Ask a beauty brand what a customer costs and you will usually get a number, delivered with confidence, that nobody in the room can trace back to a system. Ask which system produced it and the confidence evaporates. Meta says one thing. GA4 says another. Shopify quietly says a third. All three are correct inside their own definitions, and the spread between them is routinely wide enough to reverse a budget decision. We have spent this series cleaning up the scoreboard, settling efficiency, thresholds and cost. The last piece of the cleanup is the least glamorous and the most expensive to skip: deciding which system supplies the number in the first place.
Three systems, three answers, one month
The gap is not a rounding error and it is not a bug. Each system was built to answer a different question. An ad platform answers what its own media influenced. An analytics tool answers what the last recorded session did. A commerce platform answers how many orders were placed and by whom. None of them was designed to tell you what a new customer cost across every channel at once, which is the only question a growth budget actually turns on. Run the three side by side for a single month and the spread becomes obvious.
Worked example
A skincare brand spends $60,000 in a month, split across Meta and Google. Meta reports 520 purchases. Google reports 240. Platform reported total is 760 purchases, which prices a customer at $78.95.
GA4, on last click, credits 430 conversions to paid. Same spend, same month, and the number moves to $139.53. Shopify, meanwhile, recorded 610 orders in total, of which 510 came from customers who had never bought before. Blended acquisition cost on genuinely new customers is $117.65.
One month, one spend, three answers: $78.95, $117.65 and $139.53. The platform number is 33% cheaper than reality and the analytics number is 19% more expensive. Pick the wrong one and you either scale into a loss or cut a channel that was paying.
Why the platforms overcount, and why that is not lying
Meta and Google each claim a conversion when their own attribution window says they touched it. Neither knows what the other did. A customer who saw an Instagram ad on Tuesday and clicked a branded search ad on Thursday is one customer and two reported purchases. Add view through windows and the double count grows. This is not deception, it is scope: each platform is answering honestly about its own media, in isolation, and no platform has visibility into the others. The error only appears when you add the reported totals together and treat the sum as a customer count, which is exactly what most reporting sheets do.
The practical consequence is that platform reported acquisition cost is always the most flattering number available, and it gets more flattering as you add channels. A brand running two channels sees a modest gap. A brand running Meta, Google, TikTok and retail media sees a gap wide enough to make an unprofitable month look like a good one. If you want the benchmark range your account should be judged inside before you start reconciling, our Meta advertising cost benchmarks for beauty brands set it, and the 2026 skincare advertising benchmarks break it out by channel.
Why analytics undercounts
GA4 has the opposite bias. Last click credit hands the sale to whatever session closed it, which in beauty is very often branded search or direct, because the customer discovered you on social and came back later by typing your name. The paid social that created the demand gets nothing. Consent banners, browser restrictions and cross device journeys strip out more. Data thresholding quietly removes low volume rows from reports entirely, which we covered in our guide to overcoming thresholding in Google Analytics 4. The result is a system that systematically flatters your cheapest channel and punishes the one doing the discovery work.
Your store is the only system counting customers
Shopify does not attribute, and that is precisely why it should anchor the number. It records orders that happened and it knows which of those orders came from someone who had never bought before. That single distinction, new against returning, is the one your ad platforms cannot make and your analytics tool makes badly. Acquisition spend buys new customers. Returning customers were already yours. Any acquisition cost calculated without that split is measuring something else and calling it CAC.
The number that reconciles all three
Blended acquisition cost on new customers is total marketing spend divided by new customers recorded in the store. In the example above, $60,000 divided by 510 gives $117.65. It needs no attribution model, survives every privacy change, and cannot be inflated by adding channels. It is also the only version of the number your finance team will recognize, because it reconciles to money that actually left the business and orders that actually arrived. Platform figures still earn their place, but as directional signals for deciding where inside a channel to move budget, never as the count of customers. If you need the underlying definitions before layering any of this, start with our primer on what return on ad spend actually measures, then run the arithmetic in our ROAS calculator and, on the impression side, our CPM calculator.
Track the gap itself, not just the numbers
Here is the move almost nobody makes. Divide platform reported purchases by new customers in the store and record that ratio every month. In the worked example it is 760 divided by 510, or 1.49. That single figure is your duplication factor, and its stability matters more than its size. A ratio that sits near 1.5 month after month means your platform reports are usably predictive once discounted. A ratio that jumps from 1.5 to 2.1 without a change in channel mix means something broke in tracking, not in marketing, and chasing the acquisition cost movement will send you after the wrong problem for two weeks. Beauty brands lose more money to misdiagnosed measurement than to genuinely bad media.
The reconciliation, in four steps
- Take total marketing spend for the month from the invoices, not from the platform dashboards.
- Take new customer count for the same month from your store, filtered to first time buyers.
- Divide one by the other. That is your blended acquisition cost, and it is the number that decides budget.
- Record platform reported purchases divided by new customers as your duplication ratio, and watch it for movement rather than size.
Where the gap costs beauty brands the most
The damage is rarely a slow leak, it is a single decision. A brand looks at platform reported acquisition cost of $78.95, sees comfortable headroom against contribution margin, and doubles prospecting. Reality was $117.65, the headroom was thin, and the extra spend goes in at a loss that will not surface for two months. The mirror image is just as common: a team reads the GA4 number, concludes paid social does not work, cuts the channel that was doing the discovery, and watches branded search volume fall six weeks later with nobody able to explain why. Both decisions look disciplined in the moment. Both came from reading a number that was never built to answer the question asked of it. How competing brands present value at the point of purchase, which is where the conversion side of this equation gets won, is the substance of our operator review of Sweet Chemistry skincare.
Skincare, where reorders distort all three systems
Routine products break the gap open wider than any other beauty category. A serum or a cleanser generates a second and third purchase inside a year, so the store fills with returning orders that no acquisition spend bought. Count those orders in your denominator and acquisition cost looks brilliant. Strip them out and the real number can be double what the dashboard showed. Subscriptions make it sharper again, because a recurring charge is an order with no marketing attached to it at all. The discipline is to filter to first time buyers before you divide anything, then set a payback window, at 60 or 90 days depending on what your cash position tolerates, and judge the acquisition cost against contribution margin inside that window rather than against last month's figure. Our beauty and skincare growth practice starts every engagement here, because no creative or bidding decision above it can be trusted until the denominator is clean.
Medspa, where the conversion is a booking and not a sale
Aesthetics practices have the worst version of this problem, because the thing the ad platform counts is not the thing the business sells. A form fill is not a patient. A booked consultation is not revenue. Between the reported lead and the treatment sit three more leaks: the lead has to answer the phone, the booking has to show up, and the consultation has to convert. Each leak multiplies, and none of them appears in any advertising dashboard. A practice reporting a $90 cost per lead with a 45% show rate and a 50% consultation conversion is actually paying $400 per patient. That gap is far larger than any plausible movement in media price. Fix the show rate before you touch a bid. If aesthetics is your category, our medspa advertising benchmarks put the ratios in order, and our rundown of the best medspa marketing agencies covers who runs that funnel properly.
Stop trying to close the gap
A lot of measurement effort goes into forcing the three systems to agree. That work never finishes, because the systems are answering different questions and always will. The productive version is to stop treating disagreement as a fault. Name one system as the source of truth for how many customers you bought, which is your store. Use the platforms for direction inside a channel, which is what they are good at. Use analytics for on site behavior, which is what it is good at. Then watch the ratio between them as its own diagnostic. Choosing which number counts is the decision. Everything downstream, including which agency you hire to run the media, depends on it, which is the first thing we look at in our comparison of the best beauty marketing agencies.
One row to add to your sheet this week
No new tooling is required for any of this. Add a single row at the top of your monthly reporting sheet with four cells: total spend, new customers from the store, blended acquisition cost, and the duplication ratio. Put it directly above the platform reported figures so the comparison is unavoidable every time somebody opens the file. When acquisition cost moves next month, that row tells you within a minute whether you bought more expensive customers or simply changed the way you were counting them. That distinction is the entire point of the KPI cleanup, and it is worth more than any individual metric in the stack. More of this thinking sits in our insights library.