Skip to content

Google AdsPublished Updated 8 min read

Google Ads Offline Conversion Import from Your CRM

Capture click IDs, map CRM stages to conversion actions and set up imports so Smart Bidding learns from qualified leads and closed deals, not just form fills.

Muhammad Qitmeer, founder of Augere Labs
Muhammad Qitmeer
Co-founder and CEO, Augere Labs
Share

Offline conversion import lets you send outcomes that happen after the form fill, such as a qualified lead, a booked appointment or a closed deal, back to Google Ads and credit them to the original ad click. To do it you need three things: the click identifier (usually the gclid) stored on every lead in the CRM, conversion actions in Google Ads that match the CRM stages you care about, and an import method that uploads those outcomes on a schedule. Once it works, bidding can optimize toward leads that actually turn into revenue instead of every form submission.

For agencies this is the step that changes the conversation with a client. Instead of reporting form fills, you can report which campaigns produced qualified pipeline. This guide walks through the full setup and the failure points.

Updated for the June 15, 2026 change: since that date, new offline conversion and enhanced conversions for leads integrations can no longer use the Google Ads API. They go through Data Manager in Google Ads or the Data Manager API. Our guide to the Data Manager API for conversion imports covers what changed and how to migrate.

Why form fills are not enough

Smart Bidding optimizes toward whatever you call a conversion. If that is a form submission, it will find people who submit forms, including spam, job seekers, existing customers and tyre kickers. Many lead generation accounts have a wide gap between leads and qualified leads, and the gap is not the same across campaigns, keywords or audiences. Importing a later stage tells bidding which clicks led somewhere.

Importing does not replace your website conversion. You keep the form fill as a fast signal and add the later stages as additional conversion actions. How you weight them is covered below.

Step 1: capture click IDs on every lead

With auto-tagging on, Google Ads appends a gclid parameter to the landing page URL. On some iOS traffic, you may instead see gbraid or wbraid, which are click identifiers Google uses where a gclid cannot be used for privacy reasons. Store all three if present.

The capture process:

  1. Confirm auto-tagging is on in the account settings. Without it, there is no gclid.
  2. Read the parameter on landing and save it to a first-party cookie or local storage, so it survives when the visitor browses other pages before converting. A cookie lifetime of 90 days matches the maximum click-through window.
  3. Write it into a hidden form field when the form loads, so the value is submitted with the lead.
  4. Map that field to a CRM property, such as a contact property called gclid, so it is stored on the record.

The Conversion Linker tag in GTM already stores click information in a first-party cookie, but that cookie is for the Google tags themselves. For CRM imports you still need the value in a form field. A small GTM custom HTML tag or a form plugin feature typically handles it.

Where this breaks:

  • Redirects that strip query parameters before the landing page loads.
  • Forms in iframes, such as embedded scheduling tools, which cannot read the parent page's cookie without extra work.
  • CRM deduplication that overwrites the gclid when an existing contact submits again, or keeps the oldest value when you wanted the newest.
  • Leads created manually by sales staff, which never carry a click ID.

The same capture pattern works for UTMs and Meta's fbclid; see CRM lead source attribution in HubSpot and GoHighLevel for the wider setup.

Step 2: map CRM stages to conversion actions

Pick the CRM stages that represent real progress, and create one Google Ads conversion action per stage. A typical lead generation mapping:

CRM stageGoogle Ads conversion actionPrimary or secondaryValue
Form submittedWebsite lead (tag)Depends on volume, see belowLower, estimated value
Qualified lead / MQL / SQLQualified lead (import)Primary once volume allowsHigher, estimated value
Closed wonClosed deal (import)Primary or secondaryActual deal value

Choose stages that sales updates consistently. A stage that only half the team uses produces noise. Agree the definition of "qualified" with the client in writing, because their sales team will change it otherwise.

Create the import conversion actions in Google Ads as conversions imported from clicks (the option for CRMs, files or other data sources). Set the counting to One, choose a click-through window long enough to cover the client's sales cycle (up to 90 days), and set the category to something like Qualified lead or Converted lead.

Which stage should bidding optimize toward?

There is a tension between quality and volume. A closed deal is the best signal but may happen a handful of times a month, too rarely for bidding to learn from. A form fill happens often but includes junk. Common approaches:

  • Low volume: keep the form fill primary, import later stages as secondary, and use them for reporting and manual decisions.
  • Moderate volume: make the qualified lead stage primary alongside the form fill, with higher value on the qualified stage, and bid on value.
  • Healthy volume of qualified leads: make the qualified stage the main primary action and move the raw form fill to secondary.

The value side of this is covered in assigning conversion values to leads.

Step 3: choose an import method

MethodHow it worksBest for
Native CRM integrationHubSpot and Salesforce both offer Google Ads integrations that sync lifecycle stage changes as offline conversionsClients already on those CRMs; least maintenance
Google Ads Data ManagerGoogle's hub inside Google Ads for connecting data sources, including some CRMs, cloud storage and spreadsheetsAccounts that want connections managed in Google Ads itself
Scheduled files through Data ManagerA CSV, Google Sheet or SFTP file with click ID, time and value, connected in Data Manager and imported on a schedule. The older file upload flow in Google Ads is now labelled legacy.CRMs without native integrations, such as GoHighLevel via an automation tool
Data Manager APIDeveloper-built uploads from the CRM or a data warehouse. The Google Ads API only stays open to integrations that were already uploading between December 17, 2025 and June 15, 2026.Larger clients with engineering support

Whatever the method, each uploaded conversion needs the click ID (or user-provided data, see below), the conversion action it belongs to, the conversion time with its time zone, and optionally the value and currency. Files identify the action by its exact name, Data Manager connections let you map or filter actions in the setup, and the Data Manager API uses the numeric conversion action ID. Time zone mismatches are a common cause of rejected or misattributed rows.

The gclid-less alternative: enhanced conversions for leads

If capturing click IDs is impractical, enhanced conversions for leads can match outcomes using the lead's hashed email or phone instead. The website captures user-provided data at form submission, and the upload sends the same identifier when the lead progresses. Many setups use both: the gclid when it is present, and the email as a fallback. The enhanced conversions setup guide explains the website side.

A simple upload file

For scheduled uploads, the file needs a handful of columns. Google Ads Help now labels the old file upload flow in Google Ads a legacy feature and points to Data Manager, where you connect the same kind of file through a file upload, Google Sheets or SFTP connection. Column names follow Google's template or the mapping you set in Data Manager, but the content looks like this:

Google Click ID, Conversion Name, Conversion Time, Conversion Value, Conversion Currency
Cj0KCQ...abc, Qualified lead, 2026-11-02 14:32:00-05:00, 500, USD
Cj0KCQ...xyz, Closed deal, 2026-11-03 10:05:00-05:00, 8200, USD

Download the current template from Google Ads rather than building the header from memory, and keep the conversion name exactly as it appears in the account, including capitalization.

GoHighLevel and other CRMs without a native integration

For CRMs without a direct Google Ads connection, the usual pattern is an automation that writes a row to a Google Sheet when a contact or opportunity reaches the mapped stage, and a Google Sheets connection in Data Manager that imports the sheet on a schedule. Make sure the automation only triggers once per contact per stage, or you will upload duplicates. With counting set to One, Google Ads ignores repeat rows for the same click and action, but it is better not to rely on that.

Timing windows

  • Upload within the click-through window. A conversion that happens after the action's window has closed will not be accepted. Clicks older than 90 days cannot be credited at all.
  • Upload promptly. Bidding learns faster from fresh data. Daily uploads are a sensible default; real-time is better if the integration supports it.
  • New conversion actions need a short wait before they accept imports. Create the action, wait, then run the first upload.
  • Expect a delay in reporting. Imported conversions can take several hours to appear, and they are reported against the date of the click by default, which means last week's numbers keep changing as deals close. Explain this to clients.

Conversion adjustments

Google Ads supports adjustments to conversions already recorded, but what you can do depends on the upload path:

  • Restatement: changes the value, for example when the final contract value differs from the initial estimate. In the Data Manager API you re-send the event with the same transactionId (order ID) and the new value; the conversion count stays the same.
  • Retraction: removes a conversion, for example when a "closed won" deal is cancelled. The Data Manager API does not support retractions. They only remain on older Google Ads API integrations that are still allowlisted.

Because retractions are not available on the new path, map conversions to CRM stages that rarely reverse, and send an order ID with every conversion so restatements are possible later. Most agency setups do not need adjustments at first; add them once the basic import is stable.

Testing the import

  1. Check capture first. Visit the site with a test gclid parameter appended to the URL, browse a couple of pages, submit the form, and confirm the value lands in the CRM record exactly as it appeared in the URL.
  2. Check real leads. Filter recent paid search leads in the CRM. Most should have a gclid. If many do not, look for redirects, iframes or forms missing the hidden field.
  3. Run a real upload. Move a genuine paid lead to the mapped stage, or wait for one, and check the upload history in Google Ads. Test values with made-up gclids will be rejected, which is expected.
  4. Read the errors. Upload diagnostics list rows that failed and why. Typical reasons are clicks that cannot be found, clicks older than the window, conversion action names that do not match, and clicks that belong to a different Google Ads account.
  5. Reconcile weekly at first. Compare the number of leads moved to each stage in the CRM with the number of imported conversions accepted. A gap is normal (leads from other channels, missing click IDs), but it should be stable and explainable.

Common mistakes

  • Importing with conversion tracking in a different account. If the account uses cross-account conversion tracking from a manager account, the import conversion action must live where conversion tracking is managed.
  • Making the closed deal primary with too little volume. Bidding gets a handful of signals a month and becomes erratic.
  • Overwriting the gclid on repeat submissions. Decide whether the first or latest click should be credited, and configure the CRM field accordingly.
  • Forgetting consent. For EEA traffic, uploads should respect the consent the user gave, and Google's upload formats include consent fields for that purpose.
  • Not telling the client that past numbers will change. Imported conversions backfill to the click date.

What we check in an audit

  • Auto-tagging status and whether click IDs survive to the form.
  • The share of recent paid leads in the CRM that carry a gclid, gbraid or wbraid.
  • CRM stages that are used consistently enough to import.
  • Existing import actions, their windows, and upload errors.
  • Whether enhanced conversions for leads is a better fit than gclid capture.

Where to start

Most of this work sits in our CRM lead attribution service: click IDs captured, mapped to CRM properties and ready for import, starting with a free tracking check of the client's current setup.

Just one quick fix, not a whole account? Our founder takes small fixed-price jobs on Fiverr: Google Ads conversion tracking fix

Frequently asked questions

Do I need the gclid to import offline conversions?

It is the most direct method, but not the only one. Enhanced conversions for leads can match outcomes using the lead's hashed email or phone, and many setups use both.

What are gbraid and wbraid?

They are click identifiers Google Ads uses on some iOS traffic where a gclid cannot be used. Capture and store them alongside the gclid so those clicks can still be credited.

How long after the click can I upload a conversion?

Within the conversion action's click-through window, which can be up to 90 days. Upload daily if you can, because bidding learns faster from fresh data.

Should the imported stage be primary?

Only once it happens often enough for bidding to learn from. With low volume, keep it secondary for reporting and keep the form fill as the primary action.

Does HubSpot support Google Ads offline conversions?

Yes. HubSpot's Google Ads integration can sync lifecycle stage changes as offline conversions, provided the contacts were created from tracked ad clicks.

Muhammad Qitmeer

Written by

Muhammad Qitmeer, founder of Augere Labs

Augere Labs is the white-label tracking partner behind marketing agencies. If a client account has numbers you don't trust, I'm happy to look at it with you and tell you straight what's wrong.

Connect on LinkedIn