If Google Ads shows more conversions than Shopify shows orders, modeled conversions are usually part of why. This guide explains exactly what they are, how much they can inflate your reported count, and — the part most articles skip — what they do to your profit math when you use that count to set bids.
What are Google Ads modeled conversions?
Google splits every conversion into two buckets. Observed conversions use cookies and identifiers to link an ad click directly to a sale. Modeled conversions use machine learning "to assign links between ad interactions and conversions accounting for cases where cookies and identifiers weren't available," per Google Ads Help.
The mechanism is straightforward. Google takes the conversions it can see, groups them by shared traits like device, location, browser, and time of day, then applies those observed conversion rates to the ad interactions it cannot see. Where the pattern is strong enough, it estimates that a conversion occurred and reports it.
Crucially, modeled and observed conversions sit in the same Conversions column with the same attribution paths and values. You cannot tell them apart at a glance. Google says it only includes a modeled conversion "when there's a high confidence that your ad resulted in conversions," and that it won't model at all for accounts with too little volume, according to Google Ads Help.
Why Google models conversions at all
Modeling exists because direct tracking broke. Cookie-consent declines, Safari and Firefox tracking prevention, ad blockers, and cross-device journeys all sever the clean click-to-sale link Google used to rely on. Field estimates put blocker- and consent-affected traffic at roughly ten to twenty-five percent of users.
Rather than report a number it knows is undercounted, Google fills the gap with an estimate. Google's own early data claimed that conversion modeling through Consent Mode recovers "more than 70% of ad-click-to-conversion journeys lost due to user cookie consent choices" — though that figure dates to 2021 and varies widely by store.
So modeling is not a scam. It is Google answering a different question than Shopify. Google is estimating "how many sales did my ads plausibly influence?" while Shopify records "how many orders actually completed?" Both can be internally correct and still disagree. If you want the full picture of why every platform reports a different number, our guide to reconciling your ecommerce data maps all of them side by side.
How much do modeled conversions inflate your reported count?
There is no single published inflation rate for Google — it depends on your consent rates and setup. But the direction is always the same: platform-reported conversions run ahead of server-side orders, and modeling is one of the reasons.
For a sister platform, the gap is better documented. On Meta's default window, a twenty to thirty-five percent gap between platform-reported purchases and Shopify orders is considered normal, and modeled conversions plus view-through are the main drivers. Google's modeling is more conservative than Meta's view-through counting, but the same forces apply. The related consent mode impact on conversions walks through how much your accept rate alone can move the number.
Two more traits make modeled conversions slippery to reconcile:
- They lag. Modeled conversions "can take up to 5 days to fully process and stabilize," per Google Ads Help, so yesterday's number will keep changing.
- They don't reverse refunds. Google generally keeps the original conversion even after the customer refunds, so its total stays high while Shopify's net sales drop.
Modeled conversions vs. Shopify orders: a worked example
Numbers make this concrete. Say you sell a print-on-demand mug and run Google Ads for one week.
Your Google Ads account reports 120 conversions. Your Shopify admin records 100 real paid orders, and Shopify's last-click attribution credits 70 of those to Google. The other 50 of Google's reported conversions come from modeling, cross-device stitching, and view-through-style credit that no Shopify session ties back to a Google click.
Now price it out. Your mug sells for $40 plus $5 shipping. Your Printify base cost plus the shipping you pay is $18 per order. On a plan that charges roughly 2.9% + 30¢ per transaction, the fee on a $45 charge is:
$45 × 0.029 + $0.30 = $1.61
So your gross profit per order, before ad spend, is:
$40 − $18 − $1.61 = $20.39
You spent $2,000 on Google Ads that week. Here is where the modeled count bites:
- Judging by Google's 120 conversions: CPA = $2,000 ÷ 120 = $16.67. Profit per order = $20.39 − $16.67 = +$3.72. Looks like a winner.
- Judging by the 70 real Google-attributable orders: CPA = $2,000 ÷ 70 = $28.57. Profit per order = $20.39 − $28.57 = −$8.18. You are losing money on every sale.
Same campaign, same week, same bank deposit — but the metric you trust flips the campaign from profitable to underwater. That is the risk of feeding a modeled, inflated conversion count into your bid decisions.
The part every guide skips: what modeling does to your profit
Most articles stop at "the numbers won't match, and that's okay." For a store owner spending real money, that is not okay enough. Because Smart Bidding and Target CPA optimize toward the conversion count Google reports, an inflated count quietly pushes you to bid more per real sale than you think you are.
The fix is not to distrust Google — modeled conversions are a reasonable estimate of real, unobservable demand. The fix is to stop using platform-reported conversions as your profit yardstick. Reconcile them against the one number that never models: your actual Shopify order and payout data, then measure ad performance as profit per real order, not conversions per dollar.
This is also why blindly trusting any single platform's count is dangerous. Meta has the same problem from the other direction — see missing purchase events on Facebook and how to verify your Facebook Pixel is tracking purchases before you compare anything.
How to work with modeled conversions instead of against them
You cannot turn modeling off, and you mostly shouldn't want to. But you can stop letting it distort decisions:
- Compare on trailing windows, never single days. Modeled conversions restate for up to five days, so a 7-to-14-day view is the only fair comparison against Shopify orders.
- Anchor to Shopify order count and payout, not the Conversions column. Shopify's server-side record is the source of truth for how many sales happened and how much cash landed.
- Watch the ratio, not the equality. A stable Google-to-Shopify gap is healthy; a sudden jump signals a real tracking break — like a broken tag or a consent change — worth investigating.
- Improve your observed rate. Enhanced Conversions and a clean Consent Mode setup shrink how much Google has to model in the first place, which makes the modeled slice smaller and more accurate.
Where PodVector fits
This reconciliation is exactly what PodVector is built for. PodVector connects your Shopify, Meta Ads, Google Ads, Printify, and Printful accounts and computes your true per-order profit from real Shopify orders — the number no platform models.
Victor, PodVector's AI employee, reads your Google Ads and Meta data and your real order-level costs together, so a campaign that looks profitable on a modeled CPA but loses money on real orders gets flagged for what it is. Victor does not touch your ad account — he reads ad data, proposes the move, and only executes writes on the Shopify side with your approval. He is not a dashboard; he is an employee that analyzes your reconciled data and acts on it.
If you're consolidating your storefront for cleaner data at the source, our guide to migrating from Etsy to Shopify covers the move.
FAQs
Are Google Ads modeled conversions fake?
No. They are statistical estimates of real conversions Google couldn't observe directly — because of cookie consent declines, ad blockers, or cross-device journeys. Google only includes them "when there's a high confidence that your ad resulted in conversions," according to Google Ads Help. They can over- or under-shoot, but they are an estimate of something real, not invented demand.
Why does Google Ads show more conversions than Shopify shows orders?
Several reasons stack up: modeled conversions, cross-device attribution, and Google crediting clicks it saw even when Shopify's last-click model assigns the order elsewhere. Google also reports conversions on the click date and doesn't remove refunds, while Shopify records the completed order on the purchase date and reduces sales on refund. A persistent gap is normal; compare on a rolling window, not day by day.
Can I turn off modeled conversions in Google Ads?
Not directly — modeling is built into how Google reports conversions when direct observation fails. What you can do is reduce how much Google needs to model by implementing Enhanced Conversions and a proper Consent Mode setup, which raises your observed rate. You can also lean on server-side order data as your profit benchmark instead of the reported Conversions column.
How long do modeled conversions take to appear?
They can take up to five days to fully process and stabilize in reporting, per Google Ads Help. This is why recent-day conversion numbers keep rising after the fact, and why you should never judge a campaign on yesterday's count alone.
Should I trust modeled conversions for bidding?
Trust them as a directional signal of ad influence, not as your profit metric. Smart Bidding optimizes toward the reported count, so an inflated count can push your real cost-per-sale higher than you realize. Reconcile against real Shopify orders and payouts, and judge campaigns on profit per real order — the worked example above shows how a modeled CPA can flip a losing campaign into a false winner.