B2B Google Ads Conversion Tracking: Why Leads Are Not Enough

A Google Ads account can show 100 conversions.

Sales may see 20 actual leads.

And maybe two of those are companies worth talking to.

Nothing necessarily looks broken inside Google Ads.

The campaign may show a healthy cost per conversion. The conversion rate may be improving. Google may even be spending more confidently because it thinks it has learned what works.

But it may be learning from the wrong thing.

This is the core problem with B2B Google Ads conversion tracking:

Google does not know what a good lead is.

It knows what you tell it to optimise towards.

What are you actually tracking?

A conversion can be almost anything.

A form submission.

A free-trial signup.

A demo request.

A click on a submit button.

A visit to the pricing page.

A visit to the demo page.

Time spent on a page.

Someone can set up all of these as conversions.

The question is not whether the event is technically measurable.

The question is whether it tells you something useful about the chance of that person becoming a customer.

A person visiting a pricing page may be more valuable than a random website visitor.

A person filling in a demo form may be more valuable than a pricing-page visitor.

A sales-qualified lead may be more valuable than a demo request.

A paying customer is more valuable than all of them.

But Google only sees the signals you send it.

If you tell Google that a pricing-page visit is valuable, it will look for people who visit the pricing page.

If you tell it that free-trial signups are valuable, it will find people likely to sign up for free trials.

That can be useful.

Or it can become a very efficient way to bring in people who like signing up for free things and never become customers.

A lead is just somebody who has stopped being anonymous

A lead is not automatically a good lead.

Usually, a lead is simply a person who has given you some information that allows you to identify or contact them.

An email address.

A phone number.

A name and company.

A free-trial signup.

That is useful. But it is not the final outcome.

The worst type of lead is somebody who was looking for something else, misunderstood the offer, left their details and then has no realistic reason to become a customer.

They may be cheap.

They may make your cost per lead look good.

They may even make the campaign look like it is scaling.

But they are not useful.

A good lead is not just somebody who filled in a form.

It is somebody whose original need, search, company, role and commercial situation make them at least plausible as a future customer.

Google Ads cannot decide that on its own.

The company has to tell Google, directly or indirectly, what happened after the initial conversion.

Smart Bidding is only as smart as the information it receives

With manual bidding, you might decide that a click on a keyword is worth up to €5.

You are essentially saying: "I am willing to pay up to this amount for the chance to appear when somebody searches for this."

But the keyword does not tell you everything.

A CEO from a company you want to work with and a student researching the topic may search for something very similar.

Google's automated bidding is supposed to improve on that.

It can bid differently depending on the signals it has around the searcher and the likelihood that they will complete a conversion.

But it can only learn from the conversions you give it.

If the conversion is weak, Google may become better at finding weak conversions.

If the conversion is strong, Google has a better chance of finding people who look more like commercially useful customers.

Google is not deciding whether somebody will buy.

It is deciding how much it is worth bidding for the chance to show an ad.

Everything after the click happens after you have already paid.

That is why the quality of the conversion signal matters so much.

Soft conversions are useful. But only when their relationship to real outcomes is checked.

Most B2B companies do not have enough customer conversions every day for Google to use only paying customers as its main signal.

A high-ticket service may close a few clients per month.

An enterprise SaaS business may have long sales cycles and only a small number of customers entering the CRM.

So the account often needs earlier signals.

For example:

• A pricing-page visit.
• A demo-page visit.
• A product interaction.
• A free-trial signup.
• A form submission.
• A meeting booked.

These can all be useful.

But they are not equal.

The closer the event is to actual revenue, the more useful it usually is.

A person who visits a demo page is not the same as somebody who requests a demo.

A free-trial signup is not necessarily the same as an activated trial user.

A demo request is not necessarily the same as a sales-qualified opportunity.

The problem is not using softer conversions.

The problem is assuming that a soft conversion will always mean what it meant when you first looked at the data.

The 10-to-1 ratio became 45-to-1

We had a demo page where the original relationship looked reasonable.

For roughly every ten people who visited the demo page, one filled in the demo form.

That made the demo-page visit seem like a useful earlier signal.

There were not enough demo requests to give Google a lot of data on their own, so using the page visit as an additional conversion made sense.

Then Google began optimising towards the demo-page visit.

And the relationship changed.

Instead of one form submission for every ten demo-page visitors, it became closer to one for every 45 visitors.

Google had found more people who could reach the page.

But they were not the same kind of people as the original visitors who were likely to request a demo.

The original logic had stopped holding.

One adjustment was to count the demo-page visit only after someone had stayed on the page for at least 30 seconds.

That was not because 30 seconds is some universal magic number.

It was simply a way to stop treating people who opened the page and immediately left as meaningful conversion signals.

The important point is the ratio.

When you use a soft conversion, you need to keep checking whether it still predicts a harder conversion.

If one in ten pricing-page visitors starts a trial, that may be useful.

If Google starts optimising towards pricing-page visits and the ratio later becomes one in 40, the signal is no longer carrying the same meaning.

Conversion values are a way to tell Google that events are not equal

One way to handle multiple conversion events is to assign values.

For example:

• A free trial could be worth €200.
• A pricing-page visit could be worth €20.

That does not mean a click costs €200 or €20.

It means you are giving Google a relative indication of what matters more.

If a person is likely to create a high-value event, Google may be willing to bid more aggressively to show an ad.

If they are only likely to create a low-value event, it may bid less.

The numbers do not need to be mathematically perfect from day one.

But they should reflect a real logic.

A pricing-page visit should not carry the same weight as a qualified demo request.

A demo request should not automatically carry the same weight as a customer.

And the values need revisiting when the ratios change.

A technically correct conversion can still be false

One of the most dangerous tracking problems is when the conversion looks correct at first.

For example, we had a form where the conversion was tracked through a click on the submit button, rather than confirmation that the form had successfully reached the CRM.

There were technical reasons for doing it this way.

And normally, it worked reasonably well. Most people do not click submit on an empty form.

Then a Performance Max campaign was launched.

Suddenly, hundreds of people came to the website, clicked the button, did not submit the form and left.

Google Ads recorded conversions.

The CRM did not receive leads.

From Google's perspective, the campaign was working.

It had found a type of traffic that could complete the measured event.

From the business perspective, it was not working at all.

This is why you cannot just look at whether an event fires in Google Ads.

You need to ask:

• Did the form actually get submitted?
• Did the person enter the CRM?
• Did the trial account get created?
• Did the user return?
• Did sales accept the lead?
• Did anything happen after the first event?

A conversion tag can be working perfectly and still be tracking the wrong business outcome.

The best signal is usually the one closest to revenue

The strongest setup is when Google receives feedback about what happens after the initial click.

For example:

• Somebody clicks a Google ad.
• They fill in a form or start a trial.
• They enter the CRM or product system.
• Sales qualifies them, or they become an opportunity.
• They become a customer.
• That customer outcome is sent back into Google Ads.

This is what people usually mean by offline conversion tracking.

The conversion does not happen "offline" in the sense that it must happen in a physical store.

It means the outcome happens outside the Google Ads tag and needs to be imported from the CRM, product system or another internal system.

For B2B, this is often where the real quality data sits.

The website can tell you that someone filled in a form.

The CRM can tell you whether they were a relevant company.

The sales team can tell you whether they were a real opportunity.

The product system can tell you whether they actually used the trial.

The finance system can tell you whether they became a paying customer.

The more of that feedback gets back into the account, the better chance Google has of learning from something commercially meaningful.

But customer conversions alone are often too low-volume

In an ideal world, Google would optimise only towards paying customers.

In reality, a lot of B2B companies do not have enough of them per month for that to work cleanly as the only signal.

A useful benchmark is around 20 to 30 meaningful lower-funnel conversions per month. More is better.

That does not mean every account below that number cannot work.

It means Google has less data to learn from, so the setup needs more judgement.

If you only have two or three customers per month from Google Ads, customer data should still be sent back.

But you may also need stronger earlier signals:

• A trial that reaches a meaningful activation point.
• A demo that sales accepts.
• A sales-qualified lead.
• A real opportunity.
• A pricing-page visit that has historically correlated with trials or meetings.

The key is to build a hierarchy.

Do not throw every measurable action into Google Ads and call it a conversion.

Decide which actions are closer to business value, which ones happen often enough to help the system learn, and how they relate to each other.

What should be checked when lead quality drops?

When somebody says, "The leads are getting worse," start by checking the chain.

Not just the Google Ads dashboard.

1. Did the actual conversion happen?

   Was it a real form submission, a real account creation or a real meeting booking?

   Or was it a proxy event such as a button click or page visit?

2. Did it reach the next system?

   Did the person enter the CRM?

   Did the signup create a usable account?

   Did the trial user actually return?

3. Did it become a commercially meaningful stage?

   Did sales accept the lead?

   Did the person become an opportunity?

   Did they activate the product?

   Did they become a customer?

4. Has the relationship between early and late signals changed?

   Are you still seeing one demo request for every ten demo-page visitors?

   Or did that relationship become much weaker after Google started optimising towards the softer event?

5. Is Google receiving the right feedback?

   Are qualified leads, opportunities and customers being sent back into the account?

   Or is Google still learning only from the first thing that happens on the website?

For the wider audit process around search terms, location settings and whether the account is buying the right traffic in the first place, read Why B2B Google Ads Campaigns Fail: What We See in Audits.

Final point

The goal is not to track as many conversions as possible.

The goal is to give Google enough useful information that it stops optimising towards people who are easy to convert and starts learning from people who are commercially worth acquiring.

That means:

• A lead is not automatically success.
• A soft conversion is not automatically a useful signal.
• A tracked event is not automatically a real business action.
• A conversion ratio is not automatically stable.
• CRM and product outcomes matter more than what looks good inside Google Ads.

Google can only optimise towards the information it receives.

The quality of that information determines the quality of the traffic it learns to buy.

If you are not sure whether your conversion setup is sending Google the right signals, we run a standalone B2B Google Ads audit for €450 per account — it covers conversion tracking, search terms, campaign structure and landing pages, and ends with a written assessment and a prioritized action plan.

Related reading: why B2B SEO vs. PPC attribution is often misleading explains what these numbers can and cannot prove, and why we usually run Google Ads before investing in SEO covers how the same data guides organic decisions.