/projects//settings/integrations/ ).
1. The Flaw of Optimizing for Lead Volume
When an agency optimizes a B2B search campaign for raw lead volume, Google Ads machine learning optimizes for users most likely to fill out a form—often students, job seekers, and bots. Passing qualified sales stages (SQL, Opportunity, Closed-Won) back to Google Ads recalibrates the bidding engine toward real revenue.
LIVE PLATFORM SCREEN & TELEMETRY
SEO Runway Project Settings & Integrations — Active Workspace
Below is the exact production screen with live parameters and dummy data illustrating the diagnostic findings and remediation buttons.
2. Implementation Architecture with SEO Runway
Connect your tracking credentials in SEO Runway → Project Settings → Integrations to align CRM pipeline revenue directly with paid search spend.
Target System: SEO Runway PPC Engine
Execution Frequency: Automated Background Queue (UTC Sync)
Tested across enterprise datasets with zero rate limiting or data loss.
3. How GCLID-Based Offline Conversion Tracking Works
Every click on a Google Ads ad can carry a unique Google Click Identifier (GCLID), appended to the landing page URL as a gclid parameter when auto-tagging is on. Offline conversion tracking (OCT) works by catching that identifier when a lead arrives, storing it alongside the lead in your CRM, and later sending it back to Google Ads with the outcome: "this click became a qualified lead", "this click became a customer worth $12,000". Google then attributes the offline result to the original click, keyword, ad and campaign.
The process has four stages, and each one has its own failure points:
- Capture: the website reads the GCLID from the URL and keeps it, usually in a first-party cookie, so it survives page navigation.
- Store: when the visitor submits a form, the GCLID is passed through a hidden field into the CRM and saved on the lead or contact record.
- Qualify: sales moves the lead through pipeline stages, such as MQL, SQL, Opportunity and Closed-Won.
- Upload: the GCLID, conversion action name, conversion time and value are sent to Google Ads, by scheduled file upload, API or a native CRM integration.
4. Setting Up the Capture Layer Correctly
Most broken OCT setups fail at capture. Visitors land with a GCLID, browse three pages and submit a form on a page whose URL no longer contains the parameter. Unless something stored it on arrival, the identifier is lost.
- Confirm auto-tagging is on at account level, and test that your site does not strip URL parameters during redirects. Some CMS redirects and URL-rewriting rules silently drop query strings.
- Store the GCLID on landing. A small script, or a tag manager variable, reads
gclidfrom the URL and writes it to a first-party cookie with a 90-day lifetime. The Google tag's conversion linker also stores click information in first-party cookies; make sure it is deployed on every page. - Populate hidden fields. Every lead form needs a hidden field that reads the stored value on submit. Include forms in pop-ups, chat widgets and multi-step flows; teams often add the field to the main form and forget the rest.
- Handle iOS and privacy cases. Some clicks arrive with
gbraidorwbraidinstead of a GCLID, notably from iOS traffic. Capture those parameters too, and plan to send enhanced conversion data for leads where identifiers are missing. - Test end to end. Click your own ad (or append a test GCLID), submit a form and confirm the value appears on the CRM record exactly as it was in the URL.
5. Designing Conversion Actions for the Sales Pipeline
Create one offline conversion action per pipeline stage you want to measure, set to "Import" as the source. A typical B2B structure looks like this:
- Lead submitted (online tag): every form fill, used for volume tracking only.
- Marketing qualified lead: leads that meet basic fit criteria, such as company size or industry.
- Sales qualified lead: leads that sales accepts after a first conversation.
- Opportunity: deals with an active proposal.
- Closed-won: signed customers, with the actual contract value.
Decide which actions are "primary" (included in the Conversions column and used for bidding) and which are "secondary" (observed only). Bidding on closed-won alone often starves the algorithm of data, because B2B accounts may only close a handful of deals per month. A common compromise is to bid on SQLs, with assigned values, while recording opportunities and closed-won deals as secondary actions for reporting.
Assigning values to each stage
If you use value-based bidding, give each stage a value based on its historical probability of becoming revenue. If 20% of SQLs close at an average contract value of $10,000, an SQL is worth about $2,000 to the business. Values do not have to be perfect, but they must be proportional, so the algorithm knows an SQL matters far more than a raw form fill.
6. Uploading Conversions: Timing, Format and Windows
Each uploaded row needs the GCLID, the conversion action name exactly as it appears in Google Ads, the conversion time with a time zone, and optionally a value and currency. A few rules decide whether uploads count.
- Respect the click-through window. A conversion can only be imported if it happened within the conversion action's click-through window, which can be set up to 90 days. If your sales cycle is longer than 90 days, track an earlier stage that happens inside the window.
- Upload frequently. Smart Bidding learns from recent data. Daily uploads, or more often, let bidding react to pipeline changes; monthly batches leave the algorithm optimising on stale information.
- Use the correct conversion time. Send the time the stage change happened, not the upload time, and never a time earlier than the click itself, or the row will be rejected.
- Avoid duplicates. Upload each GCLID and action pair once. Use order IDs where supported, and keep a log of what has already been sent.
- Check the diagnostics. The Uploads page in Google Ads shows successful and failed rows with reasons. Review it after every scheduled upload for the first few weeks.
Automating the pipeline
Manual CSV uploads work for a pilot but rarely survive past the first busy month. Use your CRM's native Google Ads integration where one exists, or a scheduled job that queries the CRM for stage changes in the last day and pushes them through the Google Ads API. Store the upload status on each CRM record, so failures can be retried and audited later.
7. Enhanced Conversions for Leads
Browser privacy changes and cross-device journeys mean some leads arrive without a usable click identifier. Enhanced conversions for leads fill part of that gap. When a form is submitted, the Google tag captures hashed first-party data such as the email address. Later, when you upload the offline conversion with the same hashed email, Google matches it to the signed-in user who clicked the ad. Setting it up alongside GCLID capture improves match rates and makes your offline data more resilient. Make sure your privacy policy and consent handling cover this use of customer data.
8. Common OCT Failures and How to Diagnose Them
- Low match rate: fewer than 70% of paid leads carry a GCLID in the CRM. Check redirects, cookie storage and every form's hidden field.
- Uploads rejected as "click too old": the conversion happened outside the action's window. Shorten the stage you track or extend the window.
- "Conversion action not found": the name in the file does not exactly match Google Ads, including capitalisation and spaces.
- Values in the wrong currency: set currency explicitly in every row, especially in multi-region accounts.
- Bidding swings after switching goals: changing the primary conversion action resets what the algorithm optimises for. Expect a learning period and avoid switching again until it settles.
9. Switching Bidding to Pipeline Outcomes
Once offline conversions have been flowing reliably for four to six weeks, with enough volume (roughly 30+ conversions per month per campaign is a reasonable floor), move bidding from form fills to the qualified stage. Do it gradually.
- Keep form fills as primary at first, with qualified stages recorded as secondary, and compare which campaigns and keywords drive SQLs, not just leads.
- Pick one campaign with healthy SQL volume and switch its primary goal to SQLs, using Maximise conversions or Maximise conversion value with no target to begin with.
- After two to three weeks, set a target CPA or target ROAS based on the observed cost per SQL.
- Roll the change out to other campaigns one at a time, watching for budget shifts and volume drops.
Expect raw lead volume to fall. That is often the point: the algorithm stops buying cheap, unqualified form fills and starts finding the searchers who become customers. Judge success on cost per SQL, pipeline value and closed revenue, not on form submissions.
10. OCT Implementation Checklist
- Confirm auto-tagging is on and parameters survive every redirect.
- Store GCLID, gbraid and wbraid on landing in first-party cookies.
- Add hidden fields to every lead capture form and test each one.
- Create one import conversion action per pipeline stage, with values.
- Automate daily uploads and monitor the upload diagnostics.
- Enable enhanced conversions for leads for better matching.
- Move bidding to qualified stages gradually, one campaign at a time.
11. Reporting Pipeline Results Back to Stakeholders
Offline conversion tracking changes the conversation with sales and leadership. Instead of reporting "320 leads this month", you can report how many became qualified, how much pipeline paid search created and which campaigns generated closed revenue. Build a monthly report that follows the funnel stage by stage.
- Funnel by campaign: clicks, leads, MQLs, SQLs, opportunities and closed-won deals for each campaign, with the conversion rate between each stage.
- Cost per stage: cost per lead, cost per SQL and cost per closed deal. A campaign with an expensive cost per lead but a cheap cost per SQL is often your best performer.
- Pipeline and revenue: total pipeline value created and revenue won, compared with ad spend, giving a return figure finance will recognise.
- Keyword quality: the search terms that produce SQLs versus those that only produce form fills. Use the second group to build negative keyword lists.
Remember that conversions in Google Ads are reported against the date of the click, not the date of the sale. A deal that closes in March from a January click appears in January's numbers once uploaded. Explain this to stakeholders, and freeze reported figures only after enough time has passed for most deals from that period to close.
Keeping sales on board
OCT only works if sales updates CRM stages promptly and consistently. Agree on clear definitions for each stage, show sales how their updates improve lead quality, and review stage hygiene monthly. A pipeline where half the deals sit in "New" forever produces upload data that misleads the bidding algorithm.
Want the complete procedural runbook?
Visit our comprehensive step-by-step Knowledge Base for interactive screen maps, configuration parameters, and adjacent module workflows.
Open Knowledge Base Hub →

Fantastic actionable guidance. The step-by-step screenshots and UI coordinates make execution completely friction-free. As a Performance Marketing Director, this provides exactly the architectural depth we look for.