Quick Answer
Use Google Ads change history to build a timeline of relevant account changes, then compare it with the affected campaign’s performance and measurement records. Check what changed, when, through which user or tool and what else happened nearby. A change before a decline is a hypothesis to investigate, not proof that it caused the decline.
Enquiries fall after a campaign update, and the first response is to undo everything. That feels decisive, but it can remove a useful change, hide a tracking fault or create several new variables before anyone understands the original problem.
For a Pakistani business reviewing Google Ads performance, the better starting point is a shared timeline. The advertising account, website and sales process may all have changed during the same period. Change history helps inspect part of that story, while an investigation explains what the evidence does and does not connect.
What should you define before opening change history?
Define the decline in a specific metric, campaign and period. “The ads stopped working” is too broad to compare with individual account changes.
State whether the concern is fewer impressions, fewer clicks, fewer recorded conversions or fewer useful enquiries. Those are different observations and may require different supporting evidence. A dashboard decline and a business-outcome decline should not be treated as the same event without checking.
Record the first period in which the change is visible, the comparison period and why that comparison is reasonable. A single day may be an alert rather than a sufficient diagnosis, especially when the business has a limited flow of enquiries.
The conversion tracking and lead-quality guide provides context for separating recorded actions from business outcomes. This investigation should stay narrower: identify the affected advertising scope and reconstruct the changes around the observed decline.
What can Google Ads change history establish?
It can identify recorded account changes and help place them beside performance. It cannot, on its own, establish every reason performance moved.
Google’s About change history documentation describes a two-year history, date and change-type filtering, and a timeline linked to metrics such as clicks, conversions and cost. It also explains user and system or tool attribution for recorded changes.
Use those details to ask precise questions. Was the affected campaign changed, or another campaign? Did the entry represent a direct user action, an API or tool action, or a system-labelled activity? The label is evidence to inspect, not a verdict about responsibility or intent.
Keep a separate record of website releases and operational changes. A form failure, an altered offer or delayed customer responses may matter to enquiries even when the advertising settings remain unchanged. Do not expect one account log to replace those records.
How should you scope the history review?
Review the affected campaign and relevant dates first, then widen the scope only when the evidence calls for it. An unfiltered account history can bury a useful change among unrelated activity.
Google’s account-history review instructions describe selecting dates and scope, opening change details and filtering by fields including campaign, user, tool and change type. Available navigation may differ with the current interface and access.
Start the history period before the visible decline so you can see preceding changes. Include enough surrounding context to distinguish an isolated action from a sequence. Keep the report date range and the history date range clearly labelled rather than assuming they are identical.
Record the values and affected items shown in the relevant entries. “Budget changed” is less useful than a documented description of which campaign changed and how the before-and-after settings differ. Use the actual account record instead of reconstructing details from memory.
Which entries deserve attention first?
Prioritise entries that could plausibly affect the metric and campaign under review. The most recent change is not automatically the most relevant one.
| Observed problem | Entry worth examining | Additional check |
|---|---|---|
| Impressions fell | Campaign status or distribution change | Was the affected scope eligible to serve? |
| Click pattern changed | Ad, keyword or targeting change | Did the traffic mix change? |
| Reported conversions fell | Conversion-action change | Does the website action still work and record correctly? |
| Enquiry quality fell | Offer or traffic-selection change | What do the actual enquiry records show? |
| Several metrics moved together | A sequence of account edits | What changed on the website or in operations? |
The table is an investigation aid, not a list of guaranteed causes. Use it to choose what to inspect next and discard a hypothesis when the evidence does not fit.
For example, an ad-group edit is less persuasive as an explanation for an unrelated campaign’s decline than a change affecting the campaign being reviewed. Scope matters as much as timing.
How can you build a timeline that separates fact from inference?
Keep observed changes, observed performance and proposed explanations in separate columns. That prevents an early theory from becoming an accepted fact through repeated reporting.
For each relevant point, record the date, affected item, evidence location and description. Add a separate field for why it might matter. Then note what would support or weaken that explanation.
A useful entry reads: “A conversion action was edited before recorded conversions declined; website behaviour and tag receipt need checking.” A weaker entry reads: “The edit broke the leads.” The first tells the team what to verify; the second skips the verification.
Keep the investigation record with the account owner. If an external reviewer is involved, share appropriate access or a scoped export rather than distributing account credentials. The agency access and reporting checklist covers the wider working relationship.
What does a worked investigation look like?
A worked investigation should show competing explanations and the next discriminating check. It should not select a cause solely because one event happened first.
Imagine a fictional Islamabad service business notices fewer reported conversions after a website update and an advertising budget adjustment in the same week. Clicks remain broadly similar, while the sales team is unsure whether actual enquiries changed. This is an illustrative situation, not a client result.
The history confirms the budget adjustment, but that does not resolve the conversion question. The reviewer records the website release separately and asks for a controlled test of the relevant contact action and its measurement. If the action works but the event is missing, that points to a reporting investigation rather than proof that demand disappeared.
If the tracking works, the reviewer can then compare campaign scope, traffic patterns and enquiry records. The next step depends on the result of the test. Reversing both the website update and budget adjustment immediately would make it harder to learn which issue actually mattered.
How should you handle automated or unfamiliar entries?
Identify the tool or system and inspect the entry before drawing conclusions about who changed the account. An unfamiliar label does not by itself prove unauthorised interference or a performance problem.
Google’s documentation notes that automated systems and API-based changes can appear in history. Ask the account owner which approved integrations, rules or management tools were active during the period, then compare those records with the entry.
Avoid assigning blame from an email address or tool label alone. The person associated with a recorded action may need context about the requested work, the affected scope or an automated process. A review should produce a testable explanation, not an unsupported accusation.
If a change appears outside the approved scope, preserve the evidence and raise it with the responsible owner. Keep access or security decisions separate from a routine performance hypothesis; do not alter integrations simply because their names are unfamiliar.
Should you undo a change to test the explanation?
Consider a reversal only after identifying the specific change, expected effect and authorised test. A broad rollback is a poor substitute for a defined investigation.
Google currently allows most change types from the last 30 days to be undone, subject to exceptions. If one row contains several changes, Undo reverts them together. That does not mean every entry can be reversed or that reversing it will recreate the earlier trading conditions. Check the current availability and the scope of the action before proposing it.
Write down what the test is expected to affect and what else should remain stable. Agree how the team will judge the result and whether the chosen outcome has had enough time for useful review. Do not promise an immediate return to an earlier level of performance.
For a discussion of audit scope, review the digital marketing services information. Bring the timeline and the unresolved question, rather than requesting a campaign reset without evidence.
What should the investigation handover contain?
The handover should state the confirmed facts, remaining hypotheses and next check. A useful diagnosis may end with an unresolved question if the necessary evidence is not available.
- Define the affected metric, campaign and comparison dates.
- Identify the relevant history entries and their evidence.
- Include website and operational changes separately.
- List tests completed and what each established.
- State any proposed change, owner and approval requirement.
- Record what would change the current conclusion.
Preserve the original report settings and timeline so the next reviewer can follow the reasoning. Do not rewrite an uncertain finding as a confirmed cause merely to make the report sound decisive.
Frequently Asked Questions
No. Timing can identify a hypothesis, but the affected scope and a plausible explanation still need checking. Compare other account, website and operational changes before deciding what the sequence means.
Google’s current documentation describes a two-year history. Keep your own appropriate review records rather than assuming every older account detail will remain available through the same interface indefinitely.
Not by itself. Review the website’s release and test records separately. A contact-action failure or measurement fault may affect the report without appearing as the relevant advertising account change.
No. Inspect the entry, tool and affected item, then ask the account owner about approved automation. A system label alone does not establish a harmful effect or an unauthorised action.
No. Identify and test a specific explanation first. A broad rollback changes several variables and may remove useful work. Any proposed reversal needs an authorised scope and a clear review plan.
Report that limit and state the next useful check. A documented set of unresolved hypotheses is more useful than confidently naming a cause that the available account and business records do not establish.
Final Thoughts
Change history is most useful when it supports a disciplined timeline. Connect the recorded action to the affected scope, test alternative explanations and keep observed facts separate from the cause you are investigating.
For a focused account review, share the performance question and available evidence with M Khubaib Zia. Start with the metric and period that changed, then agree what the investigation should establish.