How to Track Leads from Google Ads into a CRM (Click to Closed-Won)

The pipeline in one picture

Every piece below exists to move one record through five stages. If any stage is missing, the chain breaks silently: ad platforms optimise towards whatever they can see, which is usually raw form fills. The rest of this article builds each stage, with the fields, settings and decisions each one needs.

• Click: a person clicks your ad; Google attaches a GCLID to the landing page URL.
• Capture: your form grabs the GCLID and campaign details into hidden fields.
• Record: the CRM stores them on the contact alongside the enquiry.
• Outcome: your team qualifies the lead and, eventually, wins or loses the deal.
• Return: qualified and closed-won outcomes are imported back into Google Ads as offline conversions.

GCLID and UTM: what each one does

The GCLID (Google Click Identifier) is a unique code Google appends to your landing page URL for every ad click when auto-tagging is on. It is how Google later connects a CRM outcome to the exact click that started it. Check it in Account Settings → Account settings → Auto-tagging; it should be enabled, and it is what makes offline conversion imports possible at all.

UTM parameters (utm_source, utm_medium, utm_campaign, utm_term) are the generic version of the same idea — readable labels any analytics tool or CRM can parse. Google Ads appends utm parameters through a tracking template if you configure one (for example ValueTrack parameters like {campaignid}, {adgroupid}, {keyword} and {matchtype}).

The practical division of labour: the GCLID is for Google's own import machinery — it is opaque, but exact. UTM parameters are for your CRM's own reporting and for any non-Google channel sharing the same fields. Store both; they answer different questions and cost nothing to keep.

Form submissions: capturing the click into the lead

Your forms are where an anonymous click becomes a named lead, so this is the single most important implementation step. The pattern: hidden form fields read the URL parameters and submit them alongside the name and email.

• Hidden fields to include: gclid, utm_source, utm_medium, utm_campaign, utm_term (or the ValueTrack equivalents), plus the landing page path and, if you use them, the first and referrer pages.
• Persist the parameters: visitors rarely convert on the first page. Store the parameters in a cookie or local storage on arrival (first-touch) and overwrite or keep latest-touch values as they browse, then populate the hidden fields at submit time from the stored values.
• Phone calls need their own path: a tel: click can fire a conversion, but the caller's GCLID only reaches the CRM through a call-tracking product that records it per call. If calls are a major lead source, decide this deliberately rather than assuming the form covers it.
• Google lead form extensions (lead assets) skip your website entirely, so no GCLID reaches your CRM from them — leads arrive through the native download or API with campaign attribution, not click IDs. If CRM-level tracking matters to you, treat lead assets as a separate stream with its own reconciliation, or weigh whether website forms suit you better.

One warning from experience: test with a real submission and check what actually lands in the CRM — not what the form preview shows. Hidden fields that read the URL directly fail for anyone who converted on a later page, which is most people.

CRM fields: what to add before the first lead arrives

Add these as custom properties on the contact or deal record before launch — retrofitting fields onto months of old leads is the slowest way to do this:

• GCLID (the exact click ID — non-negotiable, everything downstream depends on it).
• Campaign, ad group and keyword or search term as readable text, from the UTM/ValueTrack parameters.
• Source and medium (google / cpc for paid search; separate values for organic, direct, referral, other paid channels).
• First and last landing page, and the form or asset that produced the lead.
• Lead stage (e.g. New → Working → Qualified → Unqualified) and deal stage up to Closed-won / Closed-lost.
• Deal value — even a rough expected value per service line makes revenue-based optimisation possible later.
• Dates: lead created, qualified, closed. The intervals matter for conversion windows and for diagnosing slow stages.

Lead qualification: deciding what counts

Raw form fills are a weak optimisation target because most accounts can lower cost per lead trivially by attracting easier leads. The fix is a defined qualification stage in the CRM, agreed between whoever runs the ads and whoever takes the leads.

Keep it binary and fast to apply: Qualified / Unqualified plus a reason. A two-question checklist your sales team can apply in ten seconds survives; a lead-scoring model nobody maintains does not. If you use MQL/SQL terminology, map them to the same stages — the names matter less than the discipline of recording the outcome on every lead.

This is also where lead quality feedback starts. When a campaign produces many leads but few that qualify, that pattern is only visible once someone is marking leads — and it is often the highest-value finding in the whole exercise.

Closed-won revenue: the number worth optimising for

Closed-won is where ad decisions should ultimately be made. A campaign that produces expensive leads but closes at three times the rate is better than a cheap-leads campaign that stalls in the pipeline — but you can only see that if deals carry a value and a win/loss outcome.

Record the actual deal value (or a standard expected value per service line when deals vary). When you later import offline conversions, you can send the deal value with the outcome, and Google's bidding can weight towards larger deals — this is what value-based bidding on imported conversions does.

A caveat worth stating plainly: closed-won reporting needs enough volume to mean anything. If you close a handful of deals a month, monthly closed-won numbers will be noisy; use them as a directional signal over a quarter, not a weekly scoreboard.

Importing outcomes back into Google Ads

Once the CRM holds GCLIDs and outcomes, Google Ads can learn from them. Two mechanisms, often confused:

• Enhanced conversions for leads: sends hashed lead information (email, phone) from the lead event, then from the CRM on qualification — suited when you cannot reliably store GCLIDs (for example, calls or lead assets), and available even without GCLID capture.
• Offline conversion imports: uploads conversions against the stored GCLID — 'Qualified lead' and 'Closed-won' as separate conversion actions, each with its own window matching your real sales cycle.

Import what your bidding should chase, not everything you can measure: typically the qualified-lead event (the volume the algorithm needs) plus closed-won with value (the goal). Google's guidance is roughly 30 conversions of an action per month for smart bidding to optimise on it — if closed-won volume is far below that, optimise on qualified leads and monitor closed-won as a reporting layer instead.

The step-by-step mechanics of the import — file format, naming, windows and the scheduling options — are covered separately in offline conversion imports for B2B.

Where GA4 fits

GA4 answers a different question from the CRM pipeline: what did all channels contribute, in one comparable frame? Set it up alongside, not instead of, the CRM connection.

• Mark form starts and form submissions as key events in GA4 (the full setup is in the GA4 conversion tracking guide).
• Import GA4 key events into Google Ads for actions that happen on-site, and use offline imports for CRM outcomes — one source of truth per action, never both, or you double count.
• Use GA4's channel reporting to see how paid search interacts with organic, direct and returning visitors — something Google Ads alone cannot show you.

Expect the counts to disagree with Google Ads — different attribution models and definitions mean they always will, and a 10–20% gap is normal. The CRM, not either platform, is the source of truth for revenue.

Attribution: which number to use for what

With three systems reporting, assign each a job and stop expecting them to match:

• Google Ads (data-driven attribution on its own clicks): the signal the bidding algorithm optimises on. Use it for campaign management.
• GA4 (last non-direct click, all channels): the fair cross-channel picture for marketing reporting.
• CRM (reality): who actually became a customer and what they paid. Use it for business decisions — budget levels, which services to push, whether ads pay for themselves.

The CRM view is last-click-by-source, so it under-credits upper-funnel touches — a prospect who clicked an ad in January and typed your brand name in March looks organic. Read CRM source reports as one input alongside the platforms, not as perfect credit assignment. Directional reality beats precise fiction.

Common mistakes that break the chain

• Auto-tagging off — GCLIDs never arrive, and offline imports have nothing to attach to.
• Hidden fields reading the URL at submit time — empty for anyone who browsed first, which is most converters.
• Storing campaign names but not the GCLID — readable reporting works, but imports become impossible.
• Importing every lead as a conversion — smart bidding then optimises for volume of low-quality fills; import qualification stages instead.
• Conversion windows shorter than the real sales cycle — closed-won outcomes arriving after the window are lost to optimisation.
• No written definitions for qualification stages — two salespeople marking the same lead differently corrupts the data silently.
• Rebuilding the same fields in three tools with no owner — pick the CRM as the system of record and let the platforms receive from it.

A realistic implementation order

For most businesses this is one to two weeks of part-time work, and the order matters — each step makes the next possible:

1. Enable auto-taggingConfirm in account settings; check a live ad click carries the GCLID parameter.
2. Add CRM fieldsGCLID, campaign parameters, source, stages, value, dates — before any new lead arrives.
3. Wire the formsHidden fields fed from stored parameters; test with a real submission end-to-end into the CRM.
4. Define qualificationOne written definition, applied to every lead, recorded as a CRM stage.
5. Align GA4Form start and submit as key events; import into Ads only actions not already imported offline.
6. Start importsQualified lead first; add closed-won with value once the volume justifies it.
7. Review monthlyCompare campaign-level qualified rates and closed-won value in the CRM against Ads cost — then adjust budgets.

When to get help

If you have a CRM and running campaigns but the two have never met, the audit that matters is a data one: check auto-tagging, submit a test lead and follow it into the CRM, and confirm what the platforms are actually optimising on. Our independent paid media audit ($299) does exactly that review. If you would rather have the tracking built and the campaigns managed against qualified outcomes, paid media management starts at $399/month — and the tracking setup itself is covered under conversion tracking. We are not the cheapest option, and this is one area where cheap shows.