In the first 30 days on a new client account, preserve what is working, verify measurement, remove only clear waste and then run controlled tests. Spend week one observing, week two fixing tracking, week three correcting confirmed problems and week four testing the highest-value hypothesis.
When taking over a Google Ads account, the goal is not to prove activity through a long change log. It is to build enough understanding that every important change has a reason, a measurement plan and a way back.
Start with a baseline, not a rebuild
A new account creates pressure to move quickly. The client wants progress, and the incoming team wants the account to reflect how it works. That does not make an immediate rebuild the right move.
Inherited structure contains information. Campaign names, budget splits, match types, exclusions, audiences, conversion actions and bid strategies show how previous decisions were made. Some decisions will be outdated. Others may be carrying performance that is easy to disrupt and hard to recreate.
Before changing anything, answer five questions:
- Who owns the account, and who has administrative access?
- Which conversion actions are used for bidding and reporting?
- Where do budgets, targets and exclusions live?
- What changed recently?
- Which campaigns are stable enough to leave alone?
Account ownership deserves special attention. The client should keep owning the ad account while the agency receives the access needed to manage it. That keeps the handoff clean and avoids making account history dependent on the agency relationship. The mechanics are covered in Who owns the Google Ads account?.
Create a written baseline before editing. Record the date, active campaigns, budgets, bid strategies, conversion actions, target settings, recent performance and any known client constraints. Note open questions separately from confirmed problems.
Use a simple decision table:
| Finding | Change now | Leave alone for now |
|---|---|---|
| Broken or duplicated conversion action | Correct once the source is confirmed | Do not change bidding until the effect is understood |
| Expired payment method | Resolve immediately with the client | Do not use campaign edits to compensate for stopped delivery |
| Stable campaign aligned with the brief | Preserve it | Watch for changes in volume, cost and conversion quality |
| Naming or structure you would organize differently | Document it | Avoid rebuilding solely for preference |
| Search terms clearly outside the offer | Add precise exclusions | Do not apply a broad negative list without reviewing conflicts |
| Weak performance with little recent data | Form a hypothesis | Wait for enough evidence before making a major change |
This distinction matters when taking over a Google Ads account. “I would build it differently” is not the same as “this setup is causing a measurable problem.”
Week 1: Observe the system and protect the account
The first week is for orientation and risk control. Make changes only when waiting creates a clear operational or measurement problem.
Confirm access, ownership and billing
Verify that the client owns the Google Ads account and that the agency has the correct level of access. Review manager-account connections, user permissions, billing status and payment alerts.
Do not treat billing as an administrative footnote. An expired card can stop delivery while every campaign setting still looks correct. Record who handles payment issues and how quickly the client wants to be contacted.
Review access across the rest of the measurement chain as well. That may include analytics, tag management, the website, call tracking and the client’s lead system. Access gaps can turn a simple tracking question into days of guesswork.
Read the change history
The change history shows whether a performance shift followed a budget edit, target change, conversion adjustment or structural move. Read it before drawing conclusions from a short reporting window.
Look for:
- Large budget or bid changes
- Campaigns enabled or paused
- Conversion actions added or removed
- Broad changes to location, audience or schedule settings
- New exclusions
- Changes made shortly before the handoff
A poor recent result does not always mean the underlying campaign is poor. It may reflect a recent change that has not had time to settle, a tracking interruption or a promotion that ended.
Map the account to the client’s actual business
Ask how the client makes money, which offers are current and which leads matter. Confirm locations, service areas, sales constraints, seasonal factors and offer expiration dates.
Then compare that reality with the account. A campaign can be technically tidy and still point at the wrong offer, optimize for an unqualified action or send traffic to a page that no longer matches the ad.
A detailed Google Ads account review checklist can help organize this pass without turning the first week into a rebuild.
Send a factual first-week update
Tell the client what you confirmed, what remains uncertain and what requires immediate attention. Keep hypotheses clearly labeled.
A useful first update might say:
- Account access and ownership are confirmed.
- Billing is active.
- Two conversion actions need validation before bidding changes.
- One campaign has a clear location-setting conflict.
- The current structure will remain in place while recent changes are evaluated.
That message sets the operating expectation: deliberate changes, visible reasoning and no manufactured urgency.
Week 2: Fix measurement before judging performance
Week two is where you make the numbers trustworthy enough to guide decisions. If measurement is wrong, campaign comparisons become arguments built on bad inputs.
Inventory every conversion action
List each conversion action and identify:
- What user behavior triggers it
- Where the event is recorded
- Whether it is primary or secondary
- Whether it is included in bidding
- Whether duplicate events are possible
- Whether the action represents a lead, a sale or an earlier step
Do not assume that a conversion named “Lead” represents a qualified lead. It may fire on a form view, button click or partial submission. Test the actual user path.
Compare platform totals with the systems the client uses to receive and manage leads. The totals may not match exactly because the systems can use different attribution rules, but the differences should be understood rather than ignored.
Protect history while correcting errors
Fix confirmed tracking problems, but document the old state first. Record what changed, why it changed and when reporting should be reviewed again.
If a conversion action is duplicated, for example, confirm the duplicate before removing it from bidding. If a primary action is wrong, decide whether it should become secondary or be replaced. Avoid deleting useful historical signals simply because the naming is poor.
Example: separate a signal problem from a campaign problem
Example: Say a campaign has a $50 target CPA and appears to be reporting leads below that target. The account may look healthy until you discover that both the form submission and the thank-you page load are counted as primary conversions.
The first move is not to increase the budget. Correct the measurement, note the date and allow the account to produce a new baseline. Only then can you judge whether the campaign is meeting its target.
This is also the week to confirm that landing pages match the ads and current offers. A campaign should not be blamed for a page that promotes an expired offer or sends every audience to the same generic destination.
Week 3: Remove confirmed waste without disturbing winners
By week three, you should know which problems are real and which observations still need time. Now you can correct waste with more confidence.
Work from the largest controllable problem
Rank issues by their effect on measurement, budget control and business fit. Start with the problem you can explain clearly.
Possible candidates include:
- Search terms that do not match the offer
- Locations outside the service area
- Budgets that conflict with the agreed plan
- Active ads tied to expired offers
- Campaigns optimizing toward the wrong conversion action
- Duplicate campaigns competing for the same job
- Landing pages that do not match the promise in the ad
Make one class of change at a time where practical. If you change targeting, bidding, creative and the landing page together, you may improve the account but learn little about why.
Use account targets as fallbacks, not blunt rules
Account-level targets can provide a default, but campaigns often have different economics. Apply campaign-level CPA or ROAS targets where the business case requires them.
Example: Say the account has a general $50 target CPA. One campaign may support a higher target because its leads carry more value, while another may need a lower target because it promotes an entry-level offer. Applying the same target everywhere can hide those differences.
Group rules can help when several campaigns share the same purpose. A buyer might evaluate all TikTok traffic campaigns using link cost per click while applying a different measure to lead-generation campaigns. The point is to encode how the team actually makes decisions.
Cap the size of budget moves
Large changes can make it difficult to separate normal variation from the effect of the decision. Set a daily movement limit around campaigns that are already working, then adjust it when the account’s volume and risk justify doing so.
Keep a change register with five fields:
| Field | What to record |
|---|---|
| Decision | The exact setting, budget or asset changed |
| Reason | The evidence behind the change |
| Expected signal | What should move if the hypothesis is correct |
| Review point | When the result will be evaluated |
| Rollback | How to restore the previous state |
This makes client reporting easier because the team can explain not only what happened, but what it expected to learn.
Week 4: Test a clear hypothesis and set the next cadence
Week four is for controlled testing. Choose a question worth answering, protect the current winner and define what will count as useful evidence.
Test beside the current version
Where the platform and account structure allow it, run the proposed version beside the original. Keep the test narrow enough that the result can inform a decision.
A useful hypothesis is specific:
The landing page that matches the campaign’s current offer will produce better lead efficiency than the generic service page.
A weak hypothesis is vague:
The new setup will perform better.
State the metric, the affected campaign and the conditions that could make the result inconclusive. Low traffic, delayed conversions or uneven delivery can all prevent a clean call. “Inconclusive” is a valid result when the data does not support a decision.
Review creative across campaigns and platforms
The same asset may run in several campaigns or on several platforms. Evaluate it as one asset as well as within each placement.
Look for fatigue signals such as rising frequency, higher CPM or CPC and worsening CPA. A fatigue signal should produce a recommendation for review, not an automatic swap. The buyer still decides whether staged creative is ready and appropriate.
Set client expectations for month two
The first-month report should separate completed work from emerging evidence.
Cover:
- What was verified
- What was corrected
- What was intentionally preserved
- What was tested
- What remains inconclusive
- What will be reviewed next
Do not promise a date when performance will improve. Traffic volume, conversion lag, budget and the condition of the inherited account all affect how quickly a useful signal appears.
The strongest expectation to set is procedural: changes will be tied to evidence, reviewed after they run and reversed when they do not hold up. That gives the client a clear way to judge the work without forcing an early rebuild.
Frequently asked questions
What should you do first when taking over a Google Ads account?
Confirm access, ownership, billing, conversion tracking and recent change history before editing campaigns. Record a baseline so later decisions can be tied to actual account data.
Should you restructure an inherited account?
Restructure only when the current setup prevents clear measurement, budget control or useful testing. If the account is functioning, preserve it long enough to understand what works before replacing its structure.
How long before results improve on a new account?
There is no fixed timeline because traffic volume, conversion lag, budget and account condition vary. The first month should establish reliable measurement and produce controlled evidence, not a promised performance date.
What should you tell a new client in the first month?
Explain what you are observing, what you are correcting and which decisions still need more data. Separate confirmed problems from hypotheses and give every major change a reason, owner and review date.
What changes are safe in the first week?
Safe first-week changes are usually protective and reversible, such as fixing broken tracking, resolving billing risks and correcting settings that conflict with the agreed plan. Avoid broad structural changes until you understand why the account was built as it was.
How Ads Arsenal supports the first 30 days
Ads Arsenal keeps the approval decision with the buyer. When you approve a change, it writes the change to the account, checks a few minutes later that it landed and tracks the result. Any change can be rolled back with one click, and nothing changes until someone approves it.
The client keeps ownership of the ad account while the agency or Ads Arsenal receives administrative access, so access can be removed without losing the account’s data. Meta connects when you sign in with your Meta account. During the closed beta, Google connects through a manager-access request set up with you on a short call.
Ads Arsenal is in a free design-partner beta for agencies and in-house teams running real spend on Google, Meta and TikTok. Become a design partner.