling from over 90% in the cookie era to somewhere between a third and two-thirds of journeys today. When you can only see a fraction of what's happening and you award full credit based on the last click, that isn't measurement anymore.

Which is why an old idea is having a renaissance. Marketing mix modelling has existed since the 1960s, long before anyone could track an individual click. Rather than following individuals, it works from the top down, using statistical analysis of aggregate spend and sales over time to infer what each channel is actually contributing, including the slow, brand-building effects that click attribution was structurally unable to see. It needs no cookies and no personal data, which is exactly why it's moved from a quarterly exercise for large advertisers back toward the centre of how serious brands measure. The wheel has turned. The model digital marketing once dismissed as old-fashioned is the one that can actually see the asset being built.

How we measure instead

If the window is the wrong lens, what's the right one? The answer is to measure paid media the way you'd value any asset, by the future return it generates, not the immediate cost it incurs.

There's a finance idea underneath this. When you value a company properly, you don't look only at what it sold last month. You forecast the cash it will generate over years and discount that back to what it's worth today. A business, in the end, is worth the value of its future customer relationships minus the cost of winning them. Paid media should be judged the same way. The question isn't what a campaign returned this month. It's what the customers it acquired will be worth over their lifetime, set against what they cost to acquire.

This is the logic of customer lifetime value, and the marketing academic Peter Fader has spent a career making the case for it. His more pointed argument is about averages. Most businesses, when they think about customers at all, think about the average one, average order value, average acquisition cost, average return. But customers are not an average. They are a wildly uneven distribution, where a relatively small group of high-value customers carries most of the business, and a long tail contributes very little. Treat them as an average and that crucial difference disappears from view.

This matters enormously for how paid media gets judged. Optimise on blended ROAS inside a 28-day window and you're not just measuring over the wrong timeframe, you're optimising toward the average customer, who barely moves the business, instead of toward the high-value customers who are the entire point. Worse, the two approaches tend to attract different people. A brand that earns attention by being worth caring about tends to draw customers who stay and spend. An account engineered to convert whoever is cheapest this month tends to draw exactly that, the cheapest, least committed customers, optimised for at the expense of the ones who matter.

So the metrics that actually tell you whether an asset is being built are the ones the expense model ignores. Lifetime value against acquisition cost. Payback period. Repeat rate. The trajectory of customer value over time rather than the return on a single campaign. These are slower and less immediately satisfying than a ROAS figure that updates daily, but they answer the only question that matters: are we acquiring customers worth more than they cost, and is that gap widening or narrowing.

The mechanics of modelling this properly, predicting lifetime value for customers who buy at unpredictable intervals and never tell you when they've left, is a genuinely technical problem, and one worth its own discussion. For here, the principle is what counts. Measure the asset, not the expense. Judge paid media by the customers it builds, not the clicks it books.

Why paid media can't stand alone

All of this is why we don't run paid media as a standalone service. On its own it's just an expense, a way to spend for a short-term return, with a ceiling on how good it can be because you're renting attention rather than building anything. It becomes capital only when it sits inside something larger, a brand worth paying attention to, creative that gives people a reason to care, the data to know which customers are actually worth acquiring. Embedded in that, the same spend stops being a monthly cost and starts to compound.

It's also the thread running through everything else worth understanding about the platform right now. The mechanics increasingly reward brands that have built something and punish those trying to manufacture growth from spend alone. The platform has started enforcing the distinction between the two camps itself.

So the question was never really which camp wins. It's a quieter one. Are you treating this money as an expense, paying each month for sales you'll have to buy again next month? Or as capital, building an asset, a brand, an audience, a base of customers worth more than they cost, that keeps returning long after the spend is done. The founders who grow something lasting tend to have decided that on purpose. The rest are still trying to hack their way to a number, and wondering why it never compounds.