Google Ads Change History: How to Find the Cause of a Performance Drop

Short answer
What is Google Ads change history and how does it explain a drop?
Google Ads change history is the record of changes made to your account, campaigns, and ad groups over the past two years, including who made each change and when. During a performance drop, you read those changes next to the performance graph to find the edit that most likely caused it.
This guide focuses on one job: reading the record when results fall. When conversions suddenly decline, the first instinct is often to raise the budget or rebuild the campaign. However, that reaction can make things worse. In many cases the cause is a small edit someone made a few days earlier. So checking the record first saves both time and budget.
Think of this guide as an emergency card. First, define the drop. Next, pick the right date range. Then filter the records and find the suspicious change. Finally, test that change on its own. Each step replaces guessing with evidence.
What information does Google Ads change history show?
The Google Ads Help page About change history explains the scope clearly. For example, you can see when a campaign was paused, who added a keyword, and what the previous budget amount was.
In addition, each change appears on a timeline next to performance data such as impressions, clicks, conversions, click-through rate, and cost. That view is the heart of the analysis. Because the graph and the edits sit on the same screen, you see right away whether a break in the line follows a change.
The User column shows who made each change. Changes made in the interface show the person's email address. On the other hand, automated or system changes appear under different names. In short, the record answers three questions: what changed, when it changed, and who changed it.
Where do you find change history in your account?
According to Google, you can see all changes from the change history area within the campaigns section. To review changes for one campaign or ad group, select it first and then open change history. Changes can appear in a side panel, so you keep the main table in view while you work.
Menu names change over time. So look for the relevant area in your account rather than relying on exact button labels.
Access also matters. Users with access to your account can see who made certain changes. For instance, if you gave an agency access, its changes appear under its own email. We explain access setup in our guide on how to give agency access to Google Ads. That way you know from the start who works in the account and with which permissions.
Where should you start when performance drops?
Define the drop before you open the record. Otherwise you get lost in hundreds of entries.
- Name the metric that fell: conversions, clicks, or impressions.
- Find the day the decline started on the graph.
- Check whether the issue affects the whole account or a single campaign.
- Pick the seven to fourteen days before the start as your review window.
- Then open the change history for that window.
For example, if conversions fell while clicks stayed flat, the problem likely sits on the website or in measurement, not in the ads. In contrast, if impressions fell, a budget, bid, or targeting change becomes more likely. So choosing the right metric narrows the search.
Also note the size of the drop. A small dip can be normal daily noise. However, a clear break that lasts several days points to a real cause. If weekdays and weekends naturally differ, compare Monday with the previous Monday, not with Saturday.
How should you choose the date range?
The date range is the most important decision in the analysis. Pick it too narrow and you miss the cause. Pick it too wide and noise takes over.
In Smart Bidding campaigns, the effect of a change may not show immediately. That is because the system enters a learning period while it adapts to the new setting. Therefore look back at least one to two weeks from the start of the drop.
Conversion delay matters as well. For example, if customers buy a few days after clicking, the most recent days are not complete yet. In that case the "drop" may simply be delay.
A practical method helps here. First, set two periods of equal length before and after the drop. Then open change history across both periods together. That way you see the changes and their effect in one window. Also, comparing with the same period last year helps you separate seasonality.
Why do comparison dates matter so much?
On its own, the record answers "what happened." The question "did this change cause the drop" needs a comparison.
Performance tables in Google Ads let you set a comparison date range. For example, you compare two weeks after the drop with two weeks before it. As a result, you see which campaign, ad group, or keyword lost the most.
Next, go back to the biggest loser and open the record for that item only. That narrows the search from the whole account to one campaign. In other words, comparison dates tell you where to look, and the change log tells you what happened there. Using both together prevents decisions based on guesswork.
Which filters narrow the search fastest?
According to Google, you can filter results by type of change, such as budget adjustments or keyword edits. You can also view changes for a specific campaign or ad group.
Using filters in this order makes the work easier:
- First, select the campaign where the drop appears.
- Next, limit change types to budget and bids.
- Then review targeting and location changes.
- After that, check ad and asset changes.
- Finally, look at keyword additions and removals.
This order starts with the edits that affect performance fastest. For instance, if someone cut the budget in half, fewer impressions are almost certain. Finding that record at first glance can solve the problem without any further digging.
That said, watch for one trap. Some changes happen at the account level, not at the campaign level. A shared budget or an account-wide setting may not appear under a campaign filter. So if the campaign view shows nothing, clear the filter and scan the whole account once more.
What does the User column tell you?
The User column is the most overlooked part of the analysis. According to Google, interface changes show the person's email. Automated or Google system changes appear under names like "Google Ads system." Changes made through the API show "Google Ads API" or the name of the tool used.
This split is valuable. Some performance drops come from changes nobody made on purpose. For example, a third-party tool may have changed bids in bulk. Or an auto-applied recommendation may have added new keywords.
Here is a practical tip. Pull every record without a personal email into a separate list. Then compare that list with the date of the drop. As a result, you separate human decisions from automated ones and quickly test whether "we did not change anything" is actually true.
How do you trace auto-applied recommendations?
Google Ads generates recommendations for your account and can apply some of them automatically with your permission. These can touch budgets, bid strategies, keywords, or ad assets.
An auto-applied recommendation appears in the record under a system name, not a person. So the User filter from the previous section works here too. For example, if you see broad match keywords added by the system a few days before the drop, a shift in traffic quality becomes a strong suspect.
We cover how to turn auto-apply on or off and how to choose recommendation types in our guide on turning off auto-applied recommendations. However, keep one point in mind: switching auto-apply off does not undo past effects. Therefore check the affected items separately.
We also explain how broad match affects traffic quality in our article on broad match risks.
Which change types cause drops most often?
The drop causes we see most often follow a few patterns. The table below summarizes the change type, its likely effect, and what to look for in the record.
| Change type | Likely effect | What to look for in change history | First check |
|---|---|---|---|
| Budget cut | Fewer impressions and clicks | Previous and new budget | Limited by budget status |
| Bid strategy switch | Learning period, volatility | Strategy name and date | Learning status |
| Target CPA or ROAS change | Impression share shifts | Old and new target value | Lost impression share |
| Location or language targeting | Reach shrinks or widens | Added or removed locations | Location report |
| Keyword match change | Traffic quality shifts | New keywords and match type | Search terms report |
| Ad or asset change | CTR changes | Paused or added ads | Ad strength and approval status |
| Conversion setting change | Reported conversions shift | Primary goal changes | Conversion status |
Use the table as a checklist. For example, if the drop started in impressions, check the first three rows. In contrast, if impressions held but CTR fell, focus on the ad row. If conversions fell while clicks held, check conversion settings and the website.
Why can conversion setting changes mislead you?
Some drops are not real. Sales stay the same, yet reported conversions fall. The most common cause is a change in conversion settings.
For example, if someone moves a conversion action from primary to secondary, Smart Bidding no longer optimizes for it, and it disappears from the Conversions column. Similarly, a change in the counting setting can turn several conversions from one user into one.
Therefore check conversion settings during a drop, not only campaign settings. However, Google notes that some account-level settings changes may not appear in the record. So if you see a gap on the measurement side, review the tag and GA4 separately. We cover the GA4 side in our article on missing GA4 conversions.
What are the limits of Google Ads change history?
The record is powerful, but it does not show everything. Google lists several limits.
- Records cover only the past two years.
- Some account-level settings changes may not appear.
- Some changes made by Google representatives may not appear.
- Detail can be limited when you change a large number of items at once.
- If history looks incomplete, Google suggests clearing the browser cache or using an incognito window.
The fourth point matters in practice. Google recommends smaller changes for full detail, ideally affecting fewer than 2,000 entities at a time. For instance, if you change bids across the whole account with one bulk edit, it becomes hard to see later how each item changed. So splitting bulk changes into smaller batches makes future analysis easier.
How do you separate external causes?
If you find no record that explains the drop, the cause probably sits outside the account. That is useful information too, because it tells you to stop editing the account and look elsewhere.
The main external causes are seasonality and demand shifts, competitor bidding, site problems, and measurement issues. For example, if auction insights show a new competitor gaining impression share, competition may explain the drop. We explain how to read that report in our guide to Google Ads auction insights.
On the site side, checkout, form, or speed problems reduce conversions while clicks stay the same. So the metric split guides you here as well. To interpret a falling ROAS, see our guide on diagnosing a ROAS drop.
What should a step-by-step review checklist include?
The list below helps you read the record systematically during a drop. Work through it in order.
- Write down the falling metric and the start date.
- Use comparison dates to find where the drop concentrates.
- Open the record for that campaign for the two weeks before the drop.
- Check budget and bid changes.
- Review targeting, location, and language changes.
- Look at keyword and match type changes.
- Check ad and asset changes.
- Separate system and API records in the User column.
- Verify conversion settings separately.
- Note the suspicious change and test it on its own.
Once this routine sticks, reviewing a drop often takes less than an hour. As a result, you make a targeted fix instead of rebuilding the account in a panic. Sharing the list with your team also helps, because everyone follows the same order and no step gets skipped.
How do you test a suspicious change?
Finding a suspicious change does not prove it caused the drop. So the fix has to be controlled.
First, reverse only that change. For example, if someone raised the target ROAS, restore the old value. However, do not raise the budget at the same time. If you move two variables at once, you cannot tell which one helped.
Then wait a few days and watch the same metric. In Smart Bidding campaigns, recovery may take time because of learning. Patience is the most important rule here.
For bigger decisions, an experiment is safer. We explain the approach in our guide to drafts and experiments. That way you compare the old and new settings in the same period. Also avoid launching new campaigns during the test, because budget shifts would blur the result.
What does an example investigation look like?
Let us walk through an example scenario. An ecommerce account notices on Monday morning that conversions fell sharply over the weekend. The team says nobody changed anything.
First, comparison dates show the drop sits in one Search campaign. Next, the team opens that campaign's change history. In the User column, a Friday record under a system name stands out: new keywords were added.
Then the search terms report confirms the suspicion. The new keywords brought clicks from unrelated queries. So the fix is to pause those keywords and review the auto-apply settings.
This is not a real client case but a common pattern. Still, it shows that "nobody changed anything" cannot be confirmed without looking at the record. In addition, reviewing auto-apply settings prevents the same problem next week.
How do location and bid adjustments appear in the record?
Location targeting and location bid adjustments affect impressions directly. For example, if you removed a region from targeting or lowered its bid adjustment, impressions from that region drop.
You find these records under campaign settings. However, the level of detail varies. So when you suspect a location-related drop, also open the location report and compare by region.
We explain how regional bid adjustments work in our related article on Google Ads location bid adjustments by region. Here we stress one point: note the date of every location change. Then you know the timing even if the record shows limited detail. To measure the effect, compare impressions and conversions by region before and after that date.
What should you do when many people work in one account?
An account may have the business owner, an in-house marketer, an agency, and a software tool all working at once. That crowd makes drops harder to explain, because everyone assumes their own change was harmless.
The first rule is clear roles. For example, one person or team owns bids and budgets, and another owns ad copy. Then, when you see a record in the User column, you instantly know which area it came from.
The second rule is a shared communication channel. Anyone planning a big change announces it first. Also, if you manage several accounts from one place, the structure we describe in our guide to Google Ads manager accounts helps you organize access.
In short, the record is just a tool in a busy account. The real order comes from a workflow that defines who can change what.
How does the Smart Bidding learning period mix with a drop?
In Smart Bidding campaigns, some changes force the system to relearn. For example, switching the bid strategy or moving the target sharply can trigger this.
Performance can fluctuate during learning. That fluctuation may look like a lasting drop, but it is often temporary. Still, blaming every dip on learning is a mistake too. So when you see a strategy or target change in the record, note its date first.
Then wait a few more days and watch the trend. If performance recovers, it was learning. If not, the new setting itself may be the problem. We discuss target updates in our article on Target CPA and Target ROAS updates.
Meanwhile, adding new changes during learning extends the process. Therefore do not rush into a second edit when you see a dip. Measure the first change before anything else.
Can ad approval or policy status cause a drop?
Yes. Sometimes the cause is not your edit but a change in your ads' approval status. For example, if an ad or landing page gets flagged in policy review, it may stop serving.
These cases do not always appear as a user change. However, the ad status changes in the interface. So when the record shows nothing useful, check ad and asset statuses too.
A practical approach works well. Open the ads table in the affected campaign and look at the status column. If you find disapproved or limited ads, read the reason. For instance, an unreachable landing page can lead to a disapproval. We explain the fix in our guide on destination not working disapprovals.
Also, a website update can break the page an ad points to. Therefore logging site changes in the same change log speeds up the analysis.
When does expert support make sense?
You do not need to be an expert to read the record. With the steps in this guide, you can find the cause of many drops yourself.
However, an outside view helps in some cases. For example, when several people and tools work in one account, the record grows to thousands of lines. Or the drop comes from a mix of internal and external causes. Separating those takes experience.
In our google ads agency services, Talha Aslan and team log every major change with its date and reason. So when a drop happens, we read the official record and our own notes side by side. We do not promise results, but we can always explain why a decision was made.
The real value of expertise lies in interpretation. For instance, a budget increase and a target change on the same day can hide each other's effects. An experienced reviewer separates them with comparison dates and segments, so the fix targets one point.
Why does keeping your own change log help?
The official record shows what changed, not why. Your own change log fills that gap.
A simple table is enough: date, who, what changed, why, and the expected effect. For example, a note like "raised target ROAS because margins were falling" makes analysis two months later much easier.
In addition, your log keeps decisions older than two years. Since the official Google Ads change history stops at two years, your notebook becomes valuable memory in a long-running account. In short, the two complement each other: one is the system's record, the other is the team's memory. If you keep the log in a shared sheet, everyone has the same context, even when someone is away.
Which habits help prevent drops?
The best analysis is the one you never need. A few habits reduce the risk of drops and speed up reviews when they do happen.
- Make changes in small batches.
- Avoid several big changes on the same day.
- Review auto-applied recommendations regularly.
- Write important changes into your log.
- Do a short weekly scan of the record.
For example, a weekly scan catches an unexpected system change before the drop grows. Also, to understand budget behavior, see our article on how the daily budget limit works. That way you separate budget-driven fluctuations from real drops.
What should your next step be?
When performance drops, open Google Ads change history first. Define the falling metric, find the affected campaign with comparison dates, and separate automated changes in the User column. These three steps reveal the cause of most drops.
If your account is crowded or you cannot find the cause, our google ads management service can review the account with you. We read the records first and then share our findings with you. You make the decisions, and we make sure they rest on evidence.
For a quick overview of account health, you can also start with our Google Ads audit tool. That way you begin the review prepared.
Core guide: What Is Digital Marketing? Its types, examples and how it is done, on one page with an interactive film that travels through the brain.



