Last updated: November 17, 2026. Facts checked against the Shopify Help Center, Shopify Changelog and Meta Business Help Center in November 2026.
TL;DR
If your Facebook ads bring clicks and add-to-carts but no purchases, measure before you blame Facebook ads bot traffic. Compare link clicks with landing page views by placement in Ads Manager, filter Shopify sessions by Human or bot, and check both against your own account's recent baseline. A gap between Meta clicks and Shopify sessions does not prove bot traffic on its own, especially since Shopify changed how it counts sessions on September 21, 2026.
Quick answer
If Facebook ads bring clicks and add-to-carts but no purchases, don't start with the creative. In Ads Manager, compare link clicks with landing page views, then in Shopify filter sessions by Human or bot and compare them with your own data from the previous 30 days. A sharp gap between clicks and sessions does not prove bots by itself: after Shopify's September 21, 2026 change, session metrics shifted, while orders and sales stayed the same.
This guide is for Shopify store owners who suspect Facebook ads bot traffic and want a way to check it without buying a fraud tool first. AdAdvisor has 8+ years in media buying, more than $60M in managed ad spend and an ex-Meta engineer on the team, and the procedure below is the order in which we would rule causes in or out.
How do I tell if my Facebook ad clicks are bot traffic?
You can't tell from one number. Facebook ads bot traffic usually shows up as a pattern across several reports at once: clicks that don't turn into page loads, sessions Shopify labels as bots, and cart activity with no checkout behind it. Look for the pattern, and judge each signal against your own history.
| Signal | Where to look | What it may mean |
|---|---|---|
| Link clicks far above landing page views | Ads Manager, columns Link clicks and Landing page views | Slow or failing pages, people leaving before load, or non-human clicks |
| One placement with a much weaker click-to-load ratio than the rest | Ads Manager, Breakdown > By delivery > Placement | A placement worth investigating: accidental taps, load problems or low-quality inventory |
| Shorter sessions from paid social than your usual | Shopify Analytics, average session duration, compared with your own history | Bots, slow pages or accidental taps; real visitors also leave fast |
| Session countries outside your targeting | Shopify sessions by country vs your ad set locations | Bots, VPNs or proxies, or location mismatch between tools |
| A burst of add-to-carts within minutes, with no rise in reached checkout | Shopify add-to-cart rate and reached checkout rate, by hour | Automated cart activity, unless a promotion explains it |
| Many sessions labeled Bot | Shopify reports, Human or bot session filter or dimension | Shopify's classification of automated visits, which may not have come from ads at all |
There are no universal thresholds here, and we don't give any. No source we checked publishes a "gap above X% means bots" rule with a method behind it, so the useful comparison is your account against itself.
Bot sessions and invalid ad clicks are different things. Shopify classifies activity on your store, while Meta evaluates interactions with your ads. A crawler or price-comparison bot can show up as a Bot session without ever clicking an ad, and a click that never loads your store may leave no Shopify session at all.

Drowning in Meta Ads?
Put your campaign on autopilot with Nova.
Read moreWhy do my Facebook clicks and my Shopify sessions never match?
Because they count different events. Meta defines a link click as a click, swipe or other gesture on a link in the ad that leads to an advertiser-specified destination (Meta, Link clicks). A landing page view is counted only when the destination actually loads, measured through the Meta pixel or the in-app browser, and Meta says it may be estimated (Meta, Landing page views).
A Shopify session is a third thing again: Shopify's own record of activity on your store, which now ends after 30 minutes of inactivity and, by default, excludes sessions Shopify identifies as bots (you can switch the filter to include them) (Shopify Help, Session measurement update). These counts are not interchangeable. A click may fail to become a page load, and Meta lists two reasons: the page didn't fully load, or the ad had several destinations (Meta, Link clicks vs landing page views). Shopify, in turn, may group several visits into one session or count a journey that had no standard pageview.
So some gap is normal in almost every account. The largest gap in your reports is not the most useful one: a steady difference between clicks and sessions is usually normal, while a sudden change in one placement's click-to-load ratio is a signal you can act on. The structural reasons Meta and analytics tools disagree on conversions are covered in our guide to Meta ads attribution.
How do I measure the gap against my own baseline?
Compare the click-to-load ratio now with the same ratio before the problem started, placement by placement, and keep the Meta and Shopify comparisons separate. We call this the Account-Baseline Traffic Check. It uses only Ads Manager and Shopify Analytics, and it shows where the change sits before you touch any setting.
Set two windows.
Use the 30 days before you first noticed clicks without sales as the baseline, and the most recent 30 days as the comparison. Note any big change in between: new offer, new landing page, a sale, a new campaign.
Measure the Meta ratio.
In Ads Manager, export link clicks and landing page views for both windows with Breakdown > By delivery > Placement, using the same attribution and reporting settings. Divide landing page views by link clicks for each placement.
Compare each placement with its own baseline.
Not with an "industry norm". A placement whose ratio dropped while the others held steady is your first suspect. Ignore large swings on placements with very few clicks; they are likely noise.
Compare Shopify separately.
For sessions, add-to-cart rate and reached checkout rate, use periods that both start after September 23, 2026, with the same Human or bot filter in both.
Segment the drop.
Repeat the Meta breakdown by country and by impression device, and the Shopify reports by country and device, and check whether the change sits in one segment.
Change one thing and measure again.
Use the same baseline, so the result is comparable.
| Placement | Link clicks (baseline) | LPV (baseline) | LPV / clicks (baseline) | LPV / clicks (now) | Change (pp) |
|---|---|---|---|---|---|
| Facebook Feed | |||||
| Instagram Feed | |||||
| Instagram Reels | |||||
| Audience Network | |||||
| Other |
Ratio = landing page views / link clicks x 100%. Change (pp) = current ratio minus baseline ratio, in percentage points. A falling ratio marks a segment to investigate; it does not show that the segment contains bots. A landing-page change or a slow page can cause the same drop.
Why the Shopify windows must sit after September 23, 2026: Shopify changed its session measurement on September 21-23, 2026, and its advice is to use post-update data as the new baseline and the same Human or bot filter for both periods (Shopify Help). Meta's link clicks and landing page views were not affected by Shopify's change.
What changed in Shopify analytics on September 21, 2026?
Shopify changed how it counts sessions, not how many orders you have. From September 21, 2026, Shopify's default session reports focus on human visitors where the Human or bot filter is available, classify bot and system traffic more consistently, and keep some shopper journeys that have no standard pageview, such as a cart link straight to checkout (Shopify changelog, Sept 21, 2026).
Metrics that may look different: sessions, conversion rate, add-to-cart rate, reached checkout rate, checkout conversion rate, bounce rate, pageviews per session, online store visitors and searches. Shopify states that orders, sales and customer counts aren't changed. The rollout ran from September 21 to 23, 2026. The update can move session numbers either way: filtering identified bots may lower them, while counting previously missed journeys and no longer ending sessions at midnight UTC may raise or regroup them.
The Human or bot filter itself is older: Shopify added it on October 27, 2025 as a filter and a report dimension (Shopify changelog, Oct 27, 2025). Classification applies only to sessions from October 7, 2025 onward, and older sessions are not reclassified.
What the changelog does not say: it says nothing about Meta or any other ad platform. A widely shared Reddit thread read the update as proof that Meta sells bot clicks. Shopify's text doesn't support that reading; it describes a change to Shopify's own session counting.
Add-to-carts but no purchases: bots, tracking or the store?
Check three branches: traffic quality, tracking and the checkout. Start with bots because it is the quickest to rule in or out, but check all three, since bot activity, tracking errors and checkout problems can happen at the same time.
| Branch | Question to answer | Evidence to check | If yes |
|---|---|---|---|
| Bots | Is the odd cart activity concentrated in Bot sessions or one segment? | Shopify Human or bot session dimension on sessions and conversion reports; your Meta baseline by placement | Investigate the placement or country where it sits |
| Tracking | Does Meta report more AddToCart or Purchase events than Shopify records? | Meta events vs Shopify orders for the same days | Check event deduplication |
| Store | Do human shoppers reach checkout but not pay? | Shopify reached checkout rate vs orders, for Human sessions | Review checkout friction |
Bots. Shopify lists signs such as sudden spikes, $0.00 orders, odd customer names and sessions from data center locations (Shopify Help, Identifying bot activity). Some card-testing bots no longer create abandoned checkouts at all, so an empty abandoned-checkout list doesn't rule bots out (Shopify changelog, June 16, 2026).
Tracking. If the pixel and the Conversions API send the same event without matching deduplication keys, the action can be counted more than once (Meta for Developers, deduplication); our Meta Conversions API guide covers the setup.
The store. If humans reach checkout and the numbers agree but orders don't follow, the likely cause is checkout friction: shipping costs revealed late, forced account creation or missing payment methods. That is Layer 4 in our Facebook ads not working diagnostic.
What does Meta say about invalid clicks and charges?
Meta says it does not charge for clicks it determines to be invalid. Its help page describes two types of invalid clicks: clicks that don't show genuine interest or that show signs of ad testing, including repetitive or accidental clicks, and clicks generated through prohibited means such as fake accounts, bots, scrapers or browser add-ons (Meta, Detection and prevention of invalid clicks).
When Meta detects or is alerted to suspicious click activity, it says a manual review is performed, and "You will not be charged for clicks that are determined to be invalid." Meta also caps how often any ad is shown to the same person. That does not mean every suspicious click is classified as invalid. The page does not describe a refund form, a credit process, timelines or a report column showing invalid clicks, so check your own account's billing and support options rather than relying on figures quoted elsewhere.
What can I actually change in my campaigns?
If clicks rise without matching landing page views or purchases, first find the placement, device or country where the gap sits, then change that one setting and measure again. The steps below are practitioner advice, not Meta requirements.
- Consider excluding a placement that stays weak. Do it only after ruling out page-load problems and campaign changes, and after the drop shows up in comparable periods. For Awareness, Traffic and Engagement campaigns you can still uncheck placements manually. For Sales, Leads and App promotion, Meta is gradually moving exclusions to account controls, which cover Audience Network, Facebook Marketplace and the right column (Meta, What's new in placement settings). Our placements guide covers the trade-offs.
- Check the optimization event. If purchases are the goal, make sure the campaign optimizes for purchases rather than link clicks or landing page views. Click optimization tends to favor cheap visits, though switching doesn't by itself remove suspicious traffic.
- Review Advantage+ placements. If one placement's ratio collapsed, check whether delivery shifted toward it before excluding it everywhere.
- Tighten countries. If Shopify shows sessions from countries you don't target, review your location settings and the countries in your shipping zones.
What can't this diagnosis prove?
It can't prove that a specific click came from a bot. Ads Manager and Shopify Analytics show patterns, not the identity of a visitor. Proving that a given click was automated requires server-side logs and dedicated invalid-traffic or bot-management tools that inspect each request.
There are other limits too. Shopify's classifier is deliberately conservative ("better to miss some bots than incorrectly label real customers as bots"), so some bots will be counted as humans (Shopify Help, Bot filtering). Meta's landing page views may be modeled. And a Bot label in Shopify says nothing about where the bot came from.
If you need proof at the click level, the tool categories to look at are server-side log analysis, bot management at the CDN or firewall level, and independent ad-verification services. Treat percentages of "fake traffic" published by fraud-tool vendors as vendor claims, since they use their own definitions and have an interest in the result.
Where Nova fits
Running this check by hand once is manageable; remembering to run it every week across placements is where it usually slips. Nova, AdAdvisor's AI media buyer, checks your ad account every morning, and in Copilot mode nothing changes without your approval, so a placement exclusion waits for you with her reasoning. You can ask her to watch the click-to-landing-page ratio by placement as part of that review, and every action she takes is logged and can be rolled back (see how Nova works). Nova costs $199 a month per business, with a 7-day free trial (a card is required to start it). If you'd rather build your own alerting, our guide on how to monitor Meta ads with AI covers the setup.
Frequently asked questions
Summary
Clicks and add-to-carts without sales don't prove Facebook ads bot traffic. Compare link clicks with landing page views by placement against your own baseline, compare Shopify sessions by Human or bot using periods after September 23, 2026, and check bots, tracking and checkout, since more than one can apply. Shopify's September 2026 update changed session counting, not orders, and Meta says it doesn't charge for clicks it finds invalid. If you run a Shopify store and want this check in a daily review, with nothing changing until you approve it, Nova is built for that.
Sources
- Shopify Changelog, Shopify Analytics session measurement improvements (September 21, 2026)
- Shopify Changelog, Filter out bot traffic in your sessions related reports (October 27, 2025)
- Shopify Changelog, Reduced bot noise in abandoned checkouts (June 16, 2026)
- Shopify Help Center, Session measurement update
- Shopify Help Center, Bot filtering
- Shopify Help Center, Identifying bot activity on your online store
- Meta Business Help Center, About Meta's detection and prevention of invalid clicks
- Meta Business Help Center, Link clicks
- Meta Business Help Center, Landing page views
- Meta Business Help Center, Understand the difference between link clicks and landing page views
- Meta Business Help Center, What's new in placement settings in Meta Ads Manager
- Meta for Developers, Handling duplicate pixel and Conversions API events
Related reading

Performance Optimization
Facebook Ads Not Working? A 5-Layer Diagnostic to Find What's Actually Wrong
Facebook ads not working? Run this 5-layer diagnostic (Pixel, Audience, Creative, Offer, Budget) in the right order to find the real problem fast.
Read more
Analytics & Reporting
Meta Ads Attribution: Windows, Why Meta and GA4 Never Agree, and Which Number to Trust
Facebook attribution assigns credit by click and view windows, not last touch. Here's why Meta and GA4 always disagree, and which number to use for each decision.
Read more
Analytics & Reporting
Meta Conversions API (CAPI): Setup, Deduplication, and Why You Still Need the Pixel
Set up the Meta Conversions API correctly: Shopify, WooCommerce, and direct integration paths, plus deduplication and Event Match Quality benchmarks.
Read more
Strategy & Planning
Advantage+ Placements vs Manual Placements: Which to Use in 2026
What every Facebook and Instagram ad placement does in 2026, where each performs, and how to pick placements by objective instead of leaving it to guesswork.
Read more



