Should I feed booked job revenue back into Google Ads, or just count leads?
Revenue is the truer signal but the sparser and later one. Leads are immediate and plentiful but treat a drain snake and a system replacement as equals. The usual right answer is a middle path: send the lead event promptly with an estimated value based on that lead type's historical booking rate and average ticket, then reconcile to real revenue in your own reporting.
The tradeoff stated plainly
Feeding actual booked revenue gives the bid strategy the objective you truly care about. It also gives it fewer data points, delivered later, with a very skewed distribution — a handful of large jobs dominate the total. Skewed, sparse, delayed signals are the hardest thing to optimize against.
Feeding lead counts gives the strategy dense, immediate data about something you only partly care about. It will happily buy you a pile of cheap, low-value inquiries because every one of them scores identically.
The proxy value approach
The compromise most service businesses land on is to attach an estimated value to the lead at the moment it happens. Compute it from your own history: for each lead type, booking rate multiplied by average completed ticket. A water heater inquiry and a filter change inquiry then carry different weights, immediately, without waiting for the job to close.
This requires knowing the lead type at capture time, which is exactly what call classification and form routing produce. Deriving those rates from the field service record rather than guessing them is the part that makes it credible, and it is a standard piece of revenue intelligence work.
When to go all the way to real revenue
- You have enough closed jobs per campaign per week for the strategy to see a pattern rather than noise.
- Your lag is short and your import is daily. Weekly imports of revenue give the bidder a signal that is stale by the time it arrives.
- Your ticket distribution is not dominated by rare outliers. If one job a quarter is worth more than a month of ordinary work, consider capping the reported value so a single anomaly does not drag the whole strategy.
The failure mode nobody warns you about
Value-based bidding will chase whatever your value field says, including its errors. If the integration sends invoice totals that include parts pass-through, or credits a whole membership contract to one click, the strategy will confidently pursue more of that. It has no way to know the number is wrong.
So the value pipeline needs the same care as an accounting feed: a defined source, a defined exclusion list, and a periodic check that the total pushed to the platform reconciles to the total in the operational system. That discipline is ordinary integration engineering, and skipping it is how value-based bidding earns its bad reputation.
Topics: conversion value · offline conversions · smart bidding · revenue
Have a version of this question about your own business?
The useful answer usually depends on which systems you run and how they're connected. That's a conversation, not a blog post.