Conversions API and enhanced conversions: the first-party tracking set-up every CMO should ask for

What these two tools do

Meta Conversions API and Google enhanced conversions are two ways of sending your own conversion data to the ad platforms, without relying only on the browser. They help the platforms see more of the sales and leads your ads produce, and they let you report real business outcomes, such as a qualified lead or a signed deal, so bidding learns from the results you care about.

For a CMO the case is simple. Ad platforms optimise towards the signals you give them. If the only signal is a form submission seen by a browser tag, the system learns to find form fillers, including time-wasters. If you send back the lead that your sales team accepted, it can learn to find more of those.

The definitions come from the platforms themselves. Meta's Conversions API documentation says the API connects marketing data from your server, website platform, mobile app or CRM to Meta systems, to optimise targeting, lower cost per result and measure outcomes. Google's enhanced conversions help page says enhanced conversions improve measurement and support stronger bidding, and add to the conversions you already track.

Why this belongs on the leadership agenda

Three pressures push this up the list, and none of them depends on a particular platform announcement.

• Browser limits. Browsers and apps restrict third-party cookies and tracking in different ways, and consent banners reduce what a tag can collect. A browser-only set-up sees less than it used to.
• Automated bidding. Meta Advantage+ and Google's smart bidding make more decisions for you. Their quality depends on the data you feed them.
• Longer sales cycles. If a lead becomes revenue weeks later on a phone call or a WhatsApp chat, a pixel never sees the outcome. Only your CRM does.

The same logic sits behind our CAC payback guide: you cannot manage cost per customer if the platform only sees cost per form.

How the Meta Conversions API works

With the Conversions API, your server sends an event, such as a purchase, a lead or a qualified lead, to Meta with a POST request. Meta processes it like an event from the pixel and uses it for measurement, reporting and optimisation. You can send website events, app events, business messaging events and offline conversions through one integration.

Pixel and API together: deduplication

Most businesses keep the pixel and add the API for redundancy. That only works if Meta can tell that two reports describe the same action. According to Meta's deduplication guidance:

• Meta treats events as identical when the ID and the name match. The pixel's eventID must equal the server's event_id, and the pixel event must equal the server event_name.
• Deduplication applies within 48 hours of Meta receiving the first event with that ID.
• If the events do not meaningfully differ, Meta generally prefers the one it receives first.

This is the first thing to test. A missing or mismatched event ID is the most common cause of inflated conversion counts, and it makes a campaign look better than it is.

What to send

How Google enhanced conversions work

When a customer converts, your tag can capture first-party data they entered, such as email address, name, home address or phone number. Google says this data is hashed with SHA256, a one-way algorithm, before it is sent, and that it should be normalised first: trim spaces, lowercase text, format phone numbers to E.164 and tidy Gmail addresses. Google then matches the hashed data to signed-in Google accounts that engaged with your ad.

There are two variants:

• Enhanced conversions for web measure online conversions such as sales and events on your site.
• Enhanced conversions for leads connect a website lead to an offline outcome. You import the later conversion, such as a closed deal, from your CRM, and Google uses the hashed lead data to attribute it to the campaign.

For most Indian service, real estate, education and B2B businesses, the second variant matters most, because the money is made on a call or a site visit after the form.

A 30, 60, 90 day rollout plan

Days 1 to 30: decide the outcome and clean the data

• Agree one primary business outcome per campaign type, for example "qualified lead" with a written definition from sales.
• Check that your CRM records the ad source (UTM or click ID) for every lead. Without it nothing can be matched.
• Confirm a consent notice and purpose for sharing data with ad platforms. Our DPDP Rules checklist covers the key dates.

Days 31 to 60: build and test

• Set up Conversions API (directly, through a partner integration or a tag server) and enhanced conversions in your tag manager.
• Use the platforms' test tools to confirm events arrive and deduplicate correctly. Compare platform counts to CRM counts for a week.
• Document which fields are hashed, who can see them and where they are stored.

Days 61 to 90: switch the optimisation target

• Move one campaign to optimise for the better signal, keeping a control campaign on the old one.
• Judge by cost per qualified lead and cost per customer, not cost per form.
• If the lift is real, extend to other campaigns; if not, check data volume and delay before blaming the method.

Three ways to implement, and how to choose

There is no single right route. The choice depends on your stack, your team and how much of the customer journey happens outside the website.

A worked example, with clearly labelled assumptions

Assume a services firm spends the same monthly budget on Meta lead ads. Today the campaign optimises for form fills. Suppose sales accept about one in three leads as qualified. If the firm sends the accepted-lead event back through the Conversions API and moves the campaign to optimise on it, the system receives fewer but more meaningful signals. The numbers here are illustrative only. What you measure is whether cost per accepted lead falls against the control campaign over six to eight weeks, and whether closed revenue per rupee spent improves. If volume of accepted leads is too thin for the platform to learn, keep the earlier signal and pass the later one as a secondary event.

The same thinking applies to Google. Import offline conversions for the stage your team trusts, and keep the definition stable for at least one quarter so you can compare like with like.

Common pitfalls

• Double counting. Caused by missing event IDs. Test with the platform tools before you scale spend.
• Too few outcome events. If only a handful of deals close each month, optimise on an earlier qualified step and pass deal value separately.
• Slow feedback. A conversion reported weeks late is less useful. Upload offline conversions at least weekly.
• Dirty identifiers. Unnormalised emails and phone numbers fail to match. Standardise at capture.
• No owner. Tracking set up by an agency and never reviewed drifts. Name an internal owner and review quarterly.
• Privacy shortcuts. Sending data you have no basis to share creates legal and brand risk. Involve legal early.

How a CRM fits

The outcome data lives where your team records it. If follow-ups sit in spreadsheets and personal phones, there is nothing reliable to send back. A CRM such as Be12 CRM keeps the lead source, calls, WhatsApp chats and invoices together, which makes the offline outcome list a one-click export.

Ten questions to ask your team or agency

1. Which business outcome is each campaign optimised for today?
2. Are we using Conversions API and enhanced conversions, or only browser tags?
3. How do we deduplicate pixel and server events, and who tested it?
4. Where do click IDs and UTMs get stored for each lead?
5. How often do we upload offline outcomes, and how late are they?
6. Which personal data is hashed and shared, and under what consent?
7. What is our gap between platform-reported conversions and CRM-confirmed ones?
8. Who has admin access to the tag manager and pixel in our own name?
9. How would we know if tracking broke tomorrow?
10. What did the last change in measurement do to cost per customer?

None of this needs a large engineering team. It needs a clear definition of success, one reliable data source and disciplined testing. Start with one campaign, prove the lift, then roll it out. For wider context on judging channels, see our guide to marketing budget allocation.