Adding more conversion signals to your Google Ads account is not helping Smart Bidding perform better. For most advertisers, it is actively teaching the algorithm to optimize for the wrong outcomes. Google Ads conversion signal quality is the single most overlooked variable in paid search performance, and the industry's default advice to "track everything" is the reason so many accounts plateau or regress despite increasing budgets. The core problem: every conversion action you mark as primary becomes a training signal for Smart Bidding. When those signals include low-intent micro-conversions, duplicated events, or GA4 goals that do not correlate with revenue, Smart Bidding dutifully finds more of that noise. It does exactly what you told it to do. It just happens that what you told it to do is wrong.
This is not a niche technical issue. It is the root cause behind a significant share of accounts where cost per acquisition rises, lead quality collapses, or ROAS declines despite "more data" flowing into the system.
What Most People Believe: More Signals Mean Better Results
The conventional wisdom is straightforward and, on the surface, logical. Google's own documentation recommends enabling enhanced conversions, importing GA4 events, and adding as many conversion touchpoints as possible so Smart Bidding has a richer picture of user behavior. Agencies repeat this advice. Tools automate it. Google reps push it in every account review.
The reasoning goes like this: Smart Bidding's machine learning models need data to function. More conversion signals mean more data points. More data points mean better predictions. Better predictions mean lower CPAs and higher ROAS.
Enhanced conversions extend this logic by matching first-party data (hashed emails, phone numbers) back to ad interactions, closing attribution gaps that cookies miss. GA4 event imports let you feed scroll depth, video views, button clicks, and page engagement directly into the bidding algorithm. The promise is that these additional signals help Smart Bidding understand the full customer journey, not just the final conversion.
This is not a strawman. In accounts with clean, well-structured conversion architectures, more high-quality signals genuinely do improve bidding performance. The problem is that "more signals" and "better signals" are not the same thing, and the industry treats them as interchangeable. For most accounts, the choice of Smart Bidding strategy is only half the equation. The other half is what you are asking the strategy to optimize toward, and that is where the damage happens.
Garbage Signals Teach Smart Bidding The Wrong Lesson
Smart Bidding does not distinguish between a conversion that generates revenue and a conversion that happens to fire on your thank-you page because someone bounced and reloaded. It treats every primary conversion action with equal seriousness. When you feed it bad signals, it optimizes for bad outcomes with impressive efficiency.
The Scroll-Depth Conversion That Destroyed A B2B Account
Consider a common scenario in B2B lead generation. A marketing team imports a GA4 "scroll_depth_90" event as a conversion action, reasoning that deep scrollers are engaged prospects. Smart Bidding identifies the user profiles most likely to scroll deeply and bids aggressively for them. The account's "conversion" volume skyrockets. The CPA looks fantastic in the dashboard. But pipeline dries up because deep scrollers on a B2B content page are overwhelmingly researchers, competitors, and job seekers. Not buyers. The algorithm found exactly the pattern you gave it. The pattern just had nothing to do with revenue.
This is not hypothetical. It is one of the most common signal quality failures in B2B Google Ads, and it is the kind of problem that tanks pipeline before anyone notices the root cause.
Why Form Fills From The Wrong Audience Look Identical To Good Leads
Even "real" conversions can be bad signals. A form submission is a form submission in the data. Smart Bidding cannot see that one lead became a $50,000 deal and another was a student requesting information for a class project. Without offline conversion data tying clicks to actual revenue, the algorithm treats both identically. It optimizes for volume of form fills, which means it finds the cheapest form fills, which almost always means the lowest-quality leads.
Session Quality Signals Are Not Purchase Quality Signals
Time on site, pages per session, and engagement rate measure whether someone found your content interesting. They do not measure whether someone is ready to buy. These are fundamentally different dimensions, and treating them as interchangeable in your conversion architecture is the single fastest way to derail Smart Bidding performance.
The Three Signal Quality Mistakes That Are Most Common
Three specific patterns account for the majority of signal quality failures. If your account has any of these, Smart Bidding is learning from flawed data right now.
Importing All GA4 Goals Without Filtering For Purchase-Intent Events
GA4 makes it trivially easy to create events for everything: file downloads, video plays, outbound clicks, scroll milestones. When you import these into Google Ads as conversion actions and leave them set to "primary," each one becomes a bidding signal. The algorithm now optimizes across all of them simultaneously, weighting a PDF download the same as a purchase. The fix is surgical: only events that directly correlate with revenue belong as primary conversion actions. Everything else should be secondary (observable but not used for bidding) or removed from Google Ads entirely.
Using Page Views Or Time-On-Site As Bidding Signals
This is a variation of the GA4 problem but worth isolating because it is so pervasive. Page views and time-on-site are vanity metrics dressed up as conversion data. They inflate conversion counts, artificially lower your reported CPA, and make your account look healthier than it is while Smart Bidding chases users who browse without buying.
Not Deduplicating Conversion Sources And Inflating Conversion Counts
When you track the same action through both the Google Ads tag and a GA4 import, that single conversion gets counted twice. Smart Bidding sees double the signal volume, which distorts its learning. Duplicate conversions are invisible in most dashboards unless you specifically audit for them, which makes this one of the most quietly destructive problems in conversion tracking. A structured account audit catches this in minutes, but most teams never run one against their conversion setup specifically.
Why This Is Getting Worse In 2026, Not Better
The trend toward more campaign automation makes signal quality errors compound faster and hit harder than they did even two years ago.
Performance Max Amplifies Bad Signals Across Every Channel Simultaneously
In a Search-only campaign, a bad conversion signal misdirects bids on search terms. That is bad. In Performance Max, that same bad signal misdirects spend across Search, Display, YouTube, Discover, Gmail, and Maps simultaneously. The blast radius is the entire Google Ads ecosystem, not a single campaign. This is one of the core reasons Performance Max creates problems for many advertisers: it is an amplifier, and what it amplifies is your signal quality, good or bad.
The AI Does Not Warn You When Your Input Data Is Flawed
Google's automation layer is remarkably good at optimizing toward whatever you tell it to optimize toward. It is remarkably bad at telling you that your target is wrong. There is no alert for "your primary conversion actions do not correlate with revenue." There is no warning for "your conversion count is inflated by 40% due to deduplication." The system assumes your conversion architecture is correct and optimizes accordingly. Garbage in, garbage out, at scale, 24 hours a day.
What Strong Signal Quality Actually Looks Like
The fix is not to remove all conversion tracking. It is to be ruthlessly precise about which signals are primary (used for bidding) and which are secondary (used for observation only).
Primary Vs Secondary Conversion Actions And How To Classify Them
A primary conversion action should meet one test: does this event reliably predict revenue? For ecommerce, that is purchases. For lead gen, that is qualified form submissions or booked calls, not all form submissions. For SaaS, that is trial starts or demo requests that actually convert to pipeline. Everything else, page views, engagement events, newsletter signups, micro-conversions, belongs in the secondary category. Secondary actions still appear in your reporting. They just do not steer bidding.
Offline Conversion Imports That Tie Ad Clicks To Closed Revenue
The most powerful signal you can give Smart Bidding is actual revenue data. Offline conversion tracking imports your CRM's closed/won data back into Google Ads, matched to the original click. This lets Smart Bidding learn which clicks produce money, not just which clicks produce form fills. For any business with a sales cycle longer than a single session, offline conversion imports are not optional. They are the difference between an account that optimizes for leads and an account that optimizes for revenue.
Enhanced Conversions As A Quality Layer, Not A Replacement For Strategy
Enhanced conversions improve attribution accuracy by filling gaps left by cookie restrictions. That is genuinely valuable. But they improve the accuracy of whatever you are tracking. If what you are tracking is junk, enhanced conversions give you more accurate junk. Use them. But use them on top of a clean conversion architecture, not as a substitute for building one.
For In-House Teams: How To Run A Signal Quality Audit In One Afternoon
If you manage Google Ads in-house, here is a focused audit you can complete in a few hours.
Pull every conversion action in your Google Ads account. For each one, answer three questions. First, is this action set to primary or secondary? Second, does this action directly predict revenue, or does it measure engagement? Third, is this action also being tracked through another source (creating duplicates)?
Then cross-reference your Google Ads conversion count against your CRM or backend data. If Google Ads reports 200 conversions last month and your CRM shows 120 qualified leads, you have an 80-conversion gap that is actively misleading Smart Bidding.
Remove or reclassify every action that fails the revenue-prediction test. Deduplicate any action tracked through multiple sources. Then wait. Smart Bidding will re-enter learning phase after significant conversion changes. Resist the urge to panic during this window. The algorithm needs time to recalibrate around cleaner data, and the performance on the other side will be structurally better.
This is exactly the kind of work that groas handles inside the DWY product. The engine identifies signal quality issues across your account, and a senior strategist walks through the findings with your team on a strategy call. Your team stays in the driver's seat and makes the final decisions, but you are working from an analysis built on patterns across hundreds of billions in ad spend, not guesswork.
For Agencies: How Signal Quality Failures Silently Damage Client Accounts At Scale
Agencies face a compounding version of this problem. When a media buyer sets up conversion tracking incorrectly on one client account, the damage is contained. When the same setup pattern gets replicated across 10 or 20 accounts, the damage scales silently.
Most agency workflows do not include a conversion architecture audit as a standard onboarding step. New clients arrive with existing tracking, the agency assumes it is correct, and Smart Bidding inherits whatever mess was already in place. The account "performs" by the metrics visible in the dashboard, but the client's actual business results do not improve. Eventually the client churns, and the agency blames market conditions.
This is where the groas DIY product changes the equation for agencies. Agencies connect client accounts to the groas engine, which surfaces signal quality issues as part of its standard operation. The agency's media buyers still run everything. They keep their brand, their client relationships, and their margin. But the execution layer underneath catches the structural problems that human-only audits miss, across every connected account, continuously. Start your 7-day free trial to see what it surfaces in your client accounts.
Why Signal Strategy Should Be Owned By Someone Who Gets Paid On Outcomes
Here is the uncomfortable truth about conversion tracking: the people who set it up (developers, analytics teams, agencies billing hourly) have no financial stake in whether the signals are correct. They get paid to implement tracking, not to verify that the tracking drives profitable outcomes.
This misalignment is why signal quality problems persist for months or years. Nobody's compensation depends on fixing them.
groas operates on a fundamentally different model. Because groas is month-to-month with no long-term contract, it earns the next month by delivering results. Signal quality is not a one-time setup task. It is a continuous strategic function, because bad signals directly reduce the performance that determines whether you stay.
For businesses that want Google Ads fully handled, the DFY product means groas owns the entire conversion architecture end-to-end. A dedicated strategist audits, restructures, and maintains your signal quality as part of running your account. This includes offline conversion imports, deduplication, primary/secondary classification, and ongoing monitoring. Nothing to log into or manage on your side.
For teams that want to stay involved, the DWY product gives you the same engine running underneath, with a strategist working alongside your team to identify and fix signal issues while you retain control of execution.
If you are unsure which fits, apply for DFY and groas will figure out the right plan on the call.
The Thesis, Restated
The industry's default advice on conversion tracking is backwards. Adding more signals does not improve Smart Bidding. Adding better signals does. Every micro-conversion, every duplicated event, every engagement metric marked as primary is actively training Google's algorithm to find more of the wrong people. And in 2026, with Performance Max amplifying every signal across every channel, the cost of getting this wrong is higher than it has ever been.
Strip your conversion architecture down to the signals that predict revenue. Import offline data. Classify everything else as secondary. And if you do not have the time, expertise, or structural incentive to maintain signal quality continuously, hand it to a team that does.
groas puts a proprietary engine trained on over $500 billion in profitable ad spend underneath a senior strategist who owns your signal quality as a core function, not an afterthought. No onboarding fees. No long-term contracts. Cancel anytime.
Apply for DFY if you want groas to own it. Get started with DWY if your team wants to stay hands-on with a strategist alongside. Either way, fix the signals first. Everything else follows.
Does Adding More Conversion Actions Improve Google Ads Smart Bidding Performance?
Not automatically. Smart Bidding treats every primary conversion action as a training signal. If those actions include low-intent events like scroll depth, page views, or duplicated conversions, the algorithm optimizes for those patterns instead of revenue. More signals only improve performance when every primary action directly correlates with a purchase or qualified lead. The key distinction is primary vs. secondary classification: secondary actions appear in your reports but do not influence bidding. Only revenue-predictive events should be primary.
What Are The Most Common GA4 Conversion Tracking Mistakes In Google Ads?
Three mistakes dominate. First, importing all GA4 events as primary conversion actions without filtering for purchase intent. Second, using engagement metrics like time-on-site or pages-per-session as bidding signals. Third, failing to deduplicate conversion sources, where the same action is tracked through both the Google Ads tag and a GA4 import, inflating counts. Each of these teaches Smart Bidding to optimize for the wrong outcomes while making your dashboard look artificially healthy.
How Do Bad Conversion Signals Affect Performance Max Campaigns?
Performance Max amplifies whatever signal quality exists in your account across Search, Display, YouTube, Discover, Gmail, and Maps simultaneously. In a Search-only campaign, a bad signal misdirects bids on keywords. In Performance Max, that same bad signal misdirects spend across every Google property at once. This makes signal quality errors significantly more expensive and harder to diagnose in Performance Max than in single-channel campaigns.
What Is The Difference Between Primary And Secondary Conversion Actions In Google Ads?
Primary conversion actions are used by Smart Bidding to make bidding decisions. Secondary conversion actions are tracked and reported but do not influence bids. Only events that reliably predict revenue, such as purchases, qualified form submissions, or booked demos, should be primary. Everything else, including newsletter signups, content downloads, and engagement metrics, should be secondary. This classification is the single most important structural decision in your conversion setup.
How Do I Know If My Conversion Tracking Is Hurting My Google Ads Performance?
Compare your Google Ads conversion count against your CRM or backend revenue data. If Google Ads reports significantly more conversions than your business actually sees in qualified leads or sales, your signals are inflated. Also check for duplicate conversion sources, engagement events marked as primary, and any GA4 goals imported without intent filtering. groas identifies these issues as part of its standard operation. In the DWY product, a senior strategist reviews signal quality with your team; in DFY, groas owns and fixes the entire conversion architecture on your behalf.
Should I Use Enhanced Conversions In Google Ads?
Yes, but understand what enhanced conversions do and do not do. They improve attribution accuracy by matching first-party data to ad clicks, closing gaps created by cookie restrictions. However, they improve the accuracy of whatever you are tracking. If your conversion architecture includes junk signals, enhanced conversions give you more accurate junk. Use enhanced conversions as a quality layer on top of a properly structured conversion setup, not as a standalone fix.
How Often Should I Audit My Google Ads Conversion Tracking?
At minimum, audit your conversion architecture quarterly and after any significant website, CRM, or analytics change. Signal quality degrades over time as new events get added, tracking scripts break, or GA4 configurations change. For accounts spending at meaningful levels, continuous monitoring is better than periodic audits. groas monitors signal quality continuously across every account it manages, surfacing issues before they compound into performance problems.
Can Smart Bidding Recover After I Remove Bad Conversion Signals?
Yes, but expect a re-learning period. When you reclassify or remove primary conversion actions, Smart Bidding enters learning phase as it recalibrates around the cleaner data. This typically lasts one to two weeks depending on account volume. Do not panic or revert changes during this window. Performance on the other side will be structurally better because the algorithm is now optimizing toward signals that actually correlate with revenue.
Why Does My Google Ads CPA Look Good But My Actual Cost Per Customer Is High?
This is the classic symptom of signal quality failure. Your reported CPA is calculated against whatever conversion actions are set to primary. If those include micro-conversions, duplicated events, or low-intent actions, your CPA appears low because you are "converting" users who never become customers. Your actual cost per customer, measured from ad spend to closed revenue, tells the real story. Offline conversion imports solve this by feeding real revenue data back into Google Ads.
Who Should Own Conversion Tracking Strategy For Google Ads?
Conversion tracking should be owned by whoever has a direct financial incentive to get it right. Developers implement tags but do not verify business impact. Analytics teams configure events but do not manage bidding. The person or team responsible for profitable ad spend outcomes should own signal strategy. groas builds this into every engagement: in DFY, a dedicated strategist owns your entire conversion architecture end-to-end; in DWY, the strategist works alongside your team to ensure signal quality stays clean while you retain control.
.png)
