Many businesses invest in Meta Ads or Google Ads every month and, when asked whether it works, give the same answer: “I think so”. The platform dashboard shows thousands of impressions, hundreds of clicks and a cost per click that looks reasonable. But the important question is a different one: how much money came in thanks to those ads, and how much was left after paying for them?
In this guide we explain which metrics to look at, why platform data doesn’t always match your sales, and how to know, with numbers, whether your advertising is working.
The metrics that mislead
Ad platforms show a huge amount of data, and almost all of it is useful for optimising campaigns. But some of it gets mistaken for results when it isn’t:
- Reach and impressions: how many people saw the ad. Seeing it is not buying.
- Clicks and CTR (click-through rate): how many people came in. If the website doesn’t convince them, those clicks are lost.
- Cost per click (CPC): a cheap click is worthless if the person arriving has no intention to buy.
- Engagement: likes, comments and shares. They help the brand, but not always the till.
None of these figures is bad. The problem starts when they are the only way a campaign is measured. A campaign can have an excellent CTR and lose money, while another with expensive clicks can be the most profitable one in your business.
The metrics that do matter
A sales campaign is judged by what happens after the click. There are four concepts worth mastering:
Conversions
A conversion is the action you actually care about: a purchase, a quote request, a call or a booking. The first step is to decide what counts as a conversion and set it up properly on each platform. If you measure “visits to the contact page” instead of “forms submitted”, your data will tell you things are going better than they are.
Cost per acquisition (CPA)
This is what it costs you, on average, to get a sale or a lead: spend divided by conversions. If you invest €600 and get 20 sales, your CPA is €30. That number only makes sense when compared with what each sale leaves you.
Return on ad spend (ROAS)
This is what you invoice for every euro invested in ads: attributed revenue divided by spend. A ROAS of 4 means every euro invested brings in four in sales. Careful: ROAS is not profit. From that revenue you still have to subtract product cost, shipping, fees and all other expenses.
Break-even ROAS
There is a simple calculation every business should do before launching campaigns. If your gross margin (what you keep from each sale before advertising) is 40%, you need to bring in at least €2.50 for every euro invested just to avoid losing money (1 ÷ 0.40 = 2.5). Below that ROAS, every sale your ads bring in costs you money.
A campaign with lots of cheap clicks and few sales is usually a worse deal than one with fewer, more expensive clicks that actually sells.
Why platform data doesn’t match your sales
If you have ever compared the sales Meta or Google report with those in your online store, you will have noticed they don’t match. That’s normal, and there are several reasons:
- Each platform attributes the sale in its own way. If a customer sees an ad on Instagram, searches for your brand on Google days later, clicks an ad and buys, both platforms may claim that sale.
- Attribution windows. Platforms count sales that happen several days after the click or the view. Depending on the settings, the numbers change quite a lot.
- Tracking being blocked. Browsers, ad blockers and rejected cookies mean some conversions are never recorded.
- Multiple devices. Many people discover a product on their phone and buy on their computer.
That is why it matters to have measurement set up properly from the start: conversion tracking on each platform, a connection with your store or CRM (also server-side where possible, to recover data the browser loses) and a clear rule about which data source has the final say. Without that, every decision is a gamble.
The most honest test: comparing periods
There is a very simple way to check whether advertising is driving sales: look at what happens with and without ads, keeping everything else as similar as possible.
We saw it with a sneaker e-commerce store we work with: with Meta Ads campaigns running in November and December, its December net sales were 137% higher than in October, before we started. And when the campaigns were paused, sales went back to their previous level. You don’t need a statistical model to see the effect of advertising there. You can see the full case in our work section.
This kind of comparison has its limits (seasonality, promotions and competitors also play a part), but it forces you to look at real revenue and not just the ads dashboard. Ideally, combine it with each platform’s measurement and your store’s data.
How to know whether to raise or lower your budget
Before touching your spend, answer these questions:
- Is your real ROAS above your break-even ROAS? If it isn’t, increasing the budget only multiplies the losses.
- Do the campaigns have enough data? Platforms need a volume of conversions to learn. With very few, results are unstable and it is better to wait or simplify the structure.
- Which part of the budget performs and which doesn’t? There are almost always ads, audiences or products pulling the weight and others dragging it down.
- Have you considered customer lifetime value? If your customers buy again, you can afford to pay more for the first sale.
Signs that something is wrong
- Your cost per sale rises month after month and nobody can explain why.
- You get lots of clicks, but people leave the website within seconds.
- Nobody reviews which ads, audiences or products bring in the sales.
- The campaigns have gone months without changes or new ads.
- You don’t know the maximum you can pay to win a customer.
- The reports you receive only talk about reach, clicks and engagement.
If two or more of these sound familiar, you are probably paying for traffic, not for sales.
Frequently asked questions
What is a good ROAS?
There is no universal figure: it depends on your margin. A good ROAS is one that is clearly above your break-even ROAS (1 divided by your gross margin) and leaves a profit after all costs. A ROAS of 3 can be excellent for a high-margin business and ruinous for a low-margin one.
How long should I wait before judging a campaign?
The first weeks are for the algorithm to learn and for testing messages and audiences. As a rule of thumb, wait until you have a reasonable number of conversions before drawing conclusions, rather than deciding on two or three days of results.
Why does Meta Ads say I sold more than I see in my store?
Because it attributes sales to people who saw or clicked an ad within a window of days, even if they later bought through another route, and because other platforms may be counting the same sale. Always compare with your store’s data and use the same criteria every month.
Is it better to measure in the ad platform or in Google Analytics?
The best approach is to use several sources with a fixed rule: the platform to optimise campaigns, analytics to see the full journey, and your store or CRM as the final reference for what was actually sold.
In short
Digital advertising works when it is measured by what it leaves in the till. Before raising or lowering your budget, make sure you know what each sale costs you, what your break-even ROAS is and how much profit you keep. If you are torn between platforms, our comparison of Google Ads and Meta Ads may help. And if you want an idea of what your campaigns could deliver, try the AVAR LAB ad simulator: it’s free and tells you what is data and what is an assumption.
Written by the Avar Marketing Systems team: three partners trained in marketing, advertising and management, running their own online stores since 2019. Meet the team →
