Summary
The correct answer is Predictive. When value is generated at conversion time from propensity and churn signals rather than from a confirmed sale, it is a predicted value fed into Google Ads for value-based bidding. Actual value comes from booked revenue, proxy and manual values are fixed stand-in numbers.
Nitin Batra is a Google Ads Search certified professional who implements offline conversion imports and value-based bidding structures for lead generation accounts on a regular basis.
Last updated: August 22, 2026
A brand that calculates value at the moment of conversion using propensity and churn signals is using predictive conversion values, the value-centric category that scores expected future worth instead of recording a confirmed amount.
- Actual
- Proxy
- Predictive
- Manual
The correct answer is: Predictive
Key Takeaways
- – The correct answer is Predictive, because value is calculated at conversion time from propensity and churn signals rather than from a confirmed transaction amount.
- – Predictive values are pushed into Google Ads with the conversion itself, through offline conversion imports or the API, and the conversion action must be set to use different values for each conversion.
- – Value-centric measurement moves through stages: manual or proxy values first, actual revenue when it is available, and predictive values once you have enough closed-loop data to model outcomes.
- – A predictive value is only as good as the model behind it, so if the scoring drifts or over-weights early signals, Smart Bidding will bid hard on leads that never close.
What is the difference between proxy value and predictive value in Google Ads?
A proxy value is a single fixed number you assign to a conversion action because you do not know the real amount, for example every demo request gets 5000 rupees. A predictive value changes per conversion, calculated from that specific user’s propensity, expected order size and churn risk. Proxy value gets you started with value-based bidding, predictive value makes the bidding much sharper once you have closed-loop CRM data.
How do I send predictive conversion values into Google Ads?
The usual route is offline conversion import using the GCLID, WBRAID or GBRAID captured at lead submission and stored in your CRM. Your scoring engine writes a value against that record, and the upload sends the conversion name, time and value back to Google Ads. You can also pass a value with Enhanced Conversions for Leads if you match on hashed email or phone.
Does Smart Bidding treat a predicted value differently from real revenue?
No, and this is where people get confused. Google Ads takes whatever number arrives with the conversion and optimises to it. It has no way of knowing whether that figure is confirmed revenue or a model output, so the responsibility for accuracy sits entirely with you.
How much conversion data do I need before predictive values are worth using?
You need enough closed deals to know what a lead is actually worth by segment, which usually means several months of CRM data joined back to click IDs. Before that, start with proxy values based on average deal size and close rate by lead type. Move to predictive scoring once the model beats your flat proxy on backtested revenue.