How value-based bidding works
Most accounts start by counting. Every conversion is one conversion, and the bidding strategy tries to buy as many as it can within a cost target. Value-based bidding replaces the count with a number attached to each conversion, and asks the system to maximise the total of those numbers instead — either freely, or against a return target.
The value can come from three places. An online shop can pass the actual transaction amount, so the bidding follows real revenue. A service business can assign a fixed value to each type of enquiry, reflecting what that enquiry is typically worth once it closes. Or values can be adjusted after the fact, when a lead becomes a customer and the outcome is imported back into the account. Rules can also lift or lower value by location, device or audience where those genuinely differ.
Why value-based bidding matters
Counting treats a small order and a large one as the same event, so the system happily buys whichever is easier to get — and the easy one is almost always the cheap one. An account can improve its cost per conversion month after month while revenue stands still, or falls.
The difference shows most in businesses where enquiries vary widely in worth. A remittance or fintech advertiser signing an occasional high-value account, an education consultancy handling both short courses and full degree placements, or a trekking operator selling day hikes alongside long expeditions all have the same problem: the count is a poor description of the business. Assigning value lets the bidding chase the enquiries that pay, and it is the reason conversion value is worth setting up even when a shop’s automatic revenue tracking is not available.
Where value-based bidding goes wrong
The most common failure is invented values. If numbers are guessed to fill the field, the system optimises confidently towards a fiction, and it will do so faster than a person can notice. Values should come from something real: an average order, a known service price, a documented conversion rate from enquiry to sale.
The second is passing revenue when the business runs on margin. A high-revenue, low-margin product can outbid a profitable one and take the budget with it. The third is thin data — a strategy asked to learn from a small number of valued conversions each week has too little to work with. And the fourth is leaving values frozen while prices, product mix and close rates move on underneath them.
How to act on it
Start by writing down what each conversion type is actually worth, in a spreadsheet, before touching the account. Use margin rather than revenue where the two differ sharply. If enquiries lead to sales offline, plan how those outcomes will come back into the platform, because that loop is what makes the values honest over time.
Get the measurement right first — accurate, deduplicated conversion tracking with one primary action per outcome — then switch a single campaign to a value strategy and leave it alone through its learning period. Move to target ROAS only once maximising value has produced a stable picture of what return is achievable. Review the assigned values on a set schedule, and change them when the business changes rather than when the reporting looks disappointing.
