Optimizing campaigns with conversion values

Sending values is step one. Getting the platforms to actually optimize for them — and improving the model over time — is where the ROAS gains come from. Here's how to run it like a perf manager.

Switch to value-based bidding strategies

• Google Ads — use Maximize Conversion Value (add a target ROAS once you have volume).
• Meta — use value optimization on the dataset receiving your events.
• LinkedIn — feed the offline Conversions API and optimize toward the value-weighted action.

Passing values to a “maximize conversions” (count) strategy does nothing — the strategy has to be value-aware.

Give the algorithm a spread to learn from

A model where every lead is worth roughly the same gives the platform no signal. Aim for a distribution with real differentiation. A quick proxy: the coefficient of variation (CV) of your values — higher is better. If most leads cluster around one number, add a field that separates them (deal size, ICP fit).

Calibrate against reality

Predicted value is a hypothesis. The truth is closed-won revenue. PipeValue pulls outcomes back from your CRM and compares predicted vs. realized value per segment. If a tier is consistently over- or under-valued, the model surfaces the correction so your bids stay anchored to actual pipeline.

Iterate deliberately

Treat the model like an experiment. Version every change, compare a draft against the active version on real leads, and roll back in one click if a change hurts. Give each change a learning window before judging it — platforms need days, not hours, to adapt.