Skip to content
NagentNagent
Log inSign upHire your AI Growth Team
NagentNagent
Log inSign up

How to send Salesforce opportunity data to Google Ads so campaigns optimize for pipeline not leads

9 Minutes read
Published:
Capture Google's click ID on every lead, carry it onto the Salesforce opportunity, create import conversion actions for the stages that mean pipeline, upload them with values, and make them the bidding goal. Here are the steps, the checks and the failure modes.
PB
Pratap BeheraNov 17, 2026·9 min read
How to send Salesforce opportunity data to Google Ads so campaigns optimize for pipeline not leads

Capture Google's click ID on every web lead, store it on the Salesforce lead and carry it onto the opportunity when the lead converts. In Google Ads, create an import conversion action for each stage that means pipeline, such as qualified lead, opportunity created and closed won, with values. Upload stage changes daily through Google's data tools, a connector or the API, then make those actions the primary bidding goal so Smart Bidding chases opportunities, not form fills.

What does "optimise for pipeline" mean in Google Ads?

Google's automated bidding optimises toward whichever conversion actions you mark as primary goals. If the only primary conversion is "form submitted", the system learns to find cheap form submissions. If the primary conversions are "opportunity created" and "closed won", with values, it learns to find the searches and audiences that produce those.

So the task has two halves:

  1. Data: get Salesforce stage changes into Google Ads, matched to the original click.

  2. Bidding: tell Google those stages are what you are paying for.

Most teams do the first half and forget the second, which leaves the new data sitting in a report.

What do you need before you start?

A short checklist. If any line is missing, fix it first.

  • Auto-tagging on in Google Ads, so every ad click carries a click ID in the landing page URL.

  • A form that can hold hidden fields and a small script that reads the click ID from the URL (and from a first-party cookie on later visits) into a hidden field.

  • A custom field on the Salesforce Lead for the click ID, and a matching field on the Opportunity.

  • Lead field mapping in Salesforce so the lead's click ID is copied to the opportunity when the lead is converted.

  • Agreed stage definitions. Marketing and sales sign off on what "qualified" and "opportunity" mean.

  • Values per stage, even rough ones.

  • Consent handling that matches your markets, because Google asks you to state user consent on uploads.

Step 1: How do you capture the click ID and keep it through lead conversion?

Google's click identifier is the GCLID. Some iOS traffic carries gbraid or wbraid instead, and Google's upload format accepts all three (Google Ads API).

  • Read the parameter from the landing page URL and store it in a first-party cookie, so a visitor who browses three pages before filling the form still carries it.

  • Write it to the hidden field on every form, including demo requests, contact forms and chat-to-lead flows.

  • Save it on the Lead record. Also save the email address, because enhanced conversions for leads can match on a hashed email when the click ID is missing.

  • When sales converts the lead, the mapped field copies the click ID onto the Opportunity. Check this with a test lead before you go further. A missing mapping is the most common reason opportunity uploads match nothing.

Step 2: Which Salesforce stages should you send, and with what values?

Send more than one stage. Closed deals are the truest signal but arrive too late and too rarely for bidding to learn from.

A practical set:

  • Sales-qualified lead. Happens within days, often enough to bid on.

  • Opportunity created. The pipeline signal.

  • Closed won. The revenue signal, with the real amount.

Worked example of values (ours, not a benchmark). Average first-year contract value is $24,000. One in four opportunities closes, and one in three qualified leads becomes an opportunity.

  • Closed won: send the actual amount.

  • Opportunity created: $24,000 times 0.25 equals $6,000.

  • Qualified lead: $6,000 times 0.33 equals about $2,000.

These expected values let value-based bidding compare a search that brings one opportunity with one that brings five unqualified leads.

Step 3: How do you set up the conversion actions in Google Ads?

  • Create one import conversion action per stage, named for the stage: "SF Qualified lead", "SF Opportunity", "SF Closed won".

  • Set each to count one conversion per click, so a lead that moves through stages is not counted twice within the same action.

  • Set a default value, and send the real value on each upload.

  • Set the click-through window long enough to cover your cycle. Optmyzr notes that click IDs expire 90 days after the click (Optmyzr), so anything that happens later cannot be matched by click ID.

  • Decide which actions are primary. Start with qualified lead and opportunity as primary, and set the raw form fill to secondary so it is still reported but no longer steers bidding. Keeping both the form fill and the later stages primary double-counts the same person.

Step 4: Which upload route should you use?

There are three routes. Pick by your team's skills and volume.

  • Google's own data tools. Google describes Data Manager in Google Ads as a place to connect "all your online and offline data sources" (Google). Check which CRM connections your account lists before building anything.

  • A connector or integration platform. Many teams use an iPaaS or a marketing data tool that watches Salesforce stage changes and posts them to Google. Check that it sends values, uses the conversion time rather than the sync time, and supports consent fields.

  • Your own build on Google's APIs. Google's developer docs still describe the Google Ads API upload method, with a warning: from 15 June 2026, those upload requests fail if the developer token has never sent offline conversions or enhanced conversions for leads before, and Google points to the Data Manager API instead (Google Ads API). Data Manager API offers a validate-only mode for testing payloads before you send real data (Google).

Whatever the route, each upload needs: the click ID or hashed email, the conversion action, the time the stage changed (after the click, with a time zone offset), the value, the currency code and a unique ID per conversion, which Google recommends so you can adjust conversions later.

Step 5: How often should you upload, and when will results show?

  • Daily is enough for most B2B accounts. Hourly adds little.

  • Use the stage-change time, not the time the sync ran.

  • Expect lag. Google says imported conversions are reported against the date of the original click, and that they can take up to three hours to appear. So the last few weeks always look weaker than they are, because their opportunities have not happened yet. Judge campaigns on cohorts old enough to have matured.

  • Retract and restate. When an opportunity is lost or its amount changes, send an adjustment using the conversion's unique ID.

Step 6: When do you switch bidding to the pipeline goal?

Only once the data is flowing and there is enough of it.

  • Run the new actions as secondary for two to four weeks and compare counts with Salesforce. If Google shows far fewer opportunities than Salesforce attributes to paid search, fix matching first.

  • Make the stage that happens often enough your primary goal. A practitioner on Optmyzr's PPC Town Hall suggests about 30 qualified leads a month before offline tracking is worthwhile (Optmyzr).

  • Move to Maximize conversion value, then add a target ROAS once a few weeks of valued data exist. Change one thing at a time.

What if your sales cycle is longer than 90 days?

Then closed won will rarely match a click ID, so lean on earlier signals. Optmyzr's Town Hall on two-year cycles describes scoring early leads and sending that predicted probability back as a value, and aiming to return a signal within about two weeks rather than waiting out the window (Optmyzr, August 2026). Enhanced conversions for leads also helps here, because it can match on hashed email when no click ID survived.

What usually goes wrong, and what does each failure cost?

  • Click ID lost at lead conversion. Opportunities upload but match nothing. Cost: weeks of "no data" and a wrong conclusion that paid search produces no pipeline.

  • Sync time used as conversion time. Conversions land on the wrong days and some are rejected as before the click.

  • Duplicate uploads. No unique ID, so retries double count. Bidding overvalues the campaigns involved.

  • Form fill left as primary. Bidding keeps chasing cheap leads, and the new stages only dilute the signal.

  • Values in mixed currencies or left at defaults. Value-based bidding cannot rank anything.

  • Consent not stated. Google's docs warn conversions may not be attributable without the consent field.

Where does Nagent fit?

We make Nagent. Nagent is multiplayer AI for growth: people and AI agents in one shared context, delivering the number. NIA, the AI performance marketing lead on the Marketing team under MIRA, reads your connected Google Ads account, including conversions, cost per conversion and return on ad spend.

  • Once your Salesforce stages arrive in Google Ads as conversions, she reads them like any other conversion and can answer "which campaigns produce opportunities?" in plain English.

  • For a leads or sales objective, she will not build a new campaign without an enabled conversion action on the account.

  • She can propose a new target CPA on a campaign. Every live change waits for a person to approve on Actions.

What she does not do: move data from Salesforce to Google Ads. The click ID capture, field mapping and upload pipeline above are yours to build. In Nagent, Salesforce connects as a tool through Composio, and importing CRM records into Nagent is not built. Learn more about NIA.

Frequently Asked Questions

What are enhanced conversions for leads in Google Ads?

A way to match offline conversions using a hashed email (normalised and SHA-256 hashed) from the lead form, instead of, or as well as, the click ID. It helps when click IDs are lost.

How do I set up value-based bidding in Google Ads?

Send a value with each conversion, starting with expected values per stage, run for a few weeks, switch to Maximize conversion value, then add a target ROAS once results are steady.

Does the same approach work for HubSpot or other CRMs?

Yes. The steps are the same: capture the click ID, store it on the record, send stage changes with time and value. Only the field names and sync tool differ.

How long before Google's bidding responds to opportunity data?

Expect several weeks. The system needs a run of uploaded conversions before it shifts spend, and recent clicks will not have matured into opportunities yet.

Want NIA to read your Google Ads conversions and show which campaigns create pipeline? Book a demo.

AI agents for SaaS & Technology

Continue learning

Related agents

Agents that match this read.

Browse all 200+ agents
Select Category
Performance Marketing

How are B2B sales teams using AI to run Google Ads and LinkedIn ads?

Mostly in narrow, supervised jobs: Google's own bidding and AI Max fed with CRM outcomes, AI agents that triage search terms and draft copy for a person to approve, and LinkedIn creative variants per buyer role. Here is what works, the guardrails, and what fails.

Sales

Can an AI agent keep Salesforce opportunity data clean for a small sales team?

Mostly, yes. An AI agent can log activity, fill next steps from calls, flag stale close dates and duplicates, and propose stage changes for the owner to confirm. It cannot fix data that never reached Salesforce, and it should not change amount, stage or owner alone.

Performance Marketing

Performance Max vs Demand Gen for lead generation

Performance Max suits lead generation when you can feed it qualified-lead data, since it chases conversions across Search and all Google surfaces. Demand Gen builds demand with video and images on YouTube, Discover and Gmail, with more control and no Search.

Performance Marketing

Does AI creative testing work with Meta Ads Manager, Google Ads and Shopify data?

Mostly yes for Meta, partly for Google Ads, and usually indirectly for Shopify. Creative tools read ad-level results from the ad platforms; Shopify sales reach them through the platforms' conversion tracking or an attribution tool.

Performance Marketing

How do I set up NIA to optimise Google Ads and Meta ads campaigns?

Install the Marketing team, connect Google Ads through Google's consent screen and sync, connect your Meta ad account and set its daily and monthly hard spend cap, confirm conversion tracking, then run a diagnosis and approve NIA's proposals on Actions. Step by step.

AI Strategy

Can a ChatGPT agent run your Google Ads campaigns?

ChatGPT can draft headlines and suggest keywords, but drafting is not running. A real campaign agent needs standing account access, fixed checks, human approval and measurement loops.