Google Ads Value-Based Bidding: Beyond Cost Per Lead

Cost-per-lead bidding treats every conversion as equal, which is fine until some leads close and most do not, or close deals worth wildly different amounts. The decision this post settles is when to move a campaign from counting leads to valuing them, how to assign a defensible value to a lead before revenue actually lands, and what to check before flipping the bid strategy to Target ROAS.

What problem does cost-per-lead bidding actually have?

A Target CPA strategy optimizes to get you more conversions at a target cost, and it has no way to tell a lead that becomes a signed deal from one that never answers the phone. If your best-performing keyword by cost per lead is quietly producing your worst leads by close rate, cost-per-lead bidding will keep feeding it budget, because from the algorithm's view every lead looks the same.

This matters most where lead quality varies by source, location, or service line. If every lead is genuinely worth about the same to the business, cost-per-lead bidding is not broken and switching adds complexity without benefit.

How do you assign a value to a lead before it closes?

Start from your close rate and average deal size by lead source, then work backward: a lead source that closes at 20 percent with an average deal value of 5,000 has an expected value of roughly 1,000 per lead. Use that expected value, not the eventual deal value, as the conversion value passed back to Google Ads, since the bid strategy needs a number at the moment of conversion, weeks before the outcome is known.

• Segment expected value by lead source or form type if close rates differ meaningfully between them.
• Recalculate the expected value on a fixed schedule as close rates shift, not once at setup.
• Keep a record of how each value was derived so a future audit can defend the number.

Sending values back on a lag works through offline conversion imports, where a CRM stage change updates the value Google Ads already has on file for that click, and the bid strategy learns from the corrected number on its next training cycle rather than waiting for a perfect real-time signal that does not exist for most sales cycles.

Why is a bad value worse than no value at all?

A value-based bid strategy actively steers spend toward whatever looks most valuable in the data it is given. If the values are wrong, in either direction, the algorithm optimizes toward the mistake with more confidence than a human bidder would, because it has no way to know the number is fabricated or stale.

Flat placeholder values, like giving every lead the same value just to satisfy a checkbox for switching to Target ROAS, produce the same outcome as cost-per-lead bidding while looking like an upgrade. That is worse than staying on Target CPA, because it hides the fact that no real value signal exists yet.

Which conversion actions should carry values, and which should not?

Only mark a conversion action as primary and value-bearing once you trust the number behind it. A newsletter signup or a PDF download can stay a secondary, unvalued action for reporting; forcing an arbitrary value onto it just to include it in the value total pollutes the signal the bid strategy actually uses. Google's overview of conversion tracking setup covers how primary and secondary actions are configured before values enter the picture at all. Review how actions are grouped with the conversion action sets guide before deciding what counts toward bidding.

Keep the primary conversion list to actions with a real, defensible connection to revenue. The conversion counting guide covers how count settings interact with this once more than one action is marked primary.

What should you check before switching a campaign to Target ROAS?

Confirm the campaign has enough conversion volume to leave the learning period in a reasonable window, that values are already flowing accurately on the current bid strategy for at least a few weeks so you can see the value data before it starts driving bids, and that the value calculation itself has been reviewed by someone outside the person who built it. Compare Target CPA and Target ROAS directly before assuming ROAS is the natural next step for every account.

Also check the reporting window against the sales cycle. If deals typically take six weeks to close, judging the new strategy after two weeks measures the learning phase, not the outcome; see the conversion lag and reporting windows guide for how far back a lagging conversion can still attribute to the original click.

How do you keep values honest as sales data comes in?

Treat the value model as something that gets corrected, not set once. As actual close rates and deal sizes come in, compare them against the expected values you assigned and adjust the next cycle's numbers rather than leaving a stale model running for a year. Google's own guidance on Target ROAS bidding and on uploading conversion values through the API both assume the values you send are current, not a one-time estimate.

Do not switch to value-based bidding until you can defend the number behind every value you send. An honest cost-per-lead campaign beats a value-based campaign running on invented values.