TL;DR
Google Ads will bid toward the conversion you make primary, whether that event is a useful lead or a cheap form fill. It won’t know the difference unless you give it a better signal. Choose the earliest sales-stage event that has a real link to revenue, a clear definition, enough volume, and a short enough delay to guide bidding. Then use closed jobs, revenue, and lead-rejection reasons in the CRM to check whether that signal is still producing the customers you want.
Google Ads can be right while the business is wrong
There is a version of a Google Ads account that looks great in a monthly report and terrible everywhere else.
Cost per lead is down. Conversions are up. The campaign has hit its target CPA. On paper, someone should be celebrating.
Then the sales team tells a different story. The leads are not answering. They are looking for jobs, free advice, or something you do not sell. Appointments are not booking. The pipeline is flat.
Nobody is necessarily lying. Google Ads is reporting what happened inside Google Ads. Sales is reporting what happened after the lead left it. The trouble begins when those two versions of performance get treated as the same thing.
Google is not confused. It is following instructions, this is called “Automation Drift”. If a form submission is set as the primary conversion, Google learns that a completed form is successful. It will look for more people likely to complete forms, and it will usually find the easiest, cheapest ones first. It cannot know that one form represents a serious customer and another represents someone who will never become one unless you give it a better instruction.
That is why a lower cost per lead can be a warning sign rather than a win. The account may have become more efficient at producing inquiries while becoming less effective at producing customers.
The solution is not to stop measuring form fills, nor is it to send every number in your CRM back into Google Ads. Different numbers have different jobs. One signal should help Google decide where to bid. Another should prove whether those bids created revenue. A third should explain why performance changed when the dashboard and the business no longer agree.
How to choose the right Google Ads bidding signal
Before making any conversion action primary, put it through the same four checks: revenue link, a stable definition, timing, and enough volume.
First, ask whether the action is meaningfully connected to revenue. A form fill may start a conversation, but it doesn’t always show that the person can become a customer. A qualified lead, booked consultation, or verified service request usually has a closer link to revenue, because someone has done more than give the business their contact details.
Next, ask whether sales applies the definition the same way every time. A qualified lead might mean a homeowner in the service area who wants the right service and has a realistic budget. If two people use different standards, Google will learn from a mixed signal.
Then ask whether the event happens quickly enough to guide decisions. A closed deal that takes 60 days to appear is useful evidence, but it may arrive too late to guide bids. The account may need an earlier sales stage while it waits for revenue to land.
Finally, ask whether the event happens often enough to support learning. An HVAC account that closes four jobs from 40 form fills in two months gives Google very little to work with at the sale stage, even when those jobs are valuable. A sales-accepted lead or booked appointment may be the better primary conversion when it appears quickly, has a stable definition, and turns into revenue often enough to stay trustworthy.
For a more advance guide on conversion tracking and bidding signals, read our article Advanced Conversion Tracking Strategies
Signal 1: The signal Google Ads should bid on
The bidding signal is the one action you would pay for someone to do.
That sounds obvious, but it rules out a surprising number of account defaults. A page view is not it. A button click is not it. In many accounts, a form fill is not it either.
The right signal sits at the earliest point in the sales process where someone has demonstrated meaningful commercial potential. For a homebuilder, that might be a prospect who has supplied a real budget and timeline for purchase. For a repair company, it could be a service request in the actual service area. A B2B business, it may be a lead the sales team has accepted rather than every person who downloaded a guide.
Google needs this signal to occur often enough to find patterns, and cleanly enough that it represents the same thing every time. If sales calls something a “qualified lead” on Monday but counts a very different type of person as qualified on Friday, the account is not receiving a useful instruction. It is receiving noise with a label on it.
That is the trade-off, closed revenue is usually the truest outcome, but it can arrive months after the click and may be too rare for a campaign to learn from efficiently. A form fill arrives immediately and in volume, but may be too weak to mean much. The useful bidding signal often lives between those two extremes.
Before making an action as your “primary conversion”, ask:
- Would we want more of these exact people, even if the cost per lead rose?
- Can sales or operations apply this definition consistently?
- Does the event arrive within a time frame Google can react to?
- Does it happen often enough to give the campaign something real to learn from?
If the answer is yes, it may be the right action to optimize toward. If not, keep measuring it, but do not make it the instruction that controls the budget.
Signal 2: The signal that proves Google Ads are working
Use one number from the CRM to check whether the leads you are buying become customers. Put it beside ad spend every month.
For a lead-generation campaign, start with four figures:
- Google Ads spend
- Qualified leads
- Sold jobs or closed deals
- Revenue, or gross margin if the business can report it reliably
The useful calculation is cost per sold job. Divide the ad spend by the number of customers who actually bought. That number tells you far more than cost per lead when lead quality is the issue.
Imagine a heating company spends $5,000 in a month. Google Ads reports 50 qualified service requests, so the cost per qualified lead is $100. Fifteen of those requests become booked, paid jobs. Its cost per sold job is $333. If those jobs produced $45,000 in revenue, the business can then decide whether $333 is an acceptable acquisition cost for that work.
Now imagine next month looks better inside Google Ads. Cost per qualified lead drops to $80. The sales team only sells ten jobs, though, so the cost per sold job rises to $400. The campaign did not become more efficient. It sent sales more leads that looked qualified at first and converted less often later.
This is the validation signal. It gives the owner and agency a shared test, are lower-cost leads also becoming customers at an acceptable cost?
Keep the bidding signal from the previous section in place if it arrives quickly and has enough volume. Closed revenue may be too slow or too scarce to guide daily bidding. Use it to judge the account instead.
One detail matters here. Match the reporting window to the sales cycle. If it takes 30 days for a lead to close, do not compare this month’s ad spend with this month’s closed deals. Review leads after they have had their normal time to move through the pipeline.
Signal 3: The signal that tells Google Ads what to fix
The first two signals tell you whether Google is finding leads and whether those leads turn into customers. When those numbers drift apart, the next job is to find the cause before anyone changes bids, budgets, or campaign settings.
Build a required “not a fit” field in the CRM for rejected leads. Use a closed list of plain reasons: outside the service area, wrong service, price shopper, job seeker, duplicate or spam, and no response after follow-up. Someone on the sales team has to choose one before closing the record.
This system will turn a vague complaint, “the leads are bad,” into evidence the account manager can act on.
Say a homebuilder’s form-fill volume rises while booked viewings fall. The CRM may show that most rejected leads have budgets below the lowest available home price. That is not a bidding problem. The landing page and ad copy need to state the starting price sooner, and the form may need a budget question.
The same report can point somewhere else. A spike in leads outside the service area calls for a check of location settings, excluded postal codes, and the places named in the ads. Job seekers filling in the form send you to search terms, negative keywords, and the page they land on. When qualified leads aren’t being called back, leave Google Ads alone and fix the follow-up process first.
Review the reasons once a month by campaign, service line, and location. Look for a change in the mix. A campaign may still produce enough qualified leads overall while gradually sending more low-value work to the sales team.
Keep these fields in the CRM. Google may use some outcomes later, but the first job of diagnostic data is to show the person managing the account where to look next.
Not every conversion is a primary conversion
Google Ads gives every tracked action a setting that matters more than its name, a primary conversion can guide bidding, while a secondary conversion stays in the account for reporting and lets you watch it without asking Google to spend more money finding it.
That distinction is easy to blur. A business may track a phone call, a form fill, a booked appointment, a chat, and a closed sale. Seeing all of them in one report is useful. Telling a campaign to chase all of them at once gives Google a mixed brief.
Open the Goals section in Google Ads and make a simple list of every conversion action. Next to each one, write what it means to the business and whether you would pay for more of it.
Check campaign-level goals as well. A secondary conversion can still affect bidding when it is included in a custom goal. The label alone doesn’t settle it. Look at the goal the campaign is actually using before changing a bid strategy or judging its results.
Do this audit whenever the business changes its offer, sales process, or definition of a qualified lead. When the old setup rewarded a generic form fill and the new sales team only accepts leads with a budget, the conversion goal has to change too. Otherwise, Google will keep buying the customer you used to want.
Conclusion
Better measurement does not mean handing Google every number in the CRM. It means deciding what each number is for.
Use the primary conversion to tell Google where to spend. Use the sales result to judge whether the spend paid off. Closed jobs matter, but the reason a lead was rejected can show what needs fixing before the account loses another month of budget.
Start with the conversion goals already in the account. Check three things: which actions are primary, which campaigns use them, and whether sales would genuinely want more of those people. If you are unsure where the account’s numbers and business results have drifted apart, use the Performance Gap Checklist before changing the budget.