"Don't touch it, you'll reset the learning period" is the most repeated piece of advice in paid search, and most of the people repeating it could not tell you what actually triggers it. So they freeze the account, avoid edits that cost nothing, and make the ones that genuinely hurt.
Google documents this. The list is short.
What Google says triggers learning
A bid strategy shows a Learning status in four documented cases:
- New strategy. The bid strategy was recently created or reactivated.
- Setting change. A setting for the bid strategy was changed.
- Composition change. Campaigns, ad groups, or keywords were added to or removed from the bid strategy.
- Ad group target change. In some cases this triggers learning on Shopping campaigns.
Read that list again and notice what is missing. Budget is not on it.
So why does raising budget still hurt?
It often does, but not through learning. Our read, from the accounts we run: at its old spend level a campaign could afford to be selective, taking the cheapest conversions available to it. Give it more money and it still has to spend it, so it moves down into the auctions and queries it was previously skipping. Those cost more and convert worse. The system did not get confused. You asked it to buy the next-best inventory. We cover the safe way to do that in how to scale a Google Ads budget.
What the learning period actually costs
Google says a bid strategy can take up to three weeks, or one to two conversion cycles, to calibrate to a new objective, and that it can be faster when there is more conversion data. Two things stretch it: low conversion volume and a long sales cycle. That is why B2B accounts feel learning far more than high-volume ecommerce ones. If you get twenty conversions a month on a ninety-day cycle, "one to two conversion cycles" is most of a quarter.
This is also the argument for Smart Bidding patience in general. A target CPA that looks broken in week one is often a strategy that has not finished calibrating.
The practical rule
Batch your changes. Google's own Smart Bidding guidance is to make significant changes at once, precisely to reduce total learning time. The worst pattern is the anxious one: a target tweak Monday, a new ad group Wednesday, a bid strategy switch Friday. Each of those restarts the clock, and you never get a clean read on any of them.
Make the edits together, then leave it alone long enough to judge. And stop avoiding the changes that were never on the list.
The takeaways
- Google's documented learning triggers: new or reactivated strategy, bid strategy setting change, composition change, and Shopping ad group target changes.
- Budget changes are not on that list. If raising budget hurts, it is through what the campaign can buy, not through learning.
- Calibration can take up to 3 weeks or 1-2 conversion cycles; low volume and long sales cycles stretch it.
- Batch significant changes together rather than spreading them across the week.