Subscription lifecycle events: trial_started to renewed to churned

SaaS teams often mix pipeline events and conversion events into one noisy stream. Pipeline events describe commercial momentum: a visitor requested a demo, a lead became sales qualified, a deal moved to proposal, or procurement started. Those signals are useful because they arrive earlier and in higher volume, but they do not yet mean cash is committed. Conversion events, by contrast, should describe monetization truth such as a trial becoming paid, a subscription starting, or a contract renewing.

The distinction matters because ad platforms optimize on what you feed them. If Meta or Google receives every internal stage change as if it were revenue, bidding shifts toward leads that look busy in the CRM rather than customers who actually pay and retain. A good event model keeps pipeline progression available for scoring and reporting, while reserving monetized subscription events for value-based optimization.

TrackLayer's approach is to bridge the two layers instead of collapsing them. Pipeline changes can enrich the customer profile with plan fit, segment, rep ownership, and opportunity context. Billing or product systems then emit the canonical lifecycle events that confirm economic reality. The result is cleaner attribution, more stable bidding, and a subscription funnel that still makes sense six months later during churn and renewal analysis.

Exact event names differ by network, so the practical job is not strict one-to-one parity. The job is to map each lifecycle event into the closest optimization primitive while preserving the original canonical name in metadata. That lets reporting stay TrackLayer-native even when the ad network only supports a smaller set of standard conversion labels.

Count-based optimization is usually wrong for subscription businesses. A free trial from a micro account and an annual enterprise expansion should not train the model with equal force. The fix is to pass value with each lifecycle event and choose the field that matches the business objective. Self-serve SaaS often uses MRR. Sales-assisted SaaS may use contract_value on subscription_started and expected annual recurring revenue on trial_converted.

In practice, TrackLayer sends the canonical event plus financial context in `custom_data`, for example `value`, `currency`, `mrr`, `arr`, `contract_value`, `ltv_model`, and `mrr_delta`. Meta and TikTok mainly need the optimization value and currency. Google can use the primary conversion value for Smart Bidding while the extra fields remain available for offline analysis and warehouse joins. Keep the primary value stable per event type so models are not trained on inconsistent economics.

{
  "event_name": "subscription_started",
  "custom_data": {
    "value": 249,
    "currency": "USD",
    "mrr": 249,
    "contract_value": 2988,
    "ltv_model": 6220
  }
}

Billing systems are the safest place to derive subscription lifecycle truth because they see invoice success, payment failure, pause, resumption, upgrade, downgrade, and cancellation after the platform has committed the state change. That makes webhook ingestion more reliable than page-based tracking for any recurring revenue workflow.

Cancellation should not behave like a silent data point. If a customer churns shortly after a paid conversion, many teams send a balancing server-side cancellation signal so the media platform is not left believing the original purchase was fully valuable. A common pattern with Meta is to emit a cancellation event tied to the original subscription or order reference and attach a negative value in `custom_data.value` equal to the lost MRR or refunded amount. That makes internal attribution and modeled bidding less optimistic.

Reactivation needs more care. If the same account returns after a brief pause or dunning recovery, treat it as `subscription_reactivated` and preserve the original customer identity. If the return happens after a long churn gap and the commercial motion is effectively a new deal, you may map it to a new `subscription_started` event while still keeping the previous subscription record for retention analysis. The rule should be explicit in your data contract, not improvised inside the ad ops team.

Enterprise and sales-assisted SaaS rarely closes inside a seven-day click window. Trials may convert 30 days later, annual contracts may start 90 days later, and upsells can happen 180 days after the original campaign touch. That is why TrackLayer keeps identity and event timestamps durable enough for longer attribution windows and offline imports.

Meta generally optimizes best on the events it receives quickly, but it still benefits when later subscription starts, renewals, and qualified values are fed back for attribution learning. Google Ads is especially useful when offline conversion imports or enhanced conversions can connect long-lag revenue back to the original click. For enterprise SaaS, the important design choice is to keep the click identifiers, hashed user data, and CRM or billing keys alive long enough to join the eventual revenue event back to acquisition.

The CRM should explain why revenue happened, not invent revenue that billing never confirmed. Stage changes from Salesforce or HubSpot can enrich subscription lifecycle events with owner, segment, forecast category, source campaign, and deal metadata. Once the deal is commercially real, those fields should travel alongside the canonical billing event into ad destinations and the warehouse.