Meta says ROAS is up but Shopify revenue is flat. Which number should I trust?

Why do Meta and Shopify tell two different stories?

It is the first Monday of the month. Ads Manager says ROAS climbed from 2.5 to 3.0. Your agency's report has a green arrow on it. Then you open Shopify, and revenue is exactly where it was last month. Same spend, same team, two answers. (Illustrative numbers.)

The next question is always which one is lying. Neither is. They count different things, on different calendars, with different rules about who gets credit. Both can be accurate and still disagree by a wide margin.

So the useful question is not which number is true. It is which number answers the decision in front of you. That takes three numbers instead of one, four extra rows on your dashboard, and one check before any budget moves.

Which number answers which question?

Most of the confusion comes from asking one number to answer every question.

Platform ROAS is the right tool for picking the better of two ads, because both are measured with the same ruler. It is the wrong tool for deciding whether to add $20,000 a month, because the ruler belongs to the seller. Meta's own help page on attribution models notes that each model uses different counting mechanisms, and that if you use external analytics tools, you should evaluate performance in those tools (Meta Business Help Center).

What the gap looks like

Read it from the bottom up. The store did not grow. New-customer revenue fell, returning-customer revenue rose, and Meta's column grew by $30,000 while the total did not move. The likeliest story is not that the ads got better. It is that Meta counted more purchases from people who were already customers. (Illustrative numbers.)

What counts as a good level for each number depends on your margin. Our guide on what ROAS a supplement brand should target has the break-even math.

Why does Meta count sales that Shopify does not?

Meta credits a purchase to your ad if it happened within a set number of days after someone viewed or clicked that ad (Meta Business Help Center). Almost every gap between the two dashboards follows from that one rule.

1. The window includes views, not just clicks

The common setup, and what Meta's reporting API calls its default, is 7-day click and 1-day view (Meta for Developers). Under Meta's current standard settings, click-through counts purchases within 1 or 7 days of a link click, view-through counts purchases within 1 day of an impression, and engage-through counts purchases within 1 day of a non-link interaction, such as a video played for 5 seconds (Meta Business Help Center).

A view has no click, so it carries no UTM and no referrer. Shopify cannot see it. Meta itself notes that some third-party tools struggle to attribute conversions to impressions (Meta Business Help Center). Every view-through purchase is a sale Meta can claim and Shopify will book under another channel.

2. Meta counts people across devices

Meta measures people, not browsers, so it can connect an ad seen on a phone to a purchase made later on a laptop (Meta Business Help Center). Meta says cookie-based tools may miss those conversions, and that a buyer who returns hours later in a new browser tab loses the referrer entirely (Meta Business Help Center).

3. Some purchases are modeled

Where conversion data is missing or partial, Meta uses statistical modeling to estimate conversions, and it may also model how reported conversions are split across campaigns, ad sets and ads (Meta Business Help Center). A modeled purchase is an estimate. There is no single Shopify order you can match it to.

4. Returning customers and renewals can count too

Meta's rule asks whether a purchase followed an ad. It does not ask whether the buyer is new. Your existing customers see your ads, and many of them were going to reorder anyway. A customer who scrolls past your ad on Monday and reorders from your email on Tuesday is a Meta purchase in Ads Manager and an email order in Shopify.

Renewals depend on how your tracking is set up. Shopify's own Meta integration sends its Purchase event when a visitor completes a purchase and views the thank you page (Shopify Help Center). A renewal charged in the background has no visitor and no thank you page. Some server-side tracking setups, though, send every order to Meta, renewals included. If yours does, Meta can credit renewals from subscribers who saw an ad inside the window.

How does Shopify decide which channel gets the sale?

Shopify's marketing reports default to last non-direct click. The last channel a customer clicked before buying gets 100% of the credit, and direct visits are skipped. If the whole journey was direct, the order is booked as direct (Shopify Help Center). The Performance by marketing channel report shows the last 30 days on that model unless you change it.

Three things follow. Shopify credits clicks only, so a purchase that followed a Meta view lands on email, search or direct. Shopify has to know where a click came from, which for ads managed outside Shopify means UTM parameters (Shopify Help Center). And Shopify gives each order to one channel, while every platform counts its own. Shopify's own example: a customer who clicks both an email and a Google Shopping ad can be recorded as a conversion by each, while Shopify credits only the most recent click within the past 30 days (Shopify Help Center).

That is why adding up what Meta, Google and your email platform each report can give you more revenue than the store took in. Shopify's numbers are the only ones that have to add up to the orders you shipped.

Shopify also offers an any click model, which credits every channel a customer clicked. Shopify suggests it for analyzing one channel or reconciling with what each channel reports (Shopify Help Center), which makes it the fairer view when you hold Shopify up against Meta.

Different calendars

Meta can date a purchase to the day of the ad impression or the day of the conversion. If someone saw the ad on January 1 and bought on January 2, impression-time reporting puts the sale on January 1 (Meta for Developers). Meta tells advertisers to match the attribution window, time zone and conversion date setting before comparing tools (Meta Business Help Center). Shopify books the order on the day it was placed. Over a full week the calendars mostly even out. Around a sale or a launch, they do not.

The gap can run the other way

Meta can also miss sales. If a shopper runs an ad blocker, the Meta Pixel may not fire, and Meta reports fewer conversions than your own data (Meta Business Help Center). Shopify's Enhanced and Maximum data-sharing settings send the purchase event server to server through the Conversions API, which browser-based ad blockers cannot block (Shopify Help Center). Supplement brands should also check for Meta's health and wellness category, covered in our guide on getting that category removed.

Because the gap runs both ways, its size tells you little. Its direction over time tells you a lot.

Why does a ROAS dashboard credit media for everything?

Look at the numbers on most growth dashboards: ROAS, CPA, MER, CAC. Every one of them has ad spend in it. That makes the dashboard blind to where a gain came from. When a better landing page lifts revenue, every ratio improves, and the report hands the credit to whoever manages the spend.

Here is what that looks like when only the landing page changed. Sessions and revenue come from Shopify, for Meta-tagged sessions.

A spend-and-ROAS dashboard reports the right column as "Meta ROAS up 25%". The media did not change. Cost per session was flat. The page did the work. Run the same table the other way and you get the opposite mistake: a stockout or an ended promotion drops revenue per session, ROAS falls, and the media buyer gets blamed for a site problem. (Illustrative numbers.)

So add four rows under ROAS and never show it without them: sessions, cost per session, conversion rate and revenue per session, for Meta traffic, from Shopify. If cost per session moved, look at media. If revenue per session moved, look at the site and the offer. Our guide on a rising Meta CPA splits the media side further, and our anonymized conversion loop case study shows media buying and CRO run as one loop, read on volume-matched days.

How much of Meta's revenue can Shopify actually see?

Before you trust any CAC or payback figure built on Meta's purchases, answer one question: of the purchases Meta reports, what share can Shopify see arriving from Meta? We call that share UTM coverage. Put Meta's reported purchases for a week next to Shopify's orders that week from sessions tagged with your Meta UTMs, on the any click model so other channels do not take the credit.

Coverage will never be 100%, for all the reasons above: views, cross-device journeys, modeled purchases, ads launched without UTMs. The level matters less than the trend.

Two more cuts sharpen the read. First, split Shopify's Meta-tagged orders into first-time and returning customers; Shopify's marketing reports carry both (Shopify Help Center). Second, compare Events Manager, which shows purchase events whether or not they are attributed, with Ads Manager, which shows only those attributed to people who were shown your ads (Meta Business Help Center). If Ads Manager claims a growing share of the same purchases while Shopify revenue is flat, Meta is taking a bigger slice of the same pie.

The same rule applies to every channel: if you cannot see where a sale came from, you cannot price it. Our guide on tracking ChatGPT and AI assistant traffic in Shopify and GA4 applies it to the newest one.

What would have happened without the ads?

Attribution, Meta's or Shopify's, answers "which touchpoint gets the credit?" It cannot answer "would this sale have happened anyway?" That second question is incrementality, and only an experiment answers it cleanly.

This is not a theoretical worry. Researchers analyzed 15 US advertising experiments at Facebook, covering 500 million user-experiment observations and 1.6 billion ad impressions, and compared the results with common observational methods. The observational methods often failed to produce the same effects as the randomized experiments, even after conditioning on extensive demographic and behavioral variables (Gordon et al., Marketing Science, 2019).

Three ways to measure it

• Holdout test (Conversion Lift). Meta compares purchases in a test group with purchases in a control group held out from the ads, scaled to the same size. The result is the number of conversions that would not have happened without the ads, reported with a 90% credible interval and a cost per incremental conversion (Meta Business Help Center).
• Geo test. Ads run in some regions and pause in comparable ones, and you compare sales between them. Meta's open-source GeoLift tool was built for cases where people-based Conversion Lift is not feasible, and names healthcare advertisers as an example. Meta still recommends people-based tests where possible because they have higher statistical power (GeoLift by Meta).
• Meta's incremental attribution model. Ads Manager offers an incremental attribution model that uses machine learning to predict whether a conversion was caused by an ad, and optimizes delivery toward those conversions (Meta Business Help Center). It is useful, but it is still Meta's model of Meta's effect, not an experiment you control.

You do not need a test running all year. You need one before a large budget increase, and a fresh one when Meta and Shopify have drifted apart for weeks.

Is my agency inflating ROAS?

Usually not on purpose. Meta's ROAS is a real number from Meta. It is also the number most favorable to the ads, and when a team is judged on it, nobody has a reason to look past it.

Several choices can lift reported ROAS without adding an order: leaning on view-through credit, shifting budget to retargeting and existing customers who were going to buy anyway, or reporting through a month when an email promotion carried revenue. A change of attribution model matters too. Meta warns that results cannot be compared across ad sets with different attribution models, because each one counts differently, and points to its Compare Attribution Settings feature for a like-for-like view (Meta Business Help Center). Because the attribution model also informs delivery, that feature is a safer way to see the view-through share than switching the setting.

Ask for these five things every month:

1. Platform ROAS, blended MER and first-order MER on one page, for the same dates.
2. The attribution setting on every active ad set, and the date of any change.
3. UTM coverage: Meta's reported purchases next to Shopify orders from Meta-tagged sessions.
4. New versus returning customers among Shopify's Meta-tagged orders.
5. Sessions, cost per session, conversion rate and revenue per session for Meta traffic.

A team that sends these without being asked is managing your business. A team that only sends ROAS is managing its own scorecard. Our guides on agency red flags and questions to ask a DTC agency go further.

How we read Meta against Shopify at Succession

When a founder brings us two dashboards that disagree, we do not pick a winner. We rebuild the view from the store outward.

1. One market first. We isolate the US before reading any rate. Mixed international traffic moves every ratio and hides the real change.
2. Shopify orders are the scoreboard. Meta's reported purchases decide which ad beats which. Shopify orders decide whether the business grew.
3. UTM coverage before CAC. We check what share of Meta's reported purchases Shopify can actually see before we trust any acquisition cost built on them.
4. Revenue per session next to ROAS. Sessions, cost per session, conversion rate and revenue per session sit under every ROAS line, so a better page is never booked as better media buying, and a site problem is never blamed on media.
5. Volume-matched days. We compare days with similar spend, and the same weeks last year, so scale and season do not pass for performance.

The data pull is automated on our side. The read itself is something any founder can ask their team or agency for this week.

What should you do this week?

1. Put platform ROAS, blended MER and first-order MER side by side for the last 12 weeks, US only if you sell abroad.
2. Check the attribution setting on every active ad set, and whether it changed during those weeks.
3. Compare Meta's reported purchases with Shopify orders from Meta-tagged sessions, week by week, and write down the share.
4. Add sessions, cost per session, conversion rate and revenue per session for Meta traffic under your ROAS line.
5. Split Shopify revenue into first-time and returning customers, and find out whether renewal orders are being sent to Meta as purchases.
6. If the numbers still disagree and a budget decision rides on them, plan a holdout or geo test before you move the money, or bring all of it to a free growth audit and we will read it with you.

FAQ