Ninety days is long enough for a competent agency to prove itself and short enough that a bad one has not yet cost you a year. The problem is what most owners measure at day 90: enquiry count. That single number moves for reasons the agency did not cause — a seasonal dip, a competitor pausing, a price change on your own site — and it hides the things that decide month six.
This page is a grading tool, not a forecast. If you want to know what results are normal at each stage, read what owners should expect to see from Google Ads. Here we do the opposite job: we grade the agency's behaviour against ten rows with fixed pass marks, so a good month cannot cover for bad management and a slow month cannot bury good work.
1. What is a Google Ads agency scorecard, and why grade at 90 days?
Quick Answer: A Google Ads agency scorecard is a fixed list of ten things the agency controls, each with a pass mark at day 30, 60 and 90. You grade the agency's work, not the account's luck. Ninety days covers one full learning cycle plus two months of managed performance, which is the earliest point the pattern is real.
Grading behaviour instead of outcomes matters because the two separate early. An account can post a decent cost per lead in month two purely because the previous manager left cheap brand traffic running. An account can look flat in month two while the agency rebuilds tracking that was double-counting every enquiry. Only one of those agencies is worth keeping.
- Day 30 grades control. Access, tracking, a delivered audit and a plan. Nothing here depends on the auction.
- Day 60 grades hygiene. The rebuild is live, search terms are being worked, and the first real changes are explained in writing.
- Day 90 grades direction. Cost per lead against baseline, lead quality feedback flowing, and a report that ties spend to enquiries.
The wider hiring checks — proposals, fee models, references — sit in our Google Ads agency guide. This page starts after you have signed, when the only evidence left is what the agency actually does.
Key takeaway: Grade what the agency controls. Enquiry count alone rewards luck in month two and punishes honest rebuilding work.
The video below walks through the same idea from the client side: what to look at when you are deciding whether a PPC agency is earning its fee.
Want a second opinion before you score your own agency?
ZenWeb reviews Malaysian accounts against the same ten rows and tells you plainly which ones are failing.
See what ZenWeb's Google Ads management covers →2. The ten rows to grade your Google Ads agency on
Quick Answer: The ten rows are account ownership, conversion tracking integrity, search-term hygiene, campaign structure, cost per lead trend, lead quality feedback, landing page work, reporting, communication and documentation. Each row has a different pass mark at day 30, 60 and 90, so partial progress is still gradeable.
Print this table and mark it yourself. Every pass mark is something you can verify from inside your own account or your own inbox — no agency dashboard required. Rows one and two are the ones that make the other eight worth grading, which is why they come first.
| Row | Pass by day 30 | Pass by day 60 | Pass by day 90 |
|---|---|---|---|
| 1. Account ownership | Account in your name, agency added as manager | Still true after any rebuild | You can log in alone and see everything |
| 2. Conversion tracking integrity | Every lead action fires once, tested in front of you | Duplicates removed, values attached | Offline or CRM outcomes imported |
| 3. Search-term hygiene | First negative keyword list live | Evidence of a weekly review | Irrelevant spend share under 15% |
| 4. Campaign structure | Audit delivered, rebuild plan agreed | Rebuild live and budgeted | Structure stable, no weekly churn |
| 5. Cost per lead trend | Baseline recorded in writing | First fall against baseline | At or below the agreed target |
| 6. Lead quality feedback | A written definition of a good lead | Your sales feedback reaching the agency | Bidding reflects lead quality, not volume |
| 7. Landing page work | Page issues listed and prioritised | At least one change or test live | Result reported, next test named |
| 8. Reporting | First report on an agreed template | Report explains changes, not just numbers | Report links spend to enquiries and sales |
| 9. Communication | Named manager and agreed cadence | Queries answered within one working day | No unexplained silent weeks |
| 10. Documentation | Change log started | Change log current and shared | A handover pack exists without asking |
Source: ZenWeb scorecard, built from ZenWeb operational data across Malaysian SME Google Ads accounts taken over from a previous manager, 2024–2026.
Row one is settled by Google's own account settings rather than by your contract, so check it directly against Google's guidance on access levels in a Google Ads account, then read our fuller note on who should own your Google Ads account in Malaysia.
Key takeaway: Every pass mark is verifiable from your own account or inbox. If a row can only be confirmed by the agency, it is not a pass.
3. Day 30: has the agency taken control of the account?
Quick Answer: At day 30 you are grading control, not performance. Access is correct, conversion tracking has been tested in front of you, an audit has been delivered, and a rebuild plan is agreed in writing. Nothing on this list depends on the auction, so there is no acceptable excuse for a miss.
Owners routinely forgive a weak first month because “it takes time”. Results take time. Access does not. A tested conversion does not. An audit does not. What our Google Ads agency onboarding guide sets out as the first-30-days plan is exactly what you are grading here.
Be strictest on tracking. Watch the agency submit a test enquiry, then check the conversion appears once. Doubled conversions make every later number look better than it is, and the fix gets harder the longer bidding trains on bad data — see our Google Ads conversion tracking setup walkthrough.
Key takeaway: A day-30 miss on access, tracking or the audit is a process failure, not a market problem. Raise it in month one, not month four.
4. Day 60: is the account getting cleaner or just bigger?
Quick Answer: At day 60 the honest signal is subtraction. A good agency has removed search terms, paused weak assets and tightened targeting, and can show you the list. A weak agency has mostly added — more keywords, more campaigns, more budget — because addition looks like effort in a report.
Open the search terms report yourself. Google documents it in its own help centre under the search terms report, and it is the cheapest audit you can run without any tools. You are looking for two things: queries that obviously do not describe your business, and how much of last month's spend went to them.
- Ask for the negative list, dated. A list with no additions since week two means nobody is reviewing, whatever the report says.
- Ask what was paused, not what was launched. Named pauses with reasons are the mark of someone reading the data.
- Ask why budget moved. Any reallocation should trace back to a search term, a conversion, or a landing page result.
If the report is full of queries you would never pay for, the fix is well documented in how to fix irrelevant Google Ads search terms and in our guide to negative keywords. Neither is advanced work. Its absence is a management signal.
Key takeaway: Grade day 60 on what the agency removed. Growth in campaign count without a matching negative list is activity, not management.
5. Day 90: has cost per lead moved in the right direction?
Quick Answer: By day 90 cost per lead should be at or below the target you agreed at signing, measured against a baseline recorded in month one. Direction matters more than the absolute number. A flat cost per lead with better lead quality is a pass; a lower one built on cheaper, worse enquiries is not.
This row only works if a baseline exists, which is why it has a day-30 pass mark of its own. Without one, month three becomes an argument about what the number used to be. Agree the definition first, using our note on what cost per lead actually means.
Then judge cost and quality together. Ask your sales team how many of last month's enquiries were worth calling back, and compare that share to month one. If cost per lead fell while that share fell further, the agency bought volume — fixes in how to lower a Google Ads cost per lead and in our guide to a Google Ads agency for lead generation.
Key takeaway: Score row five on cost per lead and lead quality together. One without the other is half a result.
6. Which scorecard rows do agencies fail most often?
Quick Answer: Across Malaysian accounts ZenWeb has taken over, conversion tracking integrity fails most often, followed by documentation and search-term hygiene. Cost per lead trend fails least. Agencies are usually better at buying clicks than at proving what those clicks did.
The pattern is consistent and slightly uncomfortable: the rows that fail most are the invisible ones. Nobody notices a missing change log until the day they need it, and a double-counted conversion looks like good news until someone checks.
| Scorecard row | Failure rate at handover | % |
|---|---|---|
| Conversion tracking integrity | 68% | |
| Documentation and change log | 61% | |
| Search-term hygiene | 54% | |
| Lead quality feedback | 47% | |
| Reporting | 39% | |
| Landing page work | 33% | |
| Account ownership | 27% | |
| Communication | 22% | |
| Campaign structure | 19% | |
| Cost per lead trend | 16% |
Source: ZenWeb operational data, Malaysian SME Google Ads accounts audited at takeover from a previous manager, 2024–2026. A row counts as failing if it missed its day-90 pass mark.
If you want the full technical version of this check rather than the ten-row summary, our 12-check Google Ads audit covers the same ground at account level.
Key takeaway: The rows that fail most are the ones nobody looks at. Check tracking and documentation first, because they are also the two that quietly invalidate every other row.
Not sure whether your tracking is telling the truth?
A tracking check is the fastest way to find out whether the last three months of reporting meant anything.
Compare ZenWeb's Google Ads management plans →7. Does the day-90 score predict what happens later?
Quick Answer: In ZenWeb's client sample it does. Accounts scoring eight or more rows at day 90 saw cost per lead fall by about a third by month six. Accounts scoring five or fewer saw it rise slightly, and most of those relationships had ended before month twelve.
The gap widens rather than closes, which is the argument against giving a weak agency “one more quarter”. Poor tracking and thin documentation compound: every later decision is made on worse information than the last.
| Day-90 score | Cost per lead change by month 6 | Enquiries per month at month 6 | Still running at month 12 |
|---|---|---|---|
| 8–10 rows passed | −34% | 61 | 84% |
| 6–7 rows passed | −11% | 44 | 52% |
| 5 rows or fewer | +6% | 31 | 19% |
Source: Aggregated from ZenWeb-managed Google Ads accounts, Malaysia, 2024–2026. Medians; cost per lead change measured against each account's month-one baseline. Darker cells mark the best result.
Key takeaway: A weak day-90 score rarely recovers on its own. The gap between bands widened by month six in every cohort we tracked.
8. How the score should move across the first 90 days
Quick Answer: A healthy account gains roughly two rows a month. Agencies later retained were already passing four to five rows at day 30 and close to nine by day 90. Agencies later replaced started at two and were still under five at day 90 — the shortfall was visible in month one.
Score all three checkpoints rather than only the last one. The slope tells you more than the total, and it tells you earlier.
| Measure | Day 30 | Day 60 | Day 90 |
|---|---|---|---|
| All accounts, rows passed | 3.4 | 5.6 | 7.1 |
| Agencies later retained | 4.6 | 7.2 | 8.9 |
| Agencies later replaced | 2.1 | 3.4 | 4.2 |
| Share passing the tracking row | 41% | 73% | 88% |
Source: ZenWeb operational data, Malaysian SME Google Ads accounts under management, 2024–2026. Medians; retained and replaced measured at month twelve.
Two rows a month is the pace to expect from a properly resourced retainer, and it is one of the things a fee should buy — the trade-offs are set out in our comparison of flat fee versus percentage-of-spend management fees.
Key takeaway: Score at all three checkpoints. A day-30 score of two rows is already the warning that a day-90 score of four is coming.
9. How to run the scorecard in twenty minutes
Quick Answer: Log in yourself, mark each row pass or fail against the checkpoint you have reached, then total it. Rows one and two are non-negotiable: a fail on account ownership or conversion tracking caps the whole scorecard at a fail, whatever the other eight say.
- Log in to Google Ads with your own email. If you cannot, row one has already failed and the rest of the exercise is guesswork.
- Test one conversion. Submit an enquiry through your own site, then check it appears in the account once, with a value attached.
- Read last month's search terms. Mark row three against the share of spend on queries you would never pay for.
- Open the last three reports side by side. Mark rows five and eight on whether the same metrics appear each month and whether changes are explained.
- Ask for the change log and the handover pack. Mark row ten on how long the answer takes. A day is fine; a week is a fail.
- Total the passes and apply the cap. Eight or more is a keep, six or seven is a conversation, five or fewer is a decision.
Set the same measures out at the start of the next relationship so nobody is scored on a moving target — the method is in how to set marketing KPIs with your agency, and the reporting standard in what your agency should show you monthly.
Key takeaway: Ownership and tracking are pass-or-fail gates. Without them the other eight rows are being scored on numbers you cannot trust.
10. What to do with a failing scorecard
Quick Answer: Send the marked scorecard, name the failing rows, and ask for a dated plan on the two that matter most. Give thirty days, not another quarter. If ownership or tracking is among the failures, secure your account access before any conversation about ending the contract.
Most agencies respond well to a specific list. “Leads are down” invites excuses; “rows two, three and ten failed at day 90, here is the evidence” invites work. Send it in writing and ask for dates against each row.
- Protect access first. Confirm the account, tag manager and analytics property are in your name before anything else.
- Fix two rows, not ten. Tracking and search terms move the other rows on their own.
- Set a hard review date. Thirty days, same scorecard, same evidence standard.
- Prepare the alternative quietly. Knowing the switching sequence removes the pressure to accept a weak recovery plan.
If you do move, protect the history rather than starting from zero, using how to change Google Ads agency without losing data and the warning signs in Google Ads agency red flags. If the real question is whether to keep any agency, our note on whether a Google Ads agency is worth it works that through against what great PPC management looks like.
Key takeaway: A marked scorecard turns a vague complaint into a fixable brief. Thirty days and two named rows is a fair, testable ask.
11. Conclusion: score the work, not the weather
Ninety days is enough to know. The account will not have hit its ceiling by then, and it does not need to. The ten things that decide the next twelve months are all visible by day 90, and none of them depend on having had a good month.
ZenWeb runs this scorecard on every account it takes over, and publishes it so you can run it on us as easily as on anyone else. As a Google Partner with 500+ Malaysian clients, we would rather be graded on rows one and two than on a screenshot of a good week. Start with the Google Ads agency checklist if you are still choosing, use the Penang hiring guide if your buyers are up north, and hold whoever you appoint to these ten rows. Everything ZenWeb builds stays in your name if you ever leave.
Want your account scored against all ten rows?
ZenWeb will mark your Google Ads account row by row — ownership, tracking, search terms, reporting and the rest — and send you the completed scorecard with the evidence behind each mark.
Request a scorecard review →
12. Frequently Asked Questions
1. Is 90 days long enough to judge a Google Ads agency?
Yes, if you grade the right things. Ninety days is not always long enough for cost per lead to reach target, but it is more than long enough for account ownership, conversion tracking, search-term hygiene, reporting and documentation to be in order. In ZenWeb's client sample, agencies passing eight or more scorecard rows at day 90 went on to cut cost per lead by about a third by month six.
2. What score should I expect from a good Google Ads agency at day 90?
Eight or more rows out of ten. Across ZenWeb-managed Malaysian accounts, agencies still running the account at month twelve had a median of 8.9 rows passed at day 90, against 4.2 for those replaced. Six or seven rows is worth a conversation with a thirty-day review; five or fewer is a decision.
3. Which scorecard rows matter most if I only have time to check two?
Account ownership and conversion tracking. Ownership decides whether you keep the account's history if the relationship ends, and tracking decides whether every other number in the report means anything. Conversion tracking was the most commonly failing row at handover in ZenWeb's takeover audits, at 68% of accounts.
