Google Ads Offline Conversions: How to Improve Lead Quality

Google Ads offline conversions mean sending real sales signals from your CRM back to the original ad click. That stops Smart Bidding optimising towards every lead and starts it optimising towards the leads that actually close as business.

In this post we will answer:

• What Google Ads offline conversions are and how they work technically in 2026
• Why your bidding strategy degrades without them
• How to set the tracking up step by step
• What to watch for so you avoid the common mistakes

We are a PPC agency with nearly 15 years of experience in Google Ads and conversion optimisation. If you are reading this post, you are looking at one of the highest-leverage changes you can make to a lead generation account.

The key points in brief

• Google Ads offline conversions link closed deals in your CRM back to the original ad click
• Without offline signals, Smart Bidding optimises towards every lead, regardless of quality
• In unoptimised accounts an average of 57% of budget goes to search terms that never result in a completed conversion (Aimers, 2025)
• ECfL (Enhanced Conversions for Leads) is Google’s recommended tracking method for 2026 and encrypts email addresses and phone numbers with SHA-256 before the data is sent
• The conversion window is 90 days for GCLID uploads and 63 days for ECfL

It is mainly needed if you take in a lot of leads where the quality varies widely – for example if you have a single form that also collects irrelevant enquiries, such as support questions from existing customers and job applications.

Google Ads offline conversions work by using a unique ID to link an ad click to an event that happens away from the website – for example when a lead is qualified in your CRM, or when a deal closes.

There are two methods in parallel use in 2026:

GCLID (Google Click ID) is a URL parameter added automatically to every click on your ads. You capture the ID in a hidden form field and store it in your CRM. When a lead is qualified, the GCLID is sent back to Google Ads and matched against the original click. The method is deterministic and works regardless of whether the user is signed in to a Google account.

ECfL (Enhanced Conversions for Leads) is Google’s recommended method. Rather than relying on a URL parameter, ECfL collects first-party data – usually email address and phone number – directly from the form. The browser encrypts the information with SHA-256 hashing before it is sent to Google, which then matches it against signed-in users in its database. The match is probabilistic and works best when the user is signed in to a Google account in their browser.

In practice we recommend setting up both methods in parallel. That gives you maximum coverage regardless of which method is able to match in a given session.

Many digital channels, Google Ads among them, run automated bidding strategies.

CPA (cost per acquisition/action) is a popular bidding strategy in Google Ads where you let Google steer the bidding. The goal is to bring in a lead – depending on how you have configured it – at your target cost per conversion.

The drawback of CPA compared with traditional CPC is that Google struggles to tell the difference in quality between leads. Optimising for leads without giving Google clear feedback on which ones are good can make your results worse.

Google starts optimising towards the goals you have set and does what it can to bring in as many as possible.

If you do not distinguish between your leads, Google will not either. It is doing exactly what it is supposed to do: optimising to increase the number of leads.

What happens in practice? The algorithm seeks out the search intents with the least friction. Job seekers and support enquiries cost less to convert than a decision maker with buying intent. You get more leads, but the quality falls over time.

The data backs this up: 60–70% of B2B leads generated online are junk when offline quality data is missing (GrowthSpree, 2026). In accounts without offline tracking, an average of 57% of budget goes to search terms that never result in a completed conversion – in the worst optimised accounts the median is 73% (Aimers, 2025).

The penalty is exponential. For every 10% increase in unoptimised budget, CPA rises by 38–65% (Aimers, 2025). That means an account wasting 40% of its budget does not lose 40% of its potential – it loses considerably more, because flawed training data gradually degrades the entire bidding model.

How do you improve the efficiency of Google Ads advertising? Give the algorithm the right data. A closed feedback loop from CRM to Google Ads – where qualified leads and closed deals are sent back as conversion events – is the most effective way to improve results without raising the budget.

If you feel the quality of your enquiries is dropping, this could be the fix.

For e-commerce sites and other websites where the leads are broadly good, it is less of a priority.

Where your leads are good but the potential order value varies wildly, offline conversion tracking can be a strong option.

What you do then is send back not only the fact that the lead is qualified, but its value as well. That is called value-based bidding (VBB), and it means the algorithm optimises towards maximum gross profit rather than lead volume. You can also measure against profit on ad spend (POAS) – how much profit you generate per pound spent on advertising – instead of the more common ROAS, which only looks at revenue.

Over time it lets you compare the quality of leads from different channels.

There are two different methods we use to track offline conversions.

The first step, as we covered above, is to capture the GCLID from your enquiries.

With a CRM system it is possible to send data back to Google automatically when a lead is qualified as an SQL (sales qualified lead). You can of course name that stage whatever you like.

Salesforce and HubSpot are two of the more common CRM systems where this is very straightforward to integrate. But almost any CRM can be connected using tools such as Zapier – with one caveat we will come back to below.

We are happy to help you get started with this and with your Google Ads optimisation. Once the flow is installed, whoever owns the pipeline only has to mark each interesting lead as an SQL – and from that point on we are optimising for qualified leads rather than all leads.

That strips out job applications, duplicates, cold callers and other irrelevant leads. Your CPA bidding strategy will now optimise towards SQLs rather than every enquiry.

It also makes the data more dependable: a conversion in Google Ads now means a qualified lead.

Worth bearing in mind is that SQLs make improvements easier to compare over time. The data becomes more trustworthy because you are measuring the increase in qualified leads – and what matters more in digital marketing than data you can act on with confidence?

It usually takes a year before you see the full effect, once you can compare against the previous year. We therefore recommend keeping the old lead conversions in place so you do not lose that comparison in the meantime. But sooner or later you should switch over completely.

Zapier and Make.com are convenient ways to get started and work well for most businesses. But for organisations with high lead volumes and established CRM systems there is a more robust option: direct webhooks via cloud functions.

How it works: when a lead changes status in your CRM, a webhook triggers a cloud function that calls Google’s API directly. No intermediaries, and no connectors that quietly break when the CRM updates. It is often more cost-efficient and more reliable at high volume.

We recommend this architecture for businesses handling more than a couple of hundred leads a month.

If you do not use a CRM system, or the system you have will not send this data to Google Ads, you can handle it directly in Google Sheets. Your columns might be:

• Email
• Phone number
• Conversion Name
• Conversion Time
• Conversion Value
• GCLID

Google Ads does not currently connect directly to that document. You can, however, have Zapier or Make.com send the data whenever a new row is added.

If you run a B2B business with a long sales cycle you will often hit a concrete problem: the algorithm needs at least 30 conversion events a month to work at its best. If you only close 5–10 deals a month, Smart Bidding is starved of signals and the bidding becomes unstable.

The fix is to track intermediate milestones in the sales process, not just the final deal:

• SQL (sales qualified lead) – the lead is qualified and is assigned an estimated value
• Meeting or demo booked – the deal is moving forward and is assigned a mid-range value
• Closed deal – the actual order value is sent back

That gives the algorithm a constant flow of signals to work with, while the long-term steering happens against actual revenue.

Offline tracking is not only a technical question – it is just as much a sales management one.

An industry expert in the field puts his finger on it: “The massive and often ignored bottleneck is getting salespeople to actually update the right field in the CRM at the right time. Without that, the technology is worthless.”

Pipelines can be perfectly connected and webhooks can run flawlessly, but the system collapses if nobody updates the right status field when they should. What you can do:

• Simplify the CRM flow so the salesperson only has to change one status field
• Trigger the webhook automatically off the status change, not off manual data entry
• Make it clear to the sales team how the data shapes which leads they get next

Without that discipline in the sales team you will never get the full effect of the tracking setup, however good the technology is.

Offline tracking is not fire and forget. There are a few things worth knowing before you set it up.

The upload window is stricter than many people think. Conversions based on GCLID have to be uploaded within 90 days of the original click. For ECfL the window is 63 days. For B2B businesses with long sales cycles that can mean you never see the final deal in the data if you only track Closed Won – another reason to use multi-stage flags.

Silent deduplication is another common problem. If the same GCLID and conversion time are uploaded twice by mistake – via an incorrectly triggered webhook, for instance – the duplicate is discarded in Google’s backend without any warning message. You see “missing” data that never appears in the dashboard, which makes care in your data handling essential.

Delayed reporting occasionally happens when Google’s interface has issues. Data is uploaded correctly via the API but does not appear in the interface for several days. Avoid panic measures such as Data Exclusions in that situation – they risk disrupting the algorithm more than the delay does.

• If your leads are generally good
• If you have few leads and risk ending up with even fewer (in that case we recommend the alternatives below instead)

If this feels awkward and you cannot or would rather not go through with it, there are alternatives.

You could, for instance, rebuild the forms on your website so that you are not tracking irrelevant enquiries – either by using separate forms, or by asking the visitor to state their reason for getting in touch.

Be careful not to make the form more laborious than it needs to be. Every extra field is said to reduce completion rates. But if you cannot trust your data and paid advertising matters to your business, it can be the right route.

If you want help with your Google Ads tracking and lead quality, do get in touch. We work with clients in 10 countries and are happy to look over your setup. Request a free analysis and we will come back to you within 24 hours.