How to Send Sales Pipeline Stages as Conversions to Meta and Google

Your campaigns bring in leads, but only some of them ever become customers. The ad platform has no idea which ones: it sees the form and looks for more people who fill out forms. This guide shows how to send stages from your sales pipeline back to Meta and Google Ads as conversions, so their algorithms optimize for leads that actually buy. What a pipeline stage is and which stages are typical is covered in the glossary entry. This guide is about practice: which stages, which events, which deadlines.

Why optimizing for raw leads fails

A lead event only says one thing: someone submitted a form. It doesn't say whether the phone number is real, whether the person fits your ideal customer profile or whether they'll ever accept an offer. Optimize for that signal and you get exactly what you don't want:

• The algorithm finds form fillers: It moves budget to where leads are cheapest, not to where customers come from.
• Cheap campaigns look like winners: A low cost per lead hides a falling close rate.
• Sales notices first: More leads, fewer meetings, rising cost per acquisition.

The fix is in your CRM. It knows which lead was qualified, which one had a meeting and which one bought. Sending that information to the platform as offline conversions is the most effective lever for better lead quality. How to turn it into real ROAS is explained in the guide on how to measure ROAS in lead generation.

Which pipeline stages to send

Not every stage is a good signal. For the ad platform, three things matter:

1. Closeness to revenue: The later the stage, the more closely it tracks revenue.
2. Volume: The platform needs enough events to find patterns. A handful of deals a month gives it little to learn from.
3. Speed: The stage has to be reached soon enough after the click to arrive within the upload deadlines, and so optimization doesn't lag weeks behind.

For most lead businesses that leaves a lean set of three to four stages:

Stages like "proposal sent" or "negotiation" matter for sales forecasting, but for the ad platform they are usually too close to "won" to add anything. Leave them out.

A clear definition is essential. If "qualified" means something different to everyone in sales, the algorithm learns from noise. Give every stage you send a precise entry criterion, for example through lead scoring or a discovery call checklist, and make sure your team updates the status consistently.

Mapping stages to platform events

Meta

For leads from your website, send pipeline stages through the Conversion API as separate events, matched via hashed email, phone number and, if available, the fbc click ID. Send the lead as the standard Lead event and the won deal as Purchase or a custom event with value and currency. For intermediate stages like SQL or meeting, use a custom event. In Events Manager you can build custom conversions from events, and according to Meta's documentation you can optimize ad delivery for them (Meta for Developers).

For leads from Instant Forms, Meta has a separate performance goal, Conversion Leads, which expects CRM stages in a specific format. What Meta requires for it is covered in the post on low-quality leads from Meta Ads.

Google Ads

In Google Ads, create one conversion action of type import per stage, for example "Qualified lead" and "Deal won". Matching works through the gclid (or gbraid and wbraid) or through hashed customer data. Google calls the latter enhanced conversions for leads. Google recommends always sending the gclid when you have it. Since June 2026, new upload integrations go through the Data Manager API.

Values per stage

Value-based bidding needs values. For a won deal, that's the actual deal value. For earlier stages you can use an expected value derived from your own data:

Stage value = average deal value × close rate from this stage

Example: a deal is worth €6,000 on average and 25% of SQLs become customers. An SQL is then worth €1,500. Recalculate the rates regularly, otherwise the platform bids on outdated assumptions. More on the calculation under lead value.

Deadlines: when a stage arrives too late

Pipeline stages happen days or weeks after the click. Each platform only accepts them within a certain window:

In practice: send each stage the moment it changes in your CRM, with the actual time of the status change as the timestamp. A monthly CSV upload loses almost everything on Meta. And with very long sales cycles, the won deal may reach Google only after the deadline. Then an earlier stage as the optimization goal isn't just better, it's the only option. For a comparison of all deadlines and formats, including TikTok and LinkedIn, see conversion APIs compared.

Avoiding double counting

With several stages per lead, the risk of counting an event twice goes up. Two rules help:

• Event deduplication: If you send an event through both the pixel and the server, event_name and event_id must match on Meta. Meta then discards duplicates sent to the same pixel within 48 hours (Meta for Developers). Server-only stages need a stable event_id too, for example built from deal ID and stage, so a re-sync doesn't report anything twice.
• Each stage once per deal: If a deal moves back and forward again, don't send the stage as a conversion a second time.

Primary and secondary conversion actions

In Google Ads, you decide per conversion action whether it feeds into bidding. Primary actions appear in the "Conversions" column and are used for bidding, secondary ones only show up in "All conversions" for observation. The exception: actions in a custom goal are used for bidding regardless (Google Ads Help).

For pipeline stages, this means: set exactly the stage you want to optimize for as primary and all others as secondary. If lead, SQL and deal are all primary, one customer counts three times, and bidding still mostly optimizes for the most frequent event, the lead. On Meta, you choose the optimization event at the ad set level.

Rollout in five steps

1. Define the stages

   Pick three to four stages with clear entry criteria and check in your CRM data how many leads reach them per month and after how many days.

2. Capture click IDs and contact data

   Every lead needs its gclid, fbc/fbp and email or phone, including leads from iframes and funnel tools. Without them, the platform can't match the stage to a click.

3. Create and map events

   Create one event or conversion action per stage in Meta and Google. Set up new stages as secondary or measurement-only first.

4. Check the data

   Let the stages run for two to four weeks. Compare the platform numbers with your CRM and review the diagnostics reports.

5. Switch the optimization goal

   Only once the data checks out, switch optimization to the new stage. Judge the result by cost per acquisition, not cost per lead.

Checklist

• Three to four stages with clear criteria, maintained in the CRM
• Optimization goal has enough volume and is reached quickly
• Every stage is sent right when the status changes
• Deadlines met: 7 days for Meta, 90 or 63 days for Google
• Hashed contact data and click IDs with every event
• Stable event_id per deal and stage
• Only the optimization goal set as a primary conversion action in Google Ads
• Stage values based on your own close rates, updated regularly

How LeadMetrics helps

LeadMetrics connects campaign, lead and deal and sends the stages to the ad platforms:

• Deals from your CRM: Through the HubSpot integration, LeadMetrics picks up deal stages over time, and Close is connected natively as well. Other systems connect via webhooks, Zapier or Make.
• Stages as events: The Conversion API sends lead, purchase, deal won and deal lost events to Meta, Google Ads, TikTok and LinkedIn. You decide per project which LeadMetrics event becomes which platform event. Intermediate stages such as SQL or meeting can be captured as a custom metric, for example through a HubSpot workflow or webhook, and sent as a Conversion API event too.
• Matching and deduplication: Email, phone and names are sent SHA-256 hashed, along with click IDs such as gclid and fbc/fbp, value and currency, and an event_id. For each lead, you see a log of the events sent.
• Attributed or all data: You choose whether to send only events from attributed leads or all of them, the latter for Meta and TikTok.

LeadMetrics doesn't support Meta's Conversion Leads performance goal for Instant Forms, because Meta expects a dedicated CRM format for it. For leads from your website and funnel tools, though, you get every stage to all four platforms through one connection, and the dashboard shows per ad whether the switch actually lowers your cost per acquisition.