Beyond ROAS: How to Tell You If Your E-Commerce Is Growing

2 weeks ago

104 7 minutes read

STORY INLINE POST

Idioma Leer en Español DIA assistant 1.0x ✕

ROAS may be the most trusted number in e-commerce and the most misleading. It sits at the top of almost every dashboard, it decides where budgets go, and it is the figure teams celebrate when it climbs. Yet, I have sat in more than one meeting where the dashboard showed a healthy return on ad spend, everyone felt good about the quarter, and the business itself was not growing. Revenue was flat. New customers were not really new. The number was up and the company was standing still. That contradiction is worth understanding, because it is far more common than most leaders would like to admit.

The root of the problem is that ROAS answers the wrong question. It tells you how much revenue an ad platform decided to associate with the money you gave it. The question that actually matters to the business is different: How much revenue would you have lost if you had not spent that money at all? Those two questions sound similar. They are not. The first is a claim of credit. The second is a measure of cause. And the gap between them is where a great deal of marketing budget quietly goes to die.

That gap is not a rounding error. Incrementality research from independent measurement firms consistently finds that platform-reported ROAS overstates true incremental impact by somewhere between 20% and 60%, and for certain tactics the inflation is far worse. Haus, one of the firms that runs controlled experiments for brands, found across hundreds of tests that the incremental return on retargeting often runs 40% to 70% below what the platforms report. Think about what that means in practice. A retargeting campaign showing a 6x return on the dashboard may be delivering something closer to 2x in real, incremental terms, and the difference is money spent to reach people who were going to buy anyway.

The reason this happens is structural, not accidental. Every ad network measures conversions with its own attribution window, its own view-through rules, and its own commercial incentive to claim as much credit as possible. Meta counts an order inside its window. Google counts the same order. Your email tool counts it a third time. Add up what each channel claims and the attributed revenue can exceed the revenue the company actually booked, which is mathematically impossible and yet happens on real dashboards every day. Each platform is optimizing to look indispensable. None of them is optimizing to tell you the truth about your business.

A Tale of Two Seats

I look at this from two seats at once, and that is what made the issue impossible for me to ignore. At Infracommerce, I work close to the profit and loss statement of large brands, where the only revenue that counts is the revenue that clears, net of returns, discounts, and the cost of getting the product to the door. At Kreativo Studio, I work on the other side, where the pressure is to report a campaign result that looks impressive to a client or a board. Living in both worlds at the same time, you learn quickly that a number designed to look good in a report and a number that reflects whether the business is healthier are rarely the same number. The job of a serious marketer is to stop confusing the two.

So if ROAS is too narrow, what should a leader actually watch? I would argue for three numbers, and none of them lives inside an ad platform. The first is MER, the marketing efficiency ratio, which is simply total revenue divided by total marketing spend across every channel. MER does not care which platform claims the sale. It looks at the whole business and asks a blunt question: for every peso the company put into marketing, how many pesos came back? Because it is blended, it cannot be gamed by attribution games between channels. Triple Whale, which aggregates data across thousands of e-commerce brands, put the median MER at around 41% in 2025, which is roughly 2.4 pesos of revenue for every peso of marketing. That is a useful benchmark, but the point of MER is not the industry average. It is that a single, honest, top-down number is harder to fool than a wall of platform metrics that each flatter themselves.

The second number is contribution margin, and it is the one marketing teams most often ignore and finance teams most often obsess over. Contribution margin is what is left from a sale after you subtract the costs that scale with that sale: the product, the shipping, the payment fees, the returns. It matters because it sets the real break-even point for your advertising. There is a simple relationship every executive should keep in mind: your break-even MER is one divided by your contribution margin. At a 40% margin you break even at roughly 2.5x. At a 30% margin you need about 3.3x just to stand still. This is not a theoretical concern. Across direct-to-consumer brands, median contribution margins have fallen from around 35% in 2021 to roughly 22% in 2025, largely because paid acquisition got more expensive. When your margin compresses like that, your break-even point rises, which means the exact ad performance that was profitable last year can be losing money this year while the ROAS on the dashboard looks identical. The number did not change. The economics underneath it did.

The third number is the payback period on a customer, measured against contribution margin rather than against top-line revenue. The right question is not how much revenue a new customer generated on the first order. It is how many months of that customer's contribution it takes to earn back what you paid to acquire them. A business that recovers acquisition cost in two months can afford to grow aggressively. A business that takes fourteen months is financing its own growth with cash it does not have, no matter how attractive the ROAS looks. Customer acquisition cost is not cheap and it is not getting cheaper. Industry benchmarks now put average e-commerce acquisition cost between roughly 68 and 84 dollars, up about 60% over five years, with Google search costs alone rising nearly 13% year over year. In that environment, knowing your true payback period is not a nice-to-have. It is the difference between scaling and quietly going broke.

None of this requires a 20-person data science team, which is the objection I hear most often from leaders who assume rigorous measurement is a luxury for giants. It is not. The most reliable way to know whether spend is incremental is also one of the oldest: turn it off in some places and leave it on in others, then compare. Geographic holdouts, structured on and off tests, and simple controlled experiments will tell you more about what your marketing actually causes than any attribution model ever will. This is no longer a fringe practice. According to EMARKETER and TransUnion data from mid-2025, about 52% of brand and agency marketers already use incrementality testing, and another third plan to invest more in it over the coming year. The tools have come down in price and up in accessibility. What is usually missing is not budget or technology. It is the decision to value the truth over the flattering number.

Discipline Is an Urgent Need

The Mexican context makes this discipline urgent rather than optional. E-commerce here reached 941 billion pesos in 2025, growing 19.2% in a single year, and online sales now represent 17.7% of all retail, enough to place Mexico eighth in the world, according to AMVO. Buyer confidence is rising fast, climbing to 66% in the third quarter of 2026 from 48% the quarter before. A market growing this quickly attracts more competitors, and more competitors bid up the cost of every click and every impression. When acquisition was cheap, sloppy measurement was survivable, because even the wasted spend often paid for itself. That era is ending. In a more expensive, more crowded market, the company that cannot tell the difference between spend that creates customers and spend that merely takes credit for them will lose ground to the company that can.

This is ultimately a leadership question more than a technical one, and it comes down to a room. In most companies, the CMO and the CFO look at different numbers, defend different stories, and meet mainly to argue about budget. That has to change. The marketing efficiency ratio, contribution margin, and payback period are numbers both of them can stand behind, because they connect marketing activity to the health of the business rather than to the self-interest of a platform. When the person who owns growth and the person who owns the money are finally looking at the same dashboard, the conversation stops being a fight over credit and becomes a shared question: is the business getting stronger, and is our spend the reason why?

I want to be fair to ROAS, because it is not useless. Inside a single channel, over a short horizon, watched by someone who understands its limits, it can help steer day-to-day decisions. The mistake is not using ROAS. The mistake is letting it stand in for the truth about the whole business, and letting a number that each platform has every incentive to inflate decide how a company invests millions. A metric that answers the question of what a platform claims it caused should never be confused with the question of whether the company grew.

So the honest test for any e-commerce leader is not whether the ROAS went up. It is whether the business is worth more this quarter than it was last quarter, and whether the money you spent is the reason. Those are harder questions to answer. They do not fit neatly on a single dashboard tile, and they cannot be exported from an ad account with one click. But they are the only questions that actually matter, and the companies that learn to answer them honestly will quietly pull away from the ones still celebrating a number that was only ever telling them what they wanted to hear.