Why did my CPA increase? How to find the change behind a CPA spike

Why did my CPA increase?

Cost per acquisition is cost divided by conversions, so it only goes up for two reasons: you paid more for the same conversions, or you got fewer conversions for the same spend. Every cause you investigate pushes one of those two levers.

Google's own troubleshooting page for performance fluctuations in Search campaigns lists the usual suspects: recent changes to settings (bid strategy, bid adjustments, bids, budget, keywords, audiences, demographics, ad scheduling), conversion tracking and conversion delay, bid targets set without regard to historical data, limited budgets, ad quality, overlapping targeting, policy review status, billing issues, auction dynamics and lost impression share. Group them like this before you start digging:

The fastest diagnosis is not reading every report. It is putting all of these on one timeline and checking which one lines up with the day the CPA moved.

Is the CPA increase real, or is it conversion delay or tracking?

Check this first: a large share of "CPA spikes" in the last few days are conversions that haven't been reported yet, or conversions that stopped being recorded. Neither is a performance problem.

Google Ads reports its main conversion columns by the time of the click, not the time of the conversion. In Google's words in Understand your conversion tracking data, if an ad was clicked last week and converted this week, both the click and the conversion are reported to last week. So the most recent days carry their full cost but only part of their conversions, and CPA looks inflated until the lag catches up.

1. Segment by Conversions > Days to conversion (Google explains the segment here). It splits conversions into up to 19 rows by time to convert. If most of your conversions arrive after 3 or more days, don't judge the last week yet.
2. Add the "Conversions (by conv. time)" column next to the standard one. If the spike disappears there, you are looking at lag, not lost efficiency.
3. Check that each primary conversion action is still recording. A form that changed its thank-you URL, a tag removed in a site release or a consent banner update can cut reported conversions overnight while real leads keep coming in.
4. Compare with your CRM or backend for the same days. Google notes a 24–48 hour processing delay before comparing Google Ads with other platforms.

If CRM sales are flat and only Google Ads conversions dropped, fix measurement before you touch bids. The guide on explaining a conversion drop to a client walks through this check.

Which of my own changes can push CPA up?

Most CPA spikes trace back to an edit made in the 1–14 days before the shift: by your team, by a client, or by an automated recommendation. These are the ones to look for.

Bid targets and budgets

Raising a Target CPA tells Smart Bidding it may spend more per conversion, and it will. Lowering it too far can backfire the other way: Google's Target CPA page warns that a target set too low may make you forgo clicks that would have converted, so you get fewer total conversions. Google also says some conversions will cost more than the target and some less, and recommends judging performance over the last 30 days with at least 30 conversions, not over two days.

Bid strategy learning

According to Google's page on the learning period, a bid strategy shows "Learning" when it is new or reactivated, when one of its settings changes, or when campaigns, ad groups or keywords are added or removed. Learning typically takes 1–2 conversion cycles, so a long sales cycle means a long wobble. It doesn't apply to Manual CPC.

Targeting and structure

New broad match keywords, an expanded location, an added audience or a merged campaign bring in traffic that converts at a lower rate. Google also flags overlapping campaigns or ad groups competing in the same auctions as a source of fluctuation.

Automated changes

Auto-applied recommendations can add keywords or change bidding without anyone on the team clicking a button. See how auto-applied recommendations change your account for what to switch off.

Can CPA rise when nobody changed the account?

Yes. Competitors, seasonality and your own website can all raise CPA without a single edit in Google Ads, which is why the account's change history alone never tells the full story.

• Auction pressure. Open Auction insights for the affected campaign and compare impression share, overlap rate and outranking share before and after the spike. A new competitor showing up on the same day as your cost per click jump is a strong lead.
• Lost impression share. A rising "Lost Search IS (ad rank)" means your ads weren't shown because of ad rank: you are being outbid or ad quality fell. If the campaign is limited by budget instead, Google notes it is more likely to fluctuate, and the budget, not the bid, is the constraint.
• Website changes. Google's Target CPA page notes that actual costs depend on factors outside its control, like changes to your website. A slower page, a longer form, a new pricing page or a broken checkout step lowers conversion rate and raises CPA, and none of it appears in Google Ads.
• Demand and calendar. Holidays, sales events and news change who is searching. Compare the same period last year before blaming the account.

How do I isolate the cause of a CPA spike?

Put every change from the 14 days before the spike on one timeline, rule out measurement, then test one hypothesis at a time. Guessing from a single report is how teams roll back the wrong thing.

1. Pin the date. Find the first day CPA moved outside its normal range, using conversion time columns so lag doesn't fool you.
2. Split the lever. Did cost per click rise, did conversion rate fall, or both? That tells you whether to look at auctions and bids or at traffic quality and the site.
3. Collect the changes. Open Change history (Campaigns > Change history, which keeps 2 years and can be filtered by change type), then add what lives elsewhere: site releases, tag and consent edits, CRM or form changes, promotions and price changes.
4. Rule out tracking. Compare Google Ads conversions with CRM or backend numbers for the same days.
5. Test one hypothesis. Pick the change closest to the spike that moves the right lever. Revert it, or confirm it with a segment (device, location, keyword, audience), and wait at least one conversion cycle.

Example: CPA rises from €42 to €61 starting on a Tuesday. Cost per click is flat, conversion rate fell. Change history shows nothing that week, but the timeline shows the web team shipped a new lead form on Monday. Days to conversion is mostly same day, so it isn't lag. Restoring the old form is the first test. For a full method, read marketing root cause analysis.

How do I keep the next CPA spike from becoming a mystery?

Log changes when they happen, not when something breaks. The hardest part of every CPA investigation is the changes that left no trace in the ad platform: the landing page edit, the CRM rule, the tag that moved in a release.

• Record every change with date, owner, channel and reason, including site, tracking and CRM changes.
• Note bid target and budget changes with the old and new value, so a "Learning" status has an obvious explanation.
• Annotate promotions, price changes and seasonal events.
• Review the log in the weekly performance meeting, before anyone proposes a fix.

A shared marketing change log does this job, and the change log template is a quick way to start. Changeline keeps these changes on one to-scale timeline per client project, so you can see in seconds what happened before a spike; you can start a free change log in Changeline while it's in early access. Pair it with Google Ads change history for the in-platform edits.