What Is PPC Value Inflation and How to Fix It Before AI Smart Bidding Learns the Wrong Thing?

Smart Bidding is the most powerful bidding engine in the history of paid search. It is also completely blind to whether the numbers you’re feeding it are real. Most accounts are feeding it fiction and paying higher CPCs to get there.

Here’s what makes value inflation uniquely dangerous compared to every other Google Ads problem.

A broken keyword strategy gets visible. A bad landing page shows up in conversion rate data. A mismanaged bid cap creates obvious impression share loss. These problems surface. You can see them.

PPC Value inflation is invisible. On the surface, the account looks healthy. Campaigns are running. ROAS meets target. Conversion volumes look strong. Leadership nods at the monthly report. Meanwhile, Smart Bidding is quietly being trained on conversion data that doesn’t reflect what the business actually collects and every day it operates on that data, it gets better at chasing the wrong thing.

Smart Bidding is an AI. It learns. It builds a model of what a high-value customer looks like based on the data you give it. Feed it inflated data, and it builds an inflated model on which it bids aggressively for audiences, placements, and queries that correlate with your phantom conversions, not your actual revenue.

The longer it runs on bad data, the more precisely it optimizes toward the wrong outcome. You’re not just getting bad reporting. You’re actively training an AI to spend your budget incorrectly and the AI is getting better at it every single day.

Sarah Stemen, PPC expert and owner of Sarah Stemen LLC, outlined the exact anatomy of this problem and how to fix it in a recent Search Engine Journal piece. The technical framework is solid and worth acting on immediately. Here’s the full breakdown with the strategic context the original piece doesn’t have.

What Value Inflation Actually Is (And Why It’s Always Self-Inflicted)

Value inflation is the gap between what Google Ads reports as conversion value and what the business actually collects in revenue. It’s not fraud. It’s not the platform cheating. It’s almost always a self-inflicted tracking configuration problem, usually several small distortions stacked on top of each other, none of which trigger an alert, all of which quietly corrupt the signal Smart Bidding is optimizing toward.

The four most common sources, each one capable of inflating reported account value by 20-40% on its own:

Double-counted micro-conversions. A “form submit” and a “thank you page view” both firing as separate conversion actions, each carrying its own value, for the same lead. The business got one lead. Google Ads counted two conversions. Smart Bidding built a model around a world where form submissions generate twice the actual volume.

Mis-weighted primary vs. secondary goals. Newsletter signups, PDF downloads, and add-to-carts marked as primary conversions sitting alongside actual purchases. The bid algorithm can’t distinguish between “someone subscribed to our newsletter” and “someone paid us money” if both are set to primary. It optimizes toward volume of the noise, not quality of the signal.

Inflated offline conversion imports. Deal values uploaded from the CRM at the pipeline stage rather than the close stage. If your close rate on sales-qualified leads is 20%, importing them at full contract value tells Smart Bidding that every lead is worth five times what it actually is. The algorithm will bid accordingly spending to acquire leads at a cost that only makes sense if all of them close.

Stale conversion value rules. Location, device, or audience-based multipliers built for a campaign or promotion that ended eight months ago, still running in the background, still inflating values on every conversion that matches those criteria. These accumulate in accounts over time. Nobody removes them because nobody remembers they’re there.

Stack two or three of these together, which is the norm in accounts that haven’t been audited recently and the reported conversion value can run 40-60% above actual revenue. Smart Bidding sees 40-60% more value than exists. It bids 40-60% more aggressively. CPCs climb. Budget burns. The CFO starts asking why CAC is rising while the platform shows strong ROAS.

The answer is that ROAS is measuring something that isn’t revenue.

The Learning Period Problem: Why This Gets Worse Over Time, Not Better

Standard tracking errors are annoying. Value inflation compounds.

Smart Bidding recalibrates continuously based on recent conversion data. It doesn’t have a fixed model, it has a learning loop. New data constantly updates its understanding of which users, queries, and placements generate high-value outcomes.

Inflated conversion values don’t just distort today’s bidding. They train the model. Over weeks and months, the algorithm builds an increasingly precise picture of “the kind of customer that generates high conversion value for this account, except that picture is drawn from inflated data, so it’s a picture of the kind of customer that triggers your phantom conversions, not the kind that actually generates revenue.

By the time the discrepancy becomes visible, usually when a CFO or agency does a proper pipeline reconciliation and finds that Google Ads is claiming £400,000 in conversion value against £270,000 in actual recognized revenue, Smart Bidding has been trained on bad data for months. The model is wrong. The bidding strategy built on that model is wrong. And the learning period required to recalibrate after the fix is measured in weeks.

This is the compounding cost of value inflation: not just overpaying for bad clicks today, but paying to train an AI to get better at overpaying for bad clicks every day the problem persists.

The Five-Step Audit: Finding Exactly Where the PPC Value Inflation Lives

Run these in order. Each step either confirms a section of the data is clean or pinpoints exactly where the problem is.

Step 1: Audit Your Conversion Action Weighting

Go to Goals → Conversions → Summary. Pull every active conversion action with its category, count, value, and primary/secondary designation.

You’re looking for two specific problems: conversion actions marked primary that have no direct revenue relationship, and value-based goals sitting alongside count-based goals with no weighting differentiation.

A clean conversion setup has one to three primary conversion actions tied directly to revenue, purchase, qualified lead, booked appointment. Everything else belongs in secondary or observation status. If your primary conversion list includes newsletter signups, chat initiations, video views, or PDF downloads alongside actual purchases, the bid signal is being diluted with noise.

Fix this before anything else. Primary vs. secondary designation is the most fundamental input into Smart Bidding’s objective function. If it’s wrong, everything downstream is wrong.

Step 2: Check Attribution Model Consistency

Confirm every primary conversion action is running data-driven attribution. Inherited accounts frequently have a mix, some actions on DDA, older actions still on last-click that nobody updated when Google made the transition. Mixed attribution models across conversion actions produce wildly inconsistent value patterns for comparable conversions, and Smart Bidding interprets that inconsistency as signal rather than noise.

This is one of the most commonly missed issues in accounts that haven’t been systematically reviewed since 2022-2023.

Step 3: Hunt for Duplicate Tag Firing

Open Google Tag Manager’s Preview mode or GA4’s DebugView and fire a real test conversion through the funnel. Watch for the same conversion event firing twice, once from a hard-coded g tag snippet in the page code and once from GTM, or a “confirmation page view” firing independently of the form submission event it’s supposed to represent.

A single duplicate tag can double reported conversion volume without a single error message appearing in the interface. Google Ads has no mechanism to detect it which just sees two conversion events and records both.

Step 4: Reconcile Conversion Value to Actual Revenue

This is the step that surfaces the scale of the problem. Pull 90 days of Google Ads-attributed conversion value. Put it next to actual closed revenue or fulfilled orders from your CRM or order management system for the same window.

If Google Ads reports £400K and the business recognized £270K, you have a £130K inflation gap to explain and fix. Break the gap down by conversion action, the inflation is almost never evenly distributed. Usually one or two actions, typically offline conversion imports or a lead-value feed, are carrying the majority of the distortion.

This step is the one most teams skip because it requires pulling data from outside the ad platform. That’s exactly why the problem persists for months without being caught.

Step 5: Cross-Reference in GA4

Use GA4’s Advertising snapshot and a custom exploration comparing Google Ads-reported conversions against GA4’s own purchase or key event counts for the same campaigns and date range.

A large, consistent gap between what Google Ads claims it drove and what GA4 recorded as an actual purchase event is a red flag. Not a tracking discrepancy to dismiss, a signal that value is being inflated between the click and the conversion action firing.

The Six Fixes: Rebuilding a Clean Value Signal

Finding the inflation is half the job. The other half is rebuilding the value structure so it doesn’t creep back.

Fix 1: Set Primary vs. Secondary Correctly And Keep It That Way

Move every non-revenue action out of primary status. Newsletter signups, content downloads, chat starts, account signups, secondary or observation only. Document the intended conversion hierarchy in writing so the next person who touches the account doesn’t accidentally re-add a micro-conversion to primary status six months from now.

Fix 2: Fix Your Offline Conversion Import Logic

Import at the stage that reflects reality. If close rate on sales-qualified leads is 20%, either import at probability-weighted value 20% of contract value at SQL stage or import the full value only at deal close, with conversion adjustments for deals that subsequently fall through.

Note: Google Ads recently sunset new offline conversion imports through the legacy API in favour of the Data Manager API. If your OCT pipeline hasn’t been rebuilt for the new infrastructure, this is the moment to rebuild it correctly rather than patching the old one.

Fix 3: Fix Your E-commerce Value Feed

Confirm the dynamic value feed is passing net order value post-discount, post-tax where applicable, adjusted for returns via conversion adjustments, not gross cart value at checkout initiation. The difference between cart initiation value and net order value can be 15-30% in categories with high discount rates or return volumes. Smart Bidding optimizing toward cart value is bidding for revenue that frequently doesn’t materialize.

Fix 4: Audit and Retire Stale Value Rules

Pull every active conversion value rule and ask whether the business condition that justified it is still true. Location rules built around regional promotions, device rules from mobile tests that ended, audience rules for segments no longer being targeted, these accumulate silently. Retire anything that doesn’t reflect current business reality.

Fix 5: Remove or Merge Duplicate Tags

Once identified in Step 3, remove the hard-coded gtag snippet where GTM is firing the same event, or consolidate duplicate conversion actions into a single action. Document what was removed and why so the fix isn’t accidentally reversed during a future site update.

Fix 6: Align GA4 and Google Ads Conversion Definitions

Where GA4 key events and Google Ads conversion actions are meant to track the same behavior, confirm they’re using identical trigger conditions and value logic. Divergence between the two isn’t just a reporting problem, it prevents you from using GA4 as a reliable cross-reference to catch future inflation before it compounds.

Recalibrating tROAS After the Clean-Up: The Step That Breaks Most Accounts

This is where careful audits go wrong. When the data is fixed and inflated values disappear from reporting, the account’s apparent average conversion value drops because it’s now accurate. If the tROAS target doesn’t move with it, the algorithm enters bidding shock: it’s suddenly asked to hit a 380% ROAS target on data that realistically supports 310%, and it responds by pulling back spend dramatically, cratering impression share.

The recalibration protocol that prevents this:

Step one: Recalculate the real target before touching anything in the account. Use the pipeline reconciliation from Step 4 to establish what ROAS the account was actually generating on real revenue. If the true blended ROAS was 310% on clean data, set the target to 310%, not 380% with the hope of recovering to where the fiction said you were.

Step two: Move the target in stages, not one jump. Adjust tROAS by 15-20% every five to seven days. This gives Smart Bidding room to recalibrate without triggering a full learning-period reset. A single 70-point jump from 380% to 310% will cause a bidding shock that looks like campaign failure. A staged adjustment looks like recalibration.

Step three: Expect a temporary dip in reported conversion volume. That’s the inflation leaving the system, not the campaign failing. Measure performance against actual revenue for the 30 days post-cleanup, not against the platform’s conversion dashboard.

Step four: Hold for two full weeks before making a second round of adjustments. Judging a newly recalibrated bid strategy against a partial data set reintroduces the guessing problem the audit was designed to fix.

Step five: Brief stakeholders before the dip happens, not after. When leadership sees CPA rise and conversion volume fall immediately following a “clean-up,” they will ask hard questions. Have the pipeline reconciliation numbers ready and the explanation prepared before the first post-cleanup report lands.

The Frequency Question: When to Run This Audit

Value inflation isn’t a one-time problem with a permanent fix. Tracking configurations drift. New campaigns introduce new conversion actions. CRM integrations get updated without rechecking value import logic. Platform changes deprecate old tag implementations.

Run the full five-step audit quarterly, minimum. Set a calendar reminder. Make it a standard part of every quarterly performance review, not an emergency procedure for when the CFO notices the revenue gap.

The earlier in the drift cycle you catch it, the less training data the algorithm has built on bad inputs, and the smaller the recalibration required. An inflation problem caught at eight weeks is a two-week fix. An inflation problem that’s been running for twelve months requires months of recalibration and a difficult conversation about why nobody noticed.

Quarterly audits are the difference between those two scenarios.

The Strategic Frame: Clean Data Is Your Competitive Advantage

Here’s the angle that turns this from an operational housekeeping task into a genuine competitive differentiator.

Every advertiser in your auction is using Smart Bidding. The ones with clean, accurate conversion data are training the algorithm on reality. The ones with value inflation are training it on fiction. Over months and years, those two groups of accounts diverge significantly in bidding precision, one set gets better at finding real customers, the other gets better at finding phantom ones.

In a tROAS environment where the algorithm’s intelligence is only as good as the data feeding it, data quality is strategy. A competitor with 10% lower budget but accurate conversion data will outperform an account with 40% higher budget running on inflated signals because their algorithm is learning from truth and bidding accordingly, while the inflated account is spending to chase ghosts.

Clean data doesn’t just improve your reporting. It improves your algorithm. And a better-trained algorithm compounds its advantage every day it runs bidding smarter, wasting less, and finding the actual high-value customers your business needs.

That’s the real cost of value inflation: not the wasted spend today, but the compounding performance gap between your account and a competitor who cleaned up their data before you did.

The Quick-Reference Checklist

For the PPC teams who will use this as an operational document:

Audit checklist (run quarterly):

Primary conversion actions tied only to revenue events
Attribution model consistent across all primary actions (data-driven)
No duplicate tag firing in GTM Preview or GA4 DebugView
90-day conversion value reconciled against CRM/OMS actual revenue
GA4 conversion counts cross-referenced against Google Ads reports
All conversion value rules reviewed against current business conditions
Offline conversion import logic checked against current close rate reality
E-commerce feed passing net order value, not gross cart value

Fix sequence:

Primary/secondary designation
Duplicate tags
Value rules cleanup
OCT import logic
E-commerce feed value
tROAS target recalculation and staged adjustment

Stakeholder communication:

Brief before the post-cleanup dip, not after
Show pipeline reconciliation as the proof
Frame as precision improvement, not performance decline

The Bottom Line

Smart Bidding is the most powerful algorithmic bidding tool in paid search history. It is also completely unable to distinguish between real conversion value and inflated conversion value. That distinction is entirely your responsibility and most accounts are failing at it.

The brands that take conversion data quality seriously aren’t just improving their reporting. They’re training a more precise algorithm, making better bidding decisions, and compounding that precision advantage every week the algorithm runs on clean data.

The ones still running on inflated signals are paying more for worse outcomes, building an increasingly distorted model of what their ideal customer looks like, and widening the performance gap between themselves and competitors who already ran this audit.

Clean data is the prerequisite for everything Smart Bidding promises. Run the audit. Fix the foundation. Then let the algorithm do what it’s actually good at.

Source: Sarah Stemen, “The PPC Clean-Up: How to Audit and Fix ‘Value Inflation’ in Google Ads

Share you part of the story on how you have successfully fixed the Value Inflation or If You liked our content, pls Comment Below!