🧠 Is Google’s Aug 17 Bidding Change Behind the CPC Spikes?

Google changed how budget-limited target bidding works on August 17. Advertisers are reporting very different outcomes.

Some saw CPCs jump, budgets burn faster, or conversions fall at similar spend. Plenty of others said nothing changed. In at least two cases, a rough stretch cleared up without another edit.

The update applies to budget-limited campaigns using Target CPA or Target ROAS across Search, Shopping, Performance Max, Demand Gen, and Travel. Target CPC in Demand Gen is also included.

Before August 17, the budget constraint could make a campaign more selective than the target itself. Google's example uses a campaign with a $10 tCPA and a recent actual CPA of $5.

Google now says those campaigns will optimize more consistently toward the target you entered, even while the budget remains constrained.

That $10 target now has more weight. The system has room to bid harder and still finish near the number the advertiser entered. The same logic runs in reverse for tROAS. A campaign beating its target can accept more expensive traffic while still meeting the stated goal.

Google says the auction mechanism did not change. That can be true while prices move because the bidding behavior feeding the auction changed.

A Shopping advertiser reported average CPC moving from $4.41 to more than $12 over two days without changing bids, budgets, targeting, or creative. That works out to about 2.7 times the prior average.

A Performance Max advertiser reported roughly 50% fewer purchases at similar spend.

In the same discussions, one advertiser described CPCs as "looking like a hockey stick." Another said a normal two to three days of spend disappeared in three hours.

The recovery reports matter too. One advertiser saw CPC and CPM rise 18% to 20% while conversions fell 46% across three days. The next day, CPC, CPM, and conversions returned to normal without another change. Another tROAS Shopping campaign missed its target for several days, then hit it for two straight days without an edit.

Other advertisers reported normal accounts. These are self-reported cases without account exports or a clean causal test. Together, they show real account-level volatility and no consistent platform-wide result.

The community reports point to volatility across individual accounts, not one result across Google Ads. That makes the account audit more useful than any headline.

Treat the rollout as a plausible cause only after you confirm the campaign is in scope, line up the timing with Change history, and let conversions catch up.

If a target, budget, feed, or conversion goal moved at the same time, the account cannot separate the variables.

Google also warns that planning forecasts may be inaccurate from August 17 through August 31.

If spend is running outside your business tolerance, act. Otherwise, give the conversions time to arrive before copying a bad day into a new target.

Campaigns without a budget constraint are supposed to behave the same as before. Manual CPC and Target Impression Share are outside this update.

You may also see very little movement if actual performance was already close to the target. Low-volume campaigns need more time because one order or lead can rewrite the whole comparison.

For years, some advertisers used a loose target and a tight budget as two efficiency controls. The budget constraint often helped the campaign beat the target.

That protection is weaker now. The number in the target field needs to match the economics you are willing to accept.

A sharp CPC move deserves attention.