Ad Attribution in 2026: Platform Settings, First-Party Tracking, and What to Trust

Add up the conversions Meta, Google and Klaviyo report for last week and you will almost always get more orders than your store actually took. Nobody is lying. Each platform is applying its own ad attribution rules to the same customers, and each one gets to decide how much credit it deserves.

Ad attribution is the process of assigning credit for a sale to the ads and messages that led to it. In 2026 the practical answer to “which number do I trust” is simple to state: use each platform's own data to optimize inside that platform, and use one independent, first-party source to decide where the budget goes.

This guide walks through how each major platform attributes conversions, why their totals overlap, how first-party click tracking acts as the referee, and a decision framework you can apply on Monday morning.

What is ad attribution, and why do the numbers disagree?

Ad attribution is a set of rules that decides which ad interaction gets credit for a conversion, and the numbers disagree because every platform uses different rules and only sees its own touchpoints. Meta sees Meta clicks and impressions. Google sees Google clicks. Klaviyo sees email opens and SMS deliveries. None of them can see the full path, so each one fills the gap in its own favor.

Three variables drive almost all of the disagreement:

• The model: who gets credit when several touchpoints exist (last click, data-driven, incremental, last touch).
• The attribution window: how many days after a click or view a conversion can still count.
• The event types counted: clicks only, or also views, engagements, opens and deliveries.

Change any one of those and the reported number moves, even though the customer and the order stay the same. That is why we keep saying the ad platforms are grading their own homework. The grade is not fake. It is just graded on a curve each platform wrote for itself.

How does Meta attribute conversions?

Meta credits conversions based on an attribution model and window you choose at the ad set level. According to the Meta Business Help Center, advertisers can pick a standard or incremental model, plus a custom option that shares data from an external analytics tool.

Under standard settings, Meta can count three kinds of events:

• Click-through: conversions within 1 or 7 days after a link click.
• View-through: conversions within 1 day after someone saw the ad, with no click.
• Engage-through: conversions within 1 day after a non-link click action, which for video means a 5-second play.

The incremental model is different in kind: it optimizes for conversions Meta's models predict were caused by the ad. Meta also states plainly that results can't be compared across ad sets using different attribution models. If one ad set reports on 7-day click plus 1-day view and another reports click only, their ROAS columns are not measuring the same thing.

The view-through and engage-through categories are where most of the gap with your store comes from. A customer who scrolled past a video for five seconds, then searched your brand on Google two hours later and bought, can show up as a Meta conversion and a Google conversion. We go deeper on this in how view-through conversions inflate Meta ROAS, and our Meta attribution explainer covers the settings screen by screen.

What to do: write down the attribution setting for every active ad set before you compare anything. If they differ, stop comparing them.

How does Google Ads attribute conversions?

Google Ads uses data-driven attribution as the default for most conversion actions. Per Google Ads Help, data-driven attribution uses your account's conversion data to estimate how much each ad interaction contributed. Last click, which gives all credit to the last-clicked ad, is still available.

What changed is the menu. Google announced on April 6, 2023 that it was removing first click, linear, time decay and position-based models from Google Ads and GA4, as documented in this Google Ads Help notice. Selection of those models was removed from June 2023, and remaining conversion actions were switched to data-driven from September 2023. Google noted that fewer than 3% of Google Ads web conversions used those rules-based models at the time.

So inside Google's tools you now have data-driven, last click, or external attribution. Data-driven is a reasonable model for bidding, but it is a model Google runs on Google's view of the journey. It can't see the Meta ad, the TikTok video or the email that came first. If you want to know why Google removed four rules-based attribution models and what that means for comparing channels, start there, then read about the common Google Ads attribution gaps we see in audits.

What to do: keep data-driven for bidding if it is working, but do not use Google's conversion column as the tiebreaker when Google and Meta both claim the same order.

How do Klaviyo and other owned channels attribute sales?

Klaviyo uses a last touch model by default, with separate configurable windows for each channel. According to Klaviyo's conversion tracking documentation, the defaults for new accounts are 5 days for email opens and clicks, 5 days for SMS clicks, 12 hours for SMS deliveries and 24 hours for push opens. With delivery attribution enabled, an SMS delivery can be credited without a click.

Read that last part again. An email open within five days can earn credit. So can an SMS that landed on a phone. For a brand with a large list, that means a big share of orders driven by paid ads will also show up as flow or campaign revenue in Klaviyo.

That is not wrong for Klaviyo's purpose, which is judging which messages perform. It just means Klaviyo revenue and paid media revenue should never be added together. The AdBeacon Klaviyo integration brings email and SMS into the same click-based view as paid channels so the overlap becomes visible instead of hidden.

Platform attribution rules compared

The table below summarizes the default or standard behavior of each platform based on its own documentation.

Why do platform totals overlap?

Platform totals overlap because each platform counts every conversion it touched, and most orders were touched by more than one platform. There is no shared ledger that removes an order from Meta's column once Google claims it.

Here is a hypothetical. Say a store takes 100 orders in a week, according to Shopify. Meta reports 70 conversions. Google reports 45. Klaviyo reports 30. Add those up and you get 145 claimed conversions against 100 real orders, which is 45% more conversions than actually happened.

Now imagine you split next month's budget by those numbers. Meta looks like it drives 70% of sales, Google 45%, and email 30%. Those percentages sum to 145%, which tells you immediately they cannot all be used to allocate dollars. Our piece on why Meta, Google and TikTok ROAS never agree covers the same problem with TikTok in the mix.

The overlap is the core problem with ad attribution tracking in a multi-channel account.

What is first-party click tracking, and why does it work as a referee?

First-party click tracking captures the visitor's ad click on your own domain, stores it with your own data, and ties it to the order your store recorded. Because one system sees every channel's clicks and the real order list, each order can only be credited once.

A solid setup has three layers:

1. A first-party pixel on your store that records sessions and the UTMs on each landing.
2. Click ID capture for fbclid, gclid, ttclid and similar parameters. Our guide to click IDs like fbclid, gclid and ttclid explains what each one carries.
3. Server-side events sent back to the platforms. Meta's Conversions API creates a direct connection between your marketing data (from a server, website platform, app or CRM) and Meta's optimization systems. Google and TikTok have equivalents, covered in our server-side tracking setup guide.

AdBeacon is built on this approach. It only credits clicks it can verify, not view-throughs, and it shows AdBeacon-measured numbers side by side with platform-reported numbers. That second part matters. You don't have to throw away Meta's view. You get to see the gap and decide what to do about it. The AdBeacon reporting view is designed around that side-by-side comparison.

Click-only is a deliberate trade-off. It will undercount some genuine influence from video and display impressions. In exchange, every credited sale has a verifiable click behind it, which makes the number auditable. For budget decisions, auditable beats generous.

Which attribution windows and models should you use?

Use a window long enough to cover your real purchase cycle, and a model that matches the decision you are making. AdBeacon supports lookback windows up to 180 days and these models:

• First click and last click for spotting which channels open and close journeys.
• Linear, linear excluding direct, and paid linear for spreading credit across touchpoints.
• Full impact for a broader view of contribution.
• Lighthouse, which works like linear but credits only the platform you are optimizing. It is built for optimization rather than omnichannel reporting.

A high-ticket store with a 30-day consideration cycle should not be judging ads on a 1-day window. A low-cost impulse product probably doesn't need 180 days. Match the window to the data, not to the default.

A decision framework: what to trust for which decision

Use platform-reported data to optimize inside each platform, and use independent first-party data to allocate budget between platforms. Those are two different jobs, and most reporting problems start when one number is asked to do both.

For Meta specifically, the AdBeacon Facebook attribution tool and Chrome extension put click-verified numbers inside Meta Ads Manager, so a media buyer can make in-platform calls with both views on one screen.

Does this approach actually change results? For clients, average results after 90 days have been 15 to 20% revenue growth and 10 to 25% lower CAC. In one case with an agency partner, Sullen Clothing grew sales 53.7% and Bing revenue 568.4% over 90 days, with ROAS up 111.8%. Results vary by account and nothing is guaranteed. The underlying principle holds, though: better allocation starts with counting each order once.

What to do next

1. Audit settings. List the attribution model and window for every active Meta ad set, every Google conversion action and every Klaviyo channel. Note any mismatches.
2. Run the overlap test. For the last 30 days, add up platform-reported conversions and compare the total to store orders. The difference is your double counting.
3. Fix click capture. Confirm UTMs and click IDs survive to checkout, and that server-side events are firing to Meta, Google and TikTok.
4. Choose one budget referee. Pick a single independent, click-based source for cross-channel budget decisions and agree on its model and window.
5. Separate the jobs. Keep using platform data for in-platform optimization, and stop using it to split budget between platforms.
6. Review weekly. Put platform-reported and click-verified ROAS side by side every week and track how the gap moves.

Ad attribution in 2026 is not about finding the one correct number. Every platform is measuring something real, under rules written for its own purpose. The fix is to stop adding those numbers together and to give budget decisions to a source that sees every channel and counts each order once. If you want to see how big the overlap is in your own account, walk through your Meta, Google and Klaviyo overlap with our team.

Frequently Asked Questions

What is ad attribution?

Ad attribution is the process of assigning credit for a conversion to the ads and messages that led to it. Each platform uses its own model, window and event types to decide that credit.

Why does Meta report more sales than Shopify?

Meta can count conversions after link clicks, ad views and engagements such as 5-second video plays, and it cannot see whether another channel also touched the same order. Other platforms often claim the same order, so platform totals exceed store orders.

What attribution windows does Meta use?

Under standard settings, Meta counts click-through conversions within 1 or 7 days of a link click, view-through within 1 day of an impression, and engage-through within 1 day of a non-link click action. The setting is chosen at the ad set level.

Can I still use first click or linear attribution in Google Ads?

No. Google removed first click, linear, time decay and position-based models from Google Ads and GA4 in 2023, leaving data-driven, last click and external attribution. Independent attribution tools still offer rules-based models across all channels.

Should I trust platform data or independent attribution?

Use both for different jobs. Platform data is useful for optimizing inside that platform, while independent first-party, click-based attribution is better for deciding how to split budget between platforms because it counts each order once.

Sources

• Meta Business Help Center: About attribution models and attribution settings
• Google Ads Help: About data-driven attribution
• Google Ads Help: First click, linear, time decay, and position-based attribution models are going away
• Klaviyo Help Center: Understanding message conversion tracking
• Meta Business Help Center: About Conversions API