Today you changed the bid, tomorrow the ad, the day after that the budget. It feels like you are looking after your advertising: something is off, so it needs tweaking. But there is a problem business owners notice late: a week later enquiries went up or down, and there is no way to say what caused it. The bid, the new ad, the budget, or just the season?
First, the limits of this text. Specialists argue about whether every change resets a campaign’s learning, and below we show all sides with sources. We re-read Google’s help pages on 20 October 2026. The main argument of this article, however, does not depend on that dispute: even if the campaign forgets nothing, you lose the ability to understand what worked.
What campaign “learning” means in plain words
Many Google Ads campaigns run on automated bidding: the system decides how much to pay per click, based on data about who left an enquiry before. It needs fresh examples, and while it collects them, results can fluctuate. That time is called the learning period. We wrote about what to expect week by week in how long to wait for ad results.
How long to wait after changes: what different sources say
There is no single answer, and we will not pretend there is.
What practitioners say. In our experience and in the materials of industry teams, after a budget change of more than 20%, a change of strategy or of campaign structure, it is worth waiting 7–14 days before drawing conclusions. Some advise not touching bids and budget at all during that time; others reduce it to “step of 10–20%, then a pause of a few days”.
What Google’s help page says. In the help page on the target return on ad spend strategy, the timings are these. Before switching to value-based bidding you should wait 4 weeks or 1–2 conversion cycles (a conversion cycle is the time from click to enquiry). After you change the target, the help page says the system reacts immediately but needs some time to reach the new target, and advises waiting 1–2 conversion cycles. That page gives no timings for changing the budget or the strategy type.
What a Google manager said. On 15 October 2026 a Google Ads product manager said on the Google Ads Decoded podcast that when you change the target or budget of a stable, already running campaign, no separate learning period starts, because the system scales bids proportionally using data it already has. According to him, this does not apply to structural changes (new ad assets, new targeting, a new strategy) or brand-new campaigns. He did not say how large a change at one time is acceptable, and Google, as the outlet that reported the conversation notes, has not published data to support the claim.
The conclusion. For big changes (a new strategy, a new structure, a sharp budget jump) the advice agrees: you need a pause. For a small edit on a stable campaign, a “learning reset” is not proven. So we do not claim that every little thing harms the campaign. We claim something else.
The main problem: you will not know what worked
Imagine that in three days you changed the bid, rewrote an ad headline and raised the budget. A week later there are more enquiries. What worked? Maybe the bid, maybe the headline, or maybe it coincided with the start of the month when people have money. You do not know. And, more importantly, you cannot repeat the success on another campaign or roll back only what did the damage.
It is the same with worse results: enquiries fell, and working out whose “fault” it is has become impossible. In that case a contractor usually says “it’s the season” or “the system is learning”, and there is nothing to check those words against.
So the rules are simple:
- One change, one test. Change one thing and watch the result until you change the next.
- After big changes, wait at least a week or two. This is our practice, not a Google requirement. If enquiries come rarely, wait longer: you cannot draw conclusions from three enquiries.
- Write down what you changed and when. A template is below.
- Test hypotheses one at a time. “First we test the bid, then the ad”, not “we change everything and see”.
The change log: a template
An ordinary table in your notes or in Google Sheets is enough. Four columns:
| Date | What we changed | Hypothesis | When we check the result |
|---|---|---|---|
| 12.11 | New headline in the main ad | A number in the headline gets more clicks | 26.11 |
| 26.11 | Budget raised by 15% | The campaign is limited by budget, so there will be more enquiries | 10.12 |
| 10.12 | Added negative keywords from the search terms report | Fewer irrelevant clicks | 24.12 |
This is an illustrative example, not real data. Note the last column: you set the review date the moment you make the change. That way you will not be tempted to look at the result after a day and keep “tweaking”.
Doesn’t Google keep a history itself? It does, and it is useful. In Google Ads open the Campaigns menu and choose Change history. It shows who changed what and when: for example, previous budget amounts and the dates a campaign was paused, as well as the name of the user who made the change. The history is kept for 2 years. But there are two limitations. First, according to Google’s help page, not all account-level changes and not all changes made by Google representatives during consultations may appear in the list. Second, it does not hold your hypothesis or the date you plan to look at the result. So your own four-column log does not replace Google’s history; it complements it.
A useful habit: look at this history when a contractor reports “we optimised the campaign”. You will see how many changes there were and whether they match the story.
When you should touch a campaign right away
This is not a call to “touch nothing”. There are situations where you should not wait:
- the money in the account has run out or the campaign was rejected;
- money is being spent but there are zero enquiries and you suspect the form or enquiry tracking is broken;
- ads are shown for clearly wrong searches (negative keywords fix that, waiting does not).
In these cases you are fixing a breakage, not testing a hypothesis. But even here, write in the log what you fixed and when: you will need that date when you analyse the result.
What to check yourself in 10 minutes
- Open Campaigns → Change history for the last 30 days. Count how many changes were made in each campaign.
- If one campaign has more than five or six changes of different types in a month, you will hardly be able to say which of them had an effect.
- Start a log using the template above and enter every future change there together with its review date.
- If a contractor runs the account, ask to see their notes. Not a demand to “explain every change”, just an ordinary question: what did you change last month and what did it give?
For a similar trap with budgets, read we doubled the budget but got only 20% more customers.
Summary
A campaign may withstand frequent edits, or it may not: here practitioners, Google’s help page and Google itself disagree. But a business owner gains nothing from that dispute while they do not understand why enquiries changed. Keep four things in mind:
- one change, one test;
- after big changes, wait at least a week or two;
- write down the date, the change, the hypothesis and the review date;
- test hypotheses one at a time.
You can read how we run campaigns in the search advertising and Performance Max sections.
Want a free audit of your Google Ads account? Send us a request — we will look at your account and its change history and tell you what is worth leaving alone and what to fix.