01 / Strategy first
Choose the signal before the connector.
A cheap lead is not necessarily a good customer. Start with the sales outcome you want more of, then decide which CRM event represents it.
| Stage | What it tells you | Measurement role |
|---|---|---|
| Form submitted | Someone raised a hand | Diagnostic volume |
| Qualified lead | Sales confirms fit and intent | Potential optimization signal |
| Closed won | A customer was acquired | Revenue and CAC reconciliation |
Define qualification with sales: customer segment, buying need and a real next step. Use the same definition across campaigns. Keep rejected leads and their reasons so you can see which searches attract the wrong buyers.
For long sales cycles, a reliable qualified-lead event may arrive sooner than a sale. Report both stages separately; do not treat them as two new customers.
02 / Implementation
Build a data contract, not just a sync.
Google documents CRM connections through Data Manager, including HubSpot and Salesforce. A connector moves data; it does not decide what a qualified lead means or fix inconsistent sales records.
1. Preserve the matching signals.
Enable auto-tagging and retain available click identifiers through the landing page, form and CRM record. The GCLID (Google Click Identifier) links an eligible offline conversion to the originating Google Ads click. Preserve it through redirects and form submission, then associate it with the lead and subsequent sales events. Do not overwrite the original acquisition record every time a returning visitor clicks another ad.
For enhanced conversions for leads, configure the appropriate first-party data collection. Google recommends including GCLIDs when available; hashed customer data can supplement the match. A GCLID is a matching key, not a readable campaign name or a guarantee that every sale will be attributed.
Google: set up GCLID-based offline conversions2. Make lifecycle events explicit.
Store the lead ID, stage, actual stage-change timestamp, currency and your agreed conversion value. Create separate conversion actions for qualified and converted leads. Use stable event identifiers to prevent the same stage change from being imported repeatedly.
3. Choose an import path you can maintain.
Use a supported CRM connector where its fields and events fit. For a legacy CRM, map and transform records into a supported import format. For custom API work, check Google's current Data Manager migration guidance rather than starting from an old Google Ads API tutorial.
4. Respect consent and customer data rules.
Collect and use first-party data with the required notices, consent and permissions. Follow Google's Customer Data Terms and tagging requirements. Hashing is a matching mechanism, not a substitute for permission.
Three practical ways to connect the systems.
Choose the simplest path that preserves an auditable record. Centralized reporting and conversion import are separate jobs: joining sales to campaign data in a warehouse does not automatically send those outcomes back to Google Ads.
- A / Supported CRM connector
- Best when qualification and closed-won events already live reliably in one CRM. Map the supported lifecycle events and fields, confirm the import destination, and monitor failures. A HubSpot or Salesforce connection can reduce manual handling, but a connected contact list is not proof that downstream conversion events are flowing.
- B / Excel reconciliation and supported file import
- Useful for a low-volume pilot or legacy sales system. Export lead and sales records, join them using stable lead, customer or order IDs, and reconcile unmatched rows in a controlled Excel workbook. Keep the raw export separate from the cleaned event table, document the join rules, and prevent duplicate stage events. Excel is the preparation layer: export into Google's currently supported import template rather than assuming an arbitrary workbook can be uploaded. Assign an owner and a regular upload schedule; manual work becomes fragile as volume grows.
- C / Warehouse or centralized data system
- Best when CRM, ecommerce, invoicing and offline sales must be combined repeatedly. Load source records into a controlled warehouse, resolve customer identities, and build a separate conversion-event table. Keep lineage back to each source record, checks for missing keys, and a log of accepted and rejected exports. Choose a currently supported Data Manager source or documented API path for activation, checking the specific conversion use case rather than assuming every warehouse connector supports it.
A minimum event contract
For internal reconciliation, retain a stable lead/customer ID, acquisition matching signals, stage, actual event timestamp with time zone, event ID, value and currency. Add campaign and customer-segment fields for analysis where available. The destination's current schema determines which fields are sent to Google. Keep pipeline estimates, booked revenue and contribution margin distinct, and define how cancellations or corrected values will be handled.
For high-value leads, trace a sample from the originating interaction to qualification, opportunity and sale. Use the GCLID and permitted first-party matching signals for advertising attribution, and stable customer/order IDs for reconciliation across systems. Import promptly within Google's applicable windows; do not wait until a quarterly reporting exercise to discover that events are no longer eligible.
Google: import timing, duplicates and data guidelines03 / Quality control
Prove the loop before changing bidding.
Google recommends initially keeping a new conversion action secondary for two to three weeks. Use that period to check the signal, not to promise a performance lift.
- Follow a permitted test record from landing-page capture through the CRM to import diagnostics. Confirm the stage, timestamp and value.
- Compare eligible CRM events with accepted imports. Investigate rejected records, missing identifiers and low match coverage.
- Replay an export and check that duplicate events are not counted again.
- Reconcile one cohort using consistent dates, currency and value definitions. Separate pipeline value from booked revenue and gross profit.
- Choose the bidding goal only after checking reliable volume, sales-cycle delay and the campaign's conversion settings.
Report spend, qualified opportunities, acquired customers and revenue together. Segment by campaign and customer type when the data supports it. A dashboard of form fills alone cannot tell you whether acquisition is profitable.
04 / From the work

The CRM is rarely the only system.
The Room Alert by AVTECH case involved reconciling Google Ads, HubSpot, ecommerce and legacy CRM/offline data to make attributed revenue visible. The case reports $1.6M+ in attributed digital revenue over six months.
That is an outcome from the broader work, not a claim that a CRM connector alone generated the revenue. The practical lesson: reconcile the systems where sales actually happen.
Read the case studies05 / Common questions
Before you connect.
- What does connecting a CRM to Google Ads actually mean?
- There are two directions: deliver ad-generated leads into your CRM, and send downstream outcomes back to Google Ads. For B2B measurement, the second direction connects ad interactions to qualified leads and customers. A contact sync alone does not close that loop.
- Can I connect HubSpot or Salesforce to Google Ads?
- Google documents Data Manager connections for HubSpot and Salesforce. The right setup depends on your CRM fields, supported events and account configuration. Confirm what the connector imports before assuming a connected account is reporting qualified pipeline.
- Do I need a GCLID for every imported lead?
- Google recommends importing GCLIDs when available. Enhanced conversions for leads can also use hashed first-party data collected through the appropriate tagging setup. If a tag is not collecting user-provided data, Google's guidance requires GCLIDs. Matching is not guaranteed for every record.
- Will Google Ads revenue match my CRM revenue exactly?
- Not necessarily. Attribution, reporting dates, match coverage and conversion settings can differ. Reconcile a defined cohort and consistent values rather than forcing the totals to agree. Your CRM is the sales record; Google Ads is the advertising measurement and optimization view.
- Should qualified leads become the bidding goal immediately?
- No. First validate imports, duplicates, stage definitions and match coverage. Google's implementation checklist recommends keeping a new action secondary for the first two to three weeks. A useful bidding signal also needs enough reliable volume for your sales cycle.
Sources and implementation notes
Google's official documentation, checked October 20, 2026. Connector support and API requirements can change; confirm them before implementation. The recommendations above are a planning framework, not an account-specific configuration.
- Google Ads Help: Enhanced conversions for leads
- Google Ads Help: Matching signals and GCLID guidance
- Google Ads Help: Implementation checklist and conversion goals
- Google Ads Help: HubSpot connector
- Google Ads Help: Salesforce and Data Manager
- Google Ads Help: Current API migration guidance
- Google Ads Help: Capture and import Google Click IDs (GCLIDs)
- Google Ads Help: Offline conversion import guidelines
- Google Ads Help: Data Manager and enhanced conversions for leads