An ad account can look completely healthy and still be lying to you. Campaigns run, keywords match intent, dashboards show green — and underneath it all, Smart Bidding is optimizing toward a value signal that has quietly drifted away from what the business actually collects. Advertisers usually blame the platform when they notice the gap. The platform is not the problem. It is doing exactly what the conversion settings told it to do, which is chase a number that stopped being real months ago.

That gap has a name: value inflation. It’s what happens when the dollar figure feeding tCPA or tROAS diverges from the revenue the business actually recognizes. It rarely comes from fraud or a platform bug — it’s almost always self-inflicted, buried in tracking configuration nobody has revisited in a while. Here’s how to find it and fix it before it costs you more budget.

How value inflation builds up

Inflation is almost never one obvious error. It’s typically three or four small distortions stacked on top of each other, each nudging the algorithm a little further from reality:

• Double-counted micro-conversions — a form submit and a thank-you-page view both firing as separate conversions for the same lead, doubling reported volume for a single action.
• Mis-weighted goals — newsletter signups, PDF downloads or add-to-carts marked as primary conversions alongside actual purchases, diluting the bid signal with activity that has nothing to do with revenue.
• Inflated offline conversion imports — deal values uploaded from the CRM before a deal is won, or leads imported at full contract value when the real close rate is a fraction of that.
• Stale value rules — location, device or audience-based value multipliers built for a promotion that ended months ago, still running in the background.
• Feed drift — ecommerce feeds passing cart value instead of the actual post-discount, post-return order value.

Individually each of these looks like a rounding error. Stacked, they can inflate reported account value by 20% to 40% without triggering a single warning, because Google Ads has no way to know the number it’s being fed isn’t real — it just sees a target and works to hit it.

Why Smart Bidding can’t catch this on its own

Target ROAS bidding works backward from a value target: spend more where expected return clears the bar, less where it doesn’t. If a large share of “conversions” are inflated, the algorithm’s model of a good customer skews toward whatever behavior generated those inflated numbers — not toward the buyers actually worth acquiring. It bids up the placements and search terms producing noise, because noise is what’s being rewarded.

The learning period makes this worse. Smart Bidding recalibrates continuously on recent conversion data, so inflated inputs don’t just distort today’s bids — they train the model for weeks or months, depending on volume. By the time a CFO asks why customer acquisition cost is climbing while the platform reports healthy ROAS, the algorithm has already built an entire bidding strategy on a fiction.

A five-step audit for finding it

Run these in order — each step either confirms the data is clean or points straight at the fix. This is the same discipline behind the broader account diagnostics in the 5-pillar audit for Google Ads accounts, applied specifically to the value layer.

1. Audit conversion action weighting

Pull every active conversion action with its category, count, value and primary/secondary status. A clean setup usually has one to three primary actions tied directly to revenue — purchase, qualified lead, booked appointment. Everything else belongs in secondary, observation-only, or excluded from bidding entirely.

2. Check the attribution model

Confirm every conversion action runs data-driven attribution rather than defaulting to last-click on an older action that predates the account’s migration. Mixed attribution models across actions in the same account produce inconsistent value patterns for conversions that should behave the same way — a common oversight in inherited accounts after Google’s forced DDA transition.

3. Hunt for duplicate tags

Fire a real test conversion through the funnel and watch Tag Diagnostics or GA4’s DebugView. Look for the same event firing twice — once from a hard-coded gtag snippet still sitting in the page code and once from Google Tag Manager, or a confirmation-page load firing independently of the actual submission event it’s meant to represent.

4. Reconcile value against actual pipeline

Take the last 90 days of Google Ads-attributed conversion value and put it next to closed revenue or fulfilled orders from the CRM or order system for the same window. If Ads reports $400,000 in conversion value against $270,000 in matching revenue, that gap is worth chasing — and it’s rarely spread evenly. Usually one or two conversion actions, often an offline import or a lead-value feed, are carrying nearly all of the distortion.

5. Cross-check in GA4

Compare Google Ads-reported conversions against GA4’s own purchase or key-event counts for the same campaigns and date range. A large, consistent gap signals that value is inflating somewhere between the click and the conversion firing — worth investigating rather than dismissing as a tracking quirk.

Building the paper trail

Before changing a single setting, document exactly where the inflation is coming from. Pull the conversion value rules report to see adjusted value by rule, review conversion action diagnostics for unusual value patterns, segment conversions by action instead of relying on one blended number, run the GA4 comparisons that don’t inherit any of Google Ads’ attribution assumptions, and export CRM or order data — the only source on this list that documents what actually happened.

Rebuilding a clean value signal

Finding the inflation is half the job; the fix has to prevent it from creeping back.

Move every non-revenue action — content downloads, video engagement, chat starts, account signups — out of primary status and into secondary or observation. The bid strategy should only optimize toward actions representing real, recognized value. If a lead needs qualifying before it’s worth anything, don’t let raw lead volume drive bids on its own; import it as an offline conversion once qualified, weighted appropriately.

For offline conversion tracking, import value at the stage that reflects reality rather than the stage that looks best — either a probability-weighted value, or the full value only once a deal actually closes, with adjustments made for deals that fall through. Google has moved new offline conversion imports to the Data Manager API, so if that pipeline hasn’t been touched recently, this is the moment to rebuild it rather than patch the old one. For ecommerce, confirm the dynamic value feed passes net order value — post-discount, post-tax, adjusted for returns — not gross cart value at checkout.

Finally, audit every active conversion value rule and ask whether the business condition that justified it still holds. A regional promotion that ended, a device test that wrapped up, an audience segment no longer being targeted — these accumulate quietly. Since Google now requires a conversion value on new conversion actions, check for placeholder values nobody went back to fix.

Recalibrating targets after the clean-up

This is where careful audits often go sideways. Once the data is fixed, reported average order value drops because it’s now accurate — and if the tROAS target doesn’t move with it, the mismatch can trigger a bidding shock that tanks impression share overnight.

• Recalculate the real target before touching the account. If the inflated blended ROAS was 380% and the clean number is 310%, set the target to 310%, not 380% with hope attached.
• Move the target in stages — a 15% to 20% adjustment every five to seven days — rather than one jump, giving Smart Bidding room to recalibrate without a full learning-period reset.
• Expect a temporary dip in reported conversion volume. That’s the inflation leaving the system, not the campaign failing. Track it against real revenue, not the dashboard.
• Hold for at least two full weeks post-cleanup before making further target adjustments.
• Document the before-and-after gap for stakeholders, since a rising CPA right after a “clean-up” invites questions that the pipeline reconciliation from step four already answers.

This kind of discipline matters more as bidding automation keeps expanding — see how the landscape is shifting in where to spend your time on Google Ads bidding strategies in 2026 and how Google has responded to advertiser pushback in Google’s clarification of a recent Smart Bidding update.

Frequently asked questions

What is value inflation in Google Ads?

It’s a gap between what an account reports as conversion value and what the business actually collects in revenue, usually caused by double-counted conversions, mis-weighted goals, inflated offline imports or stale value rules.

How often should this audit run?

Quarterly is a reasonable default, not just when performance feels off. Value inflation compounds quietly — the same pattern worth watching for in how budget misallocation hides in plain sight.

Will fixing value inflation hurt reported performance?

Temporarily, yes. Reported conversion volume and ROAS typically dip once inflated data is removed. That’s the correction taking effect, not a new problem — measure against actual revenue during the adjustment window.

Can Smart Bidding detect inflated values on its own?

No. tCPA and tROAS treat whatever value they’re fed as ground truth. They have no mechanism to distinguish a real conversion from an inflated one; that verification has to happen outside the platform.