Why did your CPA go up? Once you rule out noise and conversion lag, a CPA increase has only four possible sources: media got more expensive (higher CPM), fewer people clicked (lower CTR), fewer clicks converted (lower conversion rate), or budget moved to campaigns with a higher CPA. Since CPA = CPC ÷ conversion rate, you can measure which driver explains most of the rise and which campaign caused it.
Key takeaways
- CPA = CPC ÷ conversion rate and CPC = CPM ÷ (1,000 × CTR), so every CPA change traces back to CPM, CTR, conversion rate or campaign mix.
- First confirm the rise is real: enough conversions, outside the campaign’s normal range, and not just recent days awaiting late conversions.
- Log shares show which driver matters most: in the worked example, a 20% conversion-rate drop explains about 70% of a 37.5% CPA increase.
- Account CPA can rise while every campaign holds steady, when budget moves to campaigns with a structurally higher CPA.
- A higher CPA is not automatically bad: if value per conversion rose faster, ROAS improved.
What is CPA made of?
CPA is spend divided by conversions, and it breaks down into three rates you can check one at a time: CPM, CTR and conversion rate. So every increase comes from one of four sources.
CPA = spend ÷ conversions
CPA = CPC ÷ conversion rate
CPC = CPM ÷ (1,000 × CTR)
CPA = CPM ÷ (1,000 × CTR × conversion rate)
- Media got more expensive: a higher CPM makes every click cost more.
- People click less: a lower CTR means more impressions per click, so CPC rises even at the same CPM.
- Clicks convert less: a lower conversion rate means more clicks per conversion.
- The mix changed: spend moved toward campaigns whose CPA was always higher.
Step 1: Is the CPA increase real?
It is real if it rests on enough conversions, sits outside the campaign’s normal range and isn’t just conversions that haven’t arrived yet. Check all three before looking for a cause.
Are there enough conversions?
A CPA built on a handful of conversions swings on chance alone. Example: a campaign spends $900 and gets 9 conversions, a $100 CPA. With 6 conversions for the same spend, CPA is $150, up 50%, and three conversions more or less is ordinary variation at that size. When volume is low, widen the window to 7 or 14 days.
Is it outside the campaign’s normal range?
Compare with the campaign’s own history, not with yesterday. Take four to eight weeks of daily CPA and find its typical level and spread with the median and the median absolute deviation (MAD) rather than the mean and standard deviation, so one very bad day doesn’t distort “normal”. Only a CPA well outside that spread is a signal.
Weekdays matter too. Many accounts convert differently on weekends, so compare a Monday with previous Mondays, or compare full weeks.
Are recent days still missing conversions?
Conversions often arrive days after the click. In its standard Conversions column, Google Ads credits each conversion to the date of the ad click, so the latest days keep gaining conversions and their CPA looks inflated until they do.
Example: a campaign usually runs at a $50 CPA. Yesterday it spent $500 and shows 7 conversions so far, a CPA of $71.43. If three more conversions from yesterday’s clicks arrive during the week, yesterday ends at 10 conversions and exactly $50. When conversions usually lag, leave the last few days out.
Step 2: Did clicks get more expensive, or did they convert less?
Find out by splitting CPA into CPC and conversion rate. That one split tells you whether to look at the auction or at what happens after the click.
Example: last month a campaign had a CPC of $2.00 and a conversion rate of 4.0%, so CPA was $2.00 ÷ 0.040 = $50.00. This month CPC is $2.20 (+10%) and the conversion rate is 3.2% (−20%). CPA is $2.20 ÷ 0.032 = $68.75, up 37.5%.
Notice that +10% and −20% don’t add up to +37.5%. The changes multiply: 1.10 ÷ 0.80 = 1.375. To say how much each factor contributed, use natural logarithms, which turn the division into a subtraction that adds up exactly.
ln(CPA ratio) = ln(CPC ratio) − ln(conversion-rate ratio)
ln(1.375) = ln(1.10) − ln(0.80)
0.31845 = 0.09531 + 0.22314
CPC share = 0.09531 ÷ 0.31845 = 29.9%
Conversion-rate share = 0.22314 ÷ 0.31845 = 70.1%
The conversion rate is the main driver: it explains about 70% of the increase, CPC about 30%. That sends you to step 4 first and step 3 second.
When one factor improves while CPA still rises, its share comes out negative: it offset part of the increase, and the other factor’s share goes above 100%.
Step 3: If CPC rose, is it the auction or the ads?
A CPC increase comes from a higher CPM, a lower CTR, or both, because CPC = CPM ÷ (1,000 × CTR). A higher CPM means the impressions themselves got pricier; a lower CTR means your ads earn fewer clicks from the same impressions.
Continuing the example, suppose CPM went from $40.00 to $41.80 (+4.5%) and CTR from 2.0% to 1.9% (−5%). CPC moved from $40.00 ÷ (1,000 × 0.020) = $2.00 to $41.80 ÷ (1,000 × 0.019) = $2.20. The same log split on the full chain breaks the 37.5% CPA increase into CPM 13.8%, CTR 16.1% and conversion rate 70.1%.
CPM up, CTR steady: auction pressure
You pay more for the same exposure. Typical reasons: more advertisers competing for the same searches or audiences, seasonal peaks such as Black Friday, a looser bid target, or narrower targeting. In Search campaigns, open Auction insights for new or more aggressive competitors, then the change history for bid and budget edits around the day CPM moved.
CTR down: the ads or the traffic lost relevance
Fewer people click the same impressions. Look for ad fatigue, copy or asset changes, broadened keywords or audiences, new placements and stronger competitor offers. In Search, the Search terms report shows whether you started matching queries that don’t fit what you sell.
Step 4: If the conversion rate fell, what changed after the click?
Something between the click and the conversion changed: the measurement, the landing page, the offer or the kind of traffic. Check them in roughly this order:
- Tracking. A drop that starts on one exact day, especially to near zero, usually means measurement changed, not customers: a tag lost in a site release, a conversion action switched between primary and secondary, a new consent banner. Check with Tag Assistant and compare with backend orders or GA4 key events.
- Landing page. Broken forms, checkout errors, slow mobile pages, products out of stock. Click your own ad on a phone and try to convert.
- Offer and price. A price increase, an ended promotion, higher shipping costs or a competitor’s sale all lower the conversion rate without anything breaking.
- Traffic quality. Broader match types, new audiences, automated expansion or new placements can bring clicks that convert less. Search terms and placement reports show where they come from.
- Attribution settings. A new attribution model or conversion window changes how many conversions get counted, even when nothing real changed. Compare before and after the date of the edit.
Step 5: Which campaign moved the account CPA?
Usually the campaign whose own CPA rose on a large share of spend, but sometimes none got worse. Account CPA is total spend divided by total conversions, not an average of campaign CPAs, so it also rises when spend moves toward campaigns that always had a higher CPA: the mix effect.
Example: an account spends $12,000 a month on two Search campaigns. Neither campaign’s CPA changes; only the budget split does.
| Campaign | CPA | Before: spend → conversions | After: spend → conversions |
|---|---|---|---|
| Search – Brand | $20 | $6,000 → 300 | $3,000 → 150 |
| Search – Generic | $60 | $6,000 → 100 | $9,000 → 150 |
| Account | $30 → $40 | $12,000 → 400 | $12,000 → 300 |
Account CPA rose from $12,000 ÷ 400 = $30 to $12,000 ÷ 300 = $40, up 33.3%, while both campaigns kept exactly the same CPA. Cutting brand budget, pushing money into prospecting or launching a new campaign all produce this pattern.
To separate the two effects, recompute the account CPA with this period’s campaign CPAs and last period’s spend split. Here that gives $30 again, so the whole increase is mix. If it lands near the new CPA instead, the campaigns got worse: run steps 2 to 4 on the one that moved most.
Borealis runs this diagnosis automatically for Google Ads. Every morning it compares each campaign with its own history, splits a CPA change into cost per click and conversion rate, and names the campaign that explains the account’s movement. See how it works in Google Ads monitoring.
Which metric points to which cause?
Each pattern points to a short list of causes and a first place to check.
| What moved | Likely causes | Where to look |
|---|---|---|
| CPM up, CTR steady | More competition, seasonal demand, a looser bid target, narrower targeting | Auction insights, impression share, change history |
| CTR down | Ad fatigue, copy changes, broader keywords or audiences, new placements | CTR by ad, Search terms, placement reports |
| Conversion rate down, CPC steady | Landing page problems, price or offer changes, lower-quality traffic | The landing page on a phone, recent site releases, Search terms |
| Conversions near zero overnight | Broken or changed tracking | Tag Assistant, primary and secondary conversion actions, backend orders |
| Only the last few days look bad | Conversion lag | The same days again a week later |
| Account CPA up, no campaign worse | Mix: spend moved to higher-CPA campaigns | Spend share and CPA by campaign, before vs after |
| CPA up, ROAS up | Fewer but more valuable conversions | Value per conversion and ROAS |
The last row matters: a higher CPA is not always a problem. If value per conversion rose faster than CPA, ROAS improved and the campaign earns more per dollar. Check value per conversion before cutting a campaign whose CPA went up.
Frequently asked questions
Why is my CPA always higher for the last few days?
Because conversions arrive after the click and are usually credited back to the click date. Recent days have their full spend but not yet all their conversions, which keep coming in for days or weeks, so their CPA looks inflated. Leave them out of comparisons, or recheck them a week later.
Can CPA go up when nothing changed in my campaigns?
Yes. Competitors can raise their bids, seasonal demand can push CPMs up, a site release can break a form or a tag, and budget moving between campaigns can raise account CPA while each campaign stays the same. None of these appears in your change history, so split CPA into its drivers instead.
Is it normal for CPA to rise after changing a bid strategy?
Often, for a while. Automated bid strategies recalibrate after significant changes to the strategy, target or budget, and performance is less stable in that period. Judge the result once enough conversions have gone through the new setup. A looser target, meaning a higher target CPA or a lower target ROAS, usually shows up first as higher CPCs.
Should I pause a campaign when its CPA goes up?
Not before you know why. Confirm the rise is real, then check whether value per conversion or ROAS improved, which can make a higher CPA worthwhile. If a broken landing page or tag is the cause, fix it; pausing won’t. If CPA is truly higher with steady tracking and value, lower the budget or tighten the target first.