Hi everyone, this is Neo.

Have you ever had this experience with Google Ads: nothing changed in your account, but Smart Bidding keeps getting worse — CPAs creeping up, ROAS sliding down?

Most people’s first reaction is to switch bid strategies, lower targets, or argue about primary versus secondary conversions. But a Search Engine Journal piece from August 20 asks a more upstream question: what if the conversion data feeding the algorithm is simply not real?

The author’s point hits hard: when the data pipeline breaks, conversions still show up in your reports and the campaign still looks fine. The algorithm is just quietly training on a signal that’s degraded, skewed, or partly missing.

The article breaks “conversion data corruption” into 7 traps. I’ve gone through each one below, cross-checked against Google’s own documentation, with an audit method for every trap.

Trap 1: Enhanced Conversions Data Isn’t Normalized

Enhanced conversions work by taking first-party data you collect — email or phone — hashing it with SHA256, and matching it against signed-in Google users. The match recovers conversions that cookies miss.

But the hash only works if the input is normalized first:

• Email must be lowercased and trimmed of whitespace. The hash of [email protected] is not the hash of [email protected] — it will never match.
• Phone numbers need E.164 format: country code, no punctuation.

The scary part: nothing errors when this goes wrong. Conversions still record through your normal tracking. You just silently lose the enhanced match while the match rate sits far lower than it should, and everything looks fine on the surface.

Audit: check the match rate in your conversion action diagnostics. If it’s well below what Google reports as typical, assume a normalization problem first. Send a known test conversion with a known signed-in address and confirm it matches.

Trap 2: Consent Mode Isn’t Fully Configured

In the EEA, UK, and Switzerland, enhanced conversions data flows through Consent Mode. The common failure isn’t that the tag fails to fire — it’s that ad_user_data and ad_personalization were never mapped to granted when the user accepted.

Google’s documentation describes two tiers of Consent Mode:

• Basic: tags are blocked until consent. The cost is that you lose the data from users who decline.
• Advanced: tags load with defaults and adjust behavior based on consent, enabling finer conversion modeling. But be aware: with advanced mode, a non-consented hit still fires — no cookies stored, but a ping with browser type, device type, country, page URL, and any fields you set yourself. There’s a genuine privacy debate there. Don’t switch it on casually; run it past whoever owns privacy at your company.

There’s also a sneaky timing detail: consent status doesn’t always update the moment the user clicks accept. On some setups it only updates on the next page load, so the conversion firing on the current page — the purchase the user just completed — goes out under the pre-consent state. The banner recorded the accept, but the one conversion you most wanted to match went out before the granted signal caught up.

Audit: confirm which tier you’re running; then test the exact scenario “accept consent + complete a conversion in the same page session” and verify the granted signal is in place before the conversion fires.

Trap 3: Your Target ROAS Value Is a Fiction

This is the one that hurts the most directly. Target ROAS optimizes on conversion value. If the value doesn’t reflect what the customer actually paid, it’s optimizing against a made-up number:

• A static value hardcoded while transactions are variable.
• Currency sent inconsistently, so mixed denominations land in one column.
• Gross vs. net: two orders both show €200 — one is full price, the other is mostly discounted and likely to be returned. Gross value treats them identically, so the algorithm chases more customers like the second one.
• Shipping in or out: an order with €15 of shipping folded into the value outbids an identical free-shipping order — but that €15 is your cost, not margin.

There’s a time dimension too: the value that was correct at checkout stops being correct when the customer returns two of the three items. If you never send that back, Smart Bidding keeps treating a €200 order that became €140 as a €200 win, and keeps chasing customers who look like a return you already ate. Conversion adjustments are how you close that loop: restate the value down on partial returns, retract it entirely on cancellations.

Audit: reconcile one day of reported value against actual net revenue from your backend (remember attribution windows — they should be close, not identical). If the gap looks exactly like your average shipping charge or your gross-to-net ratio, you’ve found the mistake.

Trap 4: Setup Was Right Once, Then Silently Broke

This one is the most insidious: the setup was correct and then broke. A developer ships a site change. A GTM container gets reorganized. A CMS update changes how a variable populates on the confirmation page. The email variable stops populating, or the value parameter starts returning empty.

The conversion still fires, base tracking still works — the enhanced layer just quietly stops receiving what it needs. Because nothing throws an error and conversion counts look normal, this can run for weeks. By the time anyone notices bidding has drifted, the model has trained on degraded data for a month.

Audit: watch coverage — the percentage of eligible conversion events that arrive with user data attached. It’s charted over time in the diagnostics report, and it’s the number that moves when a variable stops populating. Coverage falling while conversion count holds steady is the signature of this failure. Alert on that divergence directly; don’t rely on a human noticing.

Trap 5: Offline Conversion Match Keys Mismatch

This one crosses from the ad platform into the CRM, which is why almost nobody writes about it.

Enhanced conversions for leads captures hashed user data at the moment of the lead, then matches it later when you upload the offline conversion from your CRM — the closed deal, the qualified lead, the booked revenue. The failure: the match key changes between the click and the close. The user submits a form with their personal email, then the deal progresses under their work email. Someone fixes a typo in the phone number during qualification. The CRM stores a normalized version of a field that was captured raw on the site.

Every time, the offline conversion uploads, tries to match on a key that no longer agrees with what was captured, and fails silently. Your single most valuable signal — revenue — never reaches Smart Bidding. The algorithm trains on the lead, not the sale.

Audit: check your offline import match rate separately from your online match rate. They’re different numbers that fail for different reasons. A low offline rate almost always points to key inconsistency between capture and upload. Standardize which field is the match key and normalize it identically in both places.

Trap 6: Data Lost Across Domains

Enhanced conversions need the user data and the conversion event to end up associated with each other. On a lot of real sites, they live on different domains.

The common shape: checkout hands off to a payment processor on its own domain (Stripe, PayPal, Payoneer…), or a booking flow completes on a subdomain the main tag doesn’t fully cover. The user enters their details on your site, then the purchase confirms elsewhere. If the conversion fires on the confirmation page but the user data was only available before the handoff, the payload goes out without its match key. The conversion records. The match doesn’t happen.

It’s easy to miss because it only affects traffic running through the cross-domain path. A single-domain test purchase matches perfectly, so the setup looks correct, while a meaningful slice of real transactions takes the broken route and never matches.

Audit: map where the user data is available against where the conversion tag fires. If they’re on different domains, confirm the data is carried across explicitly rather than assumed to persist. Test through the actual cross-domain path — the simplified single-domain test is exactly the one that hides this.

Trap 7: Duplicate or Missing Order IDs

Enhanced conversions use the transaction or order ID to deduplicate. Get that ID wrong and the match layer either double-counts or collides — both hard to see from the reporting surface:

• A reloadable confirmation page fires the same order ID more than once — one purchase matches as several.
• A test and a live transaction share an ID.
• The ID field is empty or inconsistent, so deduplication can’t run and the system drops matches.

Shopify sellers, listen up: Shopify generates an abandoned-cart ID that looks similar to a transaction ID. With a sloppy offline conversion import setup, abandoned carts sometimes get counted as purchases. The tell: the abandoned-cart ID is much longer than a normal transaction ID. But nothing flags it — you have to monitor it yourself.

Audit: pull a sample of order IDs and check them against your backend for uniqueness and completeness. Every real transaction should carry exactly one ID, present and unique. Reloadable confirmation pages are the usual source of duplicates; empty fields usually trace back to a variable that doesn’t populate reliably when the tag fires.

Neo’s Take

I fully agree with the article’s closing line: before you question your bid strategy, confirm the conversion data is real. Checking your match rate is the cheapest audit in Google Ads — and the one most people have never run.

If I had to rank these 7 traps by damage: #3 (value distortion) and #4 (silent breakage) cost the most, because they directly distort the algorithm’s judgment in Target ROAS accounts. #5 (offline matching) is fatal for lead-gen accounts. The rest are chronic diseases — but they get more toxic as automation increases. An agent allocating budget on a corrupted signal doesn’t hesitate like a person would. It commits, quickly, to a pattern built on data that was wrong.

Three extra reminders for Chinese sellers going global:

1. Cross-border payment is the danger zone: independent sites almost always hand off checkout to Stripe/PayPal. Trap 6 (cross-domain) catches almost everyone — do the cross-domain test.
2. Shopify sellers: check your abandoned-cart IDs: Trap 7 is a Shopify-specific issue you can spot by ID length.
3. Don’t confuse “conversion counts look normal” with “the data is healthy”: six of these seven traps throw no errors. Starting today, put match rate and coverage into your weekly report.

The more automated the bidding, the more valuable data quality becomes. If the data is dirty, the smarter the algorithm, the faster you lose money.