How View-Through Conversions Inflate ROAS

Your Meta dashboard says ROAS 8. Your store's order list says the month was decent, not spectacular. Part of that gap has a name most store owners have never inspected: view-through conversions. These are orders Meta credits to your campaign because the customer saw an ad, scrolled past it, never clicked, and then bought within a day.

The customer may have bought because of the ad. Or they may have been on their way to buy anyway, and your ad simply appeared in the feed of someone who already knew your brand. Meta cannot tell the difference, but it counts the order either way.

What Is a View-Through Conversion?

A view-through conversion is credited when someone is shown your ad, does not click it, and later completes a purchase within the platform's view window. No visit from the ad, no click, just an impression followed by an order.

The windows are short but the volume is not. Meta's attribution setting defaults to 7-day click plus 1-day view: any purchase within 24 hours of an impression gets attributed to the campaign. Meta removed the 28-day windows in January 2021 when iOS 14 arrived, so the current system is already the restrained version. Meta has since also added a separate 1-day "engage-through" window that covers interactions like short video views, sitting between a click and a plain impression.

Google Ads has the same concept for its visual formats. Its view-through conversions apply to Display, Video, App, and Demand Gen campaigns, with a default view window of 1 day, adjustable from 1 to 30 days. Search ads have no view-through; you cannot "view" a text ad you did not click.

Where View-Through Hides in Your Reports

This is the part that actually costs money, because the two platforms report it in opposite ways.

Meta blends view-through into the numbers you look at daily. Results, cost per result, and ROAS all include view-through conversions by default. There is no asterisk in the interface. Unless you deliberately split the columns, you are reading clicks and impressions as one number.

Google keeps them apart, mostly. View-through conversions sit in their own column and stay out of "Conversions" and out of Smart Bidding targets. But there are exceptions that behave like Meta: engaged-view conversions (someone watches at least 10 seconds of your skippable YouTube ad, or 5 seconds of an in-feed or Shorts ad, then converts within 3 days) do land in the main Conversions column. Demand Gen campaigns can also opt view-through into bidding. So "Google does not count views" is only true until you run video.

GA4, for its part, is click-based. Its data-driven attribution does not see Meta impressions at all, which is one reason your GA4 and Meta numbers will never reconcile. The single nuance: GA4 does treat YouTube engaged views (30+ seconds) as a touchpoint.

The Inflation Math

Take a €100 order from a returning visitor. She bought from you twice before. Yesterday she scrolled past your retargeting ad on Instagram without slowing down. Today she typed your store name into her browser and ordered.

Under Meta's default setting, that is a view-through conversion: €100 of revenue attributed to the retargeting campaign. Multiply across a month and your retargeting ROAS looks heroic, precisely because retargeting targets the people most likely to buy anyway.

This is not a hypothetical bias. The largest controlled study on the question, 15 randomized Facebook experiments covering 1.6 billion impressions published in Marketing Science, found that observational attribution methods routinely overstate advertising's true effect compared to randomized lift tests. More recently, Seer Interactive tested Meta's own incrementality setting on $1.05M of ad spend and found Meta's default attribution claimed a meaningfully higher share of conversions than a GA4 cross-reference supported: 87% claimed as incremental versus 67% implied.

The honest take: view-through is not fake. Some of those people genuinely bought because the ad reminded them. But the platform grading its own homework has no incentive to tell you which ones, and the correlation runs the wrong way. The warmer the audience, the more view-through conversions, and the less the ad actually changed.

How to See Your View-Through Share

You do not have to guess. Both platforms let you split the number in a few clicks.

In Meta: open Ads Manager, click Columns, then Compare attribution settings. Tick 7-day click and 1-day view separately. The 1-day view column is your view-through share per campaign. Run it on your retargeting campaigns first; that is where the inflation concentrates.

In Google Ads: add the "View-through conv." column to any Display, Video, or Demand Gen report. It is already excluded from your Conversions column, so here the check is the reverse: make sure nobody on your team is quoting "All conversions" (which includes view-through) as if it were "Conversions."

Then do one more comparison that no platform column can do for you: put platform-reported revenue next to your actual orders. If Meta plus Google claim more revenue than your store recorded, the surplus has to come from somewhere, and view-through plus double-counting between platforms are the two usual suspects. Your blended ROAS is the ceiling check: total real revenue over total spend cannot be argued with.

Where Order-Based Attribution Fits

Enalitica does not count view-through conversions, by design. Attribution starts from the confirmed order in your WooCommerce or Shopify store, and an order is only tied to a campaign when there is a captured click behind it: a gclid, fbclid, gbraid or another click ID stored with that specific order. An impression cannot produce a click ID, so an impression can never claim an order.

In practice you see three numbers side by side for the same campaign: what Meta claims (clicks plus views), what Google claims, and the revenue from orders that carry an actual click from that campaign. The first time store owners see this, the reaction is usually the same: the Meta number shrinks, the "Direct" bucket shrinks, and the sum finally matches the bank statement. Your ROAS will look lower than Meta's. It is also the one you can defend in a spreadsheet with order numbers in it.

That does not make view-through useless. It is a soft signal for prospecting reach, and if you want Meta's algorithm to keep learning from view events, you should keep sending them. The rule is narrower: never let a view-through number decide where next month's budget goes. Book a demo and we will pull up your own campaigns and show you the click-backed revenue next to the platform-claimed revenue.

Frequently Asked Questions

Should I change my Meta attribution setting to 7-day click only?

Changing the ad set attribution setting changes what the delivery algorithm optimizes for, not just reporting, so treat it as a media decision rather than a reporting fix. The safer first step is to use the Compare attribution settings columns for honest reporting while leaving optimization untouched. If a campaign's results turn out to be mostly 1-day view, fix the strategy, not just the column.

Are engaged-view conversions the same as view-through conversions?

No, and the difference matters for reporting. Engaged-view requires real engagement (at least 10 seconds of a skippable video ad, or the full ad if shorter) and Google counts it in the main Conversions column with a 3-day default window. View-through requires nothing but the impression and stays in its own column. Meta's newer engage-through window is the same idea on the Meta side.

Do view-through conversions show up in GA4?

No. GA4 only attributes across clicks (plus YouTube engaged views on Google's own inventory), so a Meta view-through conversion appears in GA4 under whatever channel the buyer actually arrived from, often Direct or Organic. This mismatch alone can explain a large share of the Meta-versus-GA4 gap.

What is a normal view-through share for a store?

There is no published benchmark worth citing, and the share varies mostly with audience temperature: cold prospecting tends to show a low view-through share, retargeting a high one. That variance is the point. Measure your own split with the comparison columns before assuming your ROAS is click-earned.

Does turning off view-through events break Meta's optimization?

You cannot fully "turn off" impressions being used for attribution without changing the attribution setting, but you can and should keep sending complete conversion events (including server-side ones) to Meta regardless of how you report internally. Signal for the algorithm and truth for the budget meeting are two different jobs. Use both.