Measure your conversion lag before you judge a target CPA

Your target CPA looks worse than it is because the last few days of conversions haven't arrived yet. Measure your conversion lag, the gap between a click and the conversion it produces, then leave that many days out of the window you judge the target on. Google is blunt about the method: with a two-day conversion delay, review your historical CPA over a 28-day period and exclude the past two days.

That one subtraction changes most target CPA decisions. A campaign sitting 40% over target on Tuesday morning is often on target by Thursday, once conversions land and get credited back to Tuesday's clicks.

Find your conversion lag in the Days to conversion report

Google Ads already measures this for you. Go to Campaigns, Ad groups, or Search keywords, click the segment icon, choose Conversions, then Days to conversion. It splits your conversion columns into up to 19 rows, one per delay bucket, so you can see what share of conversions closed same day, on day two, on day seven, and later.

Two things make those numbers usable. End the date range at least 30 days ago, further back if you run a longer conversion window, so every bucket is full. And read the shape of the distribution instead of an average. An account where 70% of conversions close inside 24 hours and the rest trickle in over three weeks behaves nothing like one where they arrive evenly across ten days, even when both average five.

On Smart Bidding campaigns the bid strategy report carries the same signal. Hover the Conversions and Cost/conv. figures and it tells you whether more conversions are still expected inside the dates you're looking at.

Subtract the lag from the window you judge on

Say 80% of your conversions land within nine days. A Monday review of the last seven days is then a review of roughly a fifth of the conversions those clicks will report. The cost per conversion you read is inflated by arithmetic, and every decision you take from it inherits the error.

Move the window. Thirty days, ending nine days ago. Same campaign, same target, and the cost per conversion in front of you is close to the one you will still read a month from now. Keep the recent days on screen for spend and click volume, which are complete the day they happen. Judge efficiency on the older, settled window.

A long conversion cycle makes a slow feedback loop

Google says Smart Bidding starts working toward a new target within minutes, then asks you to wait one to two conversion cycles before you assess the result. A conversion cycle is the typical time a click takes to convert. Calibrating to a new goal can take around 50 conversion events or three conversion cycles.

This is how the money gets lost. You raise the target on Monday. Thursday looks bad, because Thursday always looks bad, so you raise it again. Google warns against more than one target change inside a single conversion cycle for that reason: you cannot read the first change once the second one is in. Each change also restarts the clock, the same trap as moving the target too often right after you switch to target CPA bidding.

When your sales cycle outruns the conversion window

Google Ads counts clicks for 30 days by default and can be set as far out as 90. A B2B deal that takes four months to close falls outside both. The clicks that become your largest contracts convert in month five, after the window shut, so the bidder never learns they were the valuable ones and never goes looking for more of them.

The move here is to bid on an earlier action that predicts revenue: a booked demo, a trial started on a work email, a lead your sales team marks qualified in the CRM. Test the prediction before you commit to it. Pull 90 days of that action alongside the deals it produced and check whether more of the action reliably means more closed revenue. If it does not, the target CPA you build on it will be precise and wrong. This is also why the target has to come from what a customer is worth to you rather than from the cost of whatever action is easiest to count.

Set the review cadence from the lag, not the calendar

A nine-day lag turns a Monday-morning weekly review into a reading of noise. Put the cadence on the cycle: one conversion cycle for a health check, two before you touch the target. Write the lag figure next to the target where the team can see both, so nobody reacts to a week that had not finished. Update it quarterly, or whenever your price or offer changes enough to move how long people take to decide.

Conversion lag sets the speed limit on every paid decision you make. Once the number is written down, the arguments about a bad week that was still filling in stop happening.

If you want help turning this into a measurement and bidding routine you run every week, that is the kind of work we do.