Your Ads Dashboard Says $400K. Your Bank Says $270K.

A Google Ads dashboard says the account returned $400,000. The books say $270,000 came in. Those are real figures from a paid search audit published this month, and nothing in that account is broken. Both numbers were produced by systems doing exactly what they were configured to do.

Reported conversion value is a claim about worth, not a record of cash. When those two drift apart, you do not just get a reporting inconvenience. You get a bidding system spending real budget to buy more customers at a price nobody ever collected, and a budget approval built on money that was never in the account.

Most articles on this question will tell you why Google Ads and GA4 disagree with each other. That is a real answer to a different question. The gap between reported value and the cash actually banked runs 20% to 40% wide as a separate problem from attribution, Smart Bidding treats the inflated figure as its goal, and closing it takes a specific reconciliation. That is what I have spent ten years of Google Ads work untangling across 300+ businesses in medical, ecommerce and local services, and it is what this article covers.

Your ROAS Is Funding a Number You Never Banked

Conversion value is not measured. It is assigned. Every conversion action in the account carries a value that either comes from your website's tracking code, a static number someone typed in during setup, a rule that multiplies it, or an import from your CRM. Add all of those together and the platform prints a total. That total is only as honest as the weakest assumption feeding it.

Sarah Stemen, a paid search marketer with more than 17 years in paid search, wrote in Search Engine Journal on August 13, 2026 that stacked tracking errors commonly inflate an account's reported conversion value by 20% to 40% above the revenue actually collected, with no alert and no obvious symptom. On an account reporting $400,000, that is roughly $67,000 to $114,000 of value that never reached the bank.

An inflated value figure does not sit quietly in a report. It sets your target ROAS, which sets how aggressively the platform bids, which sets what you spend. It justifies next quarter's budget in a meeting where nobody has the bank statement open. It also decides which campaigns get killed, because a campaign whose value is inflated by a stale rule will always look better than the one selling your actual best product at honest numbers.

Picture your own business. A home goods store runs Shopping campaigns, the dashboard shows $400,000 in conversion value against the spend, and the owner signs off on a bigger budget for Q4. The accountant recognizes $270,000 for the same period. The $130,000 difference is about a third of everything the platform reported, and every decision made off that dashboard was priced as if the money existed. That example sits above the 20% to 40% band, which is what a bad case looks like once several of these errors stack in one account.

Attribution Explains the GA4 Gap, Not the Bank Gap

Search this problem and you will land on a solid, correct explanation of why the two reporting platforms disagree. McCord Web Services covers it well: Google Ads counts clicks while GA4 counts sessions, Ads credits a conversion back to the click date while GA4 logs it on the day it happened, the two use different attribution models, and tag settings, currency configuration and whether tax and shipping are included all pull the numbers apart.

Every word of that is true. It explains why two dashboards disagree with each other. It does not explain why the dashboard disagrees with your bank, and reconciling Ads against GA4 will not catch this problem, because both platforms are usually fed by the same tag, sending the same value, carrying the same mistakes. Two systems inheriting one bad number will agree with each other beautifully all the way to the wrong conclusion.

This is also a different problem from whether your tracking fires correctly in the first place. I have written separately about the accounts where conversion tracking is miscounting or missing conversions entirely, which is the question of whether the counting works at all. What follows here assumes the counting works fine and asks what the counted number is actually worth.

That sentence is the whole risk in seven words. The bidding system optimizes toward the value you hand it. It has no way to know that $130,000 of that value was double counted, estimated, or already refunded, and it will never come back and tell you the goal was wrong. It just gets better and better at hitting a target that does not exist, which is a fair description of what automated bidding does to a budget when nobody is checking its inputs.

Five Ways the Value Inflates While the Tracking Works Fine

None of these look like errors on a normal Monday review. They look like a well-configured account. Stemen's analysis names five, and I see the same five, often two or three at once in the same account, each adding a few percent until the total is unrecognizable.

Double-counted micro-conversions come first because they are the most common. Somebody set up "viewed the pricing page" or "clicked the phone button" as a conversion with a value attached, and it stayed in the primary column. One customer now generates value twice, once for looking and once for buying, and the account looks like it produced two sales.

Mis-weighted goals are the same failure with better manners. Google separates primary goals, which bidding optimizes toward, from secondary goals, which are for observation only. Secondary actions leak into the primary column during a routine settings change, and nobody notices because the report just gets healthier.

Offline imports are where lead-gen accounts inflate. The CRM sends a value back to Google when a lead is created or an opportunity is opened, priced at pipeline value or an average deal estimate rather than money that closed. A pipeline is a hope. Note for anyone rebuilding this plumbing: Google has sunset new offline conversion imports through the legacy Ads API in favor of the Data Manager API, so that job is not a quick copy of the old setup.

Stale value rules deserve a paragraph of their own, because they are invisible from the main reporting screen. A value rule multiplies or overrides conversion value based on device, location or audience, and it was created for a reason that expired. Google reports an "original conversion value" metric showing what the conversion was worth before rules were applied, and Dataslayer's 2025 explainer of that metric is the clearest description of what the two columns mean. If those two totals differ by a wide margin, a rule is writing your revenue for you.

Gross versus net is the one that hurts ecommerce most. The platform books the full cart at checkout. Refunds happen three weeks later, discount codes come off the top, tax and shipping were never yours, and none of that flows back. In a category with a high return rate, that single gap can carry most of the inflation on its own.

The Fix Starts With Your Bank Statement, Not Your Tags

The instinct is to go fix the tags. Do not start there. Start by finding out how big the lie is, because the size of the gap decides whether this is a Tuesday afternoon cleanup or a quarter's worth of budget being misallocated, and because you cannot reset a target without knowing your true number.

Pull 90 days. Put Google Ads reported conversion value next to recognized revenue from your accounting system or your CRM for the same window, not GA4, which can be inflated by the exact same errors. In Google Ads, set the date range to the last 90 days and read the Conv. value column; in your accounting system, pull recognized revenue for the same window. Divide the Google Ads number by the banked number. In the worked example, 400,000 divided by 270,000 is 1.48, so the dashboard is claiming 48% more than the business collected. Everything after that is subtraction.

Step five is the one people skip, and skipping it is how a correct fix turns into a bad month. Slam the target from 380% down to 310% overnight and the bidding system treats it as a new problem, pulls back hard while it relearns, and volume drops in a way that gets the whole project blamed. Give it five to seven days per move. Deriving that number properly is its own piece of work, and I have laid out how to calculate a target ROAS from your actual margins rather than picking one that sounds ambitious separately.

If an agency runs the account, you do not need to learn any of this to hold them to it. Ask two questions. Show me every conversion action currently in the primary column and what value each one carries. Then show me our reported conversion value next to the revenue we recognized last quarter, and explain the difference. A good partner will already have that spreadsheet. Anyone who answers by defending the ROAS chart has told you which number they manage.

Navah Hopkins was making an adjacent argument in the same publication a week earlier, writing on August 6 that AI-era PPC campaign structures depend on clean value signals.

The First Week After the Reset Will Look Worse. That Is the Fix Working.

Expect the dip and hold your nerve. You removed the fake value before the real value had a chance to be measured against a corrected target, so bidding is recalibrating, and judging the reset on day three is how good accounts get reverted back to comfortable lies.

Make the reconciliation a monthly habit, not a one-time project. Same two columns, same subtraction, one number that matters: how close reported value sits to banked revenue. Watch it drift back up and you have found a new value rule, a new goal someone promoted, or a returns problem the marketing team never sees.

The vanity trap that catches owners hardest is conversion count. More conversions on a screen feels like progress, and after this cleanup you will have fewer, because the page views stopped pretending to be sales. Fewer, honest conversions tied to money you can spend is the win.

Frequently Asked Questions

Why does Google Ads report more conversion value than my actual revenue?

Because conversion value is assigned by your setup, not read from your bank account. Small errors stack: a micro-conversion like a pricing page view carries a value alongside the purchase it led to, a secondary goal ends up in the primary column, a CRM sends pipeline estimates instead of closed deals, an old value rule keeps multiplying, and cart value is reported gross with refunds, discounts, tax and shipping never subtracted. Sarah Stemen wrote in Search Engine Journal on August 13, 2026 that this combination commonly inflates reported value by 20% to 40% above collected revenue with no alert. That is her practitioner range from audits rather than a controlled study, so use it as the size of the problem to look for.

Should I use GA4 to check whether my Google Ads revenue is real?

No, and this is the most common wasted week I see. Ads and GA4 disagree for their own structural reasons, including different attribution models, clicks versus sessions, and whether a conversion is credited to the click date or the day it happened. Those differences are real, but both platforms are fed by the same tag sending the same value, so both inherit the same inflated numbers. Reconcile against your accounting system or your CRM, because those hold money that actually arrived.

Will lowering my target ROAS crash my sales volume?

It will if you do it in one move. Smart Bidding treats a large target change as a new problem and pulls back sharply while it relearns, which shows up as a volume drop within days. Stemen's recalibration example is worth copying: if the inflated data showed 380% but your true blended ROAS is 310%, set 310% as the destination, cut no more than 15% to 20% at a time, and give each cut five to seven days before judging it. That specific move is a single step; a bigger gap takes three or four. Calculate the destination from your own margins rather than from what the dashboard was reporting.

Nothing in this fix requires new software, a new agency, or a bigger budget. It requires one spreadsheet with two columns and the willingness to find out that a number you have been reporting for a year was never yours. In ten years of auditing ad accounts, I rarely see this comparison already built when I start, which means the ROAS people trust most is often the one figure in the business that has never been checked against reality.

If you want a second set of eyes on the gap between what your account reports and what you actually collected last quarter, book a call and bring both numbers. That conversation gets interesting after one look at the primary conversion column.