Lesson 3.3: The Value Ladder (When Revenue Isn't Ready Yet)
By the end of this lesson, you'll know how to assign relative values to pipeline stages without needing perfect LTV data, and how to choose a bridge metric that keeps Smart Bidding from starving while you build toward better signals.
Maybe you don't know exactly what each lead type is worth in dollars. Maybe your CRM isn't connected to billing yet. Maybe different deal sizes make a simple dollar value feel misleading.
Good news: you don't need perfect dollar values to stop treating all leads as equal. You need relative values: a hypothesis, explicitly labeled as such, about how outcomes differ in importance.
Here's what that looks like:
| Stage | Hypothesis value (relative points) | Why |
|---|---|---|
| Raw submit | 1 | Baseline; mostly noise |
| Booked call (showed) | 20 | Higher intent signal |
| Qualified | 80 | Sales agrees this one matters |
| Won | 500 | Money in the door |
Smart Bidding using Maximize Conversion Value or tROAS can work with these point values to prefer better outcomes, even before your finance team is ready to sign off on precise LTV math. You're not claiming perfect knowledge. You're encoding honest directional judgment.
Choosing a bridge metric
Pick one stage that (a) correlates with money, (b) happens often enough not to starve your campaigns of learning signals, and (c) you can define without triggering a week-long debate between teams. Common bridges: qualified lead, booked appointment, opportunity created. You know your business — pick the one that's both meaningful and measurable.
On "starving the account"
This is a real concern, and one of the most common objections when you propose moving away from counting everything. The ladder exists precisely so you don't flip overnight from "count everything" to "count three events per month," which would genuinely break bidding. The sequence matters: volume to keep learning, truth to set direction. Stage your way there.
Value rules
Google allows you to adjust conversion values by geography, audience, or custom dimensions. If some segments systematically outperform others (certain industries, certain markets, certain device types), you can encode that directionally. Don't wait for perfect data. "Directionally right and reviewed quarterly" beats "technically perfect and never shipped."