Free tool · Google Ads · Runs in your browser
Account audit scorecard: one health score, measured against your own targets
Most account scorecards compare you to a benchmark that somebody else invented, and then sell you the plan to reach it. That is a sales technique, not a measurement. This one asks for your targets and scores you against those, so the number answers the only question that matters: are you performing to the plan you actually set?
1. What happened
One full billing period. Use the same window every month, or the trend means nothing.
Total spend for the period, all campaigns.
What you intended to spend.
From the search-term miner, not from a guess.
Figures are shown in dollars, and the shared formatter is deliberately currency-locked. Nothing here depends on that choice: cost per acquisition, return on ad spend and conversion rate are all ratios, so the currency cancels out of every one of them and out of the score. Only the displayed money figures would change. Enter anything, as long as spend and revenue are in the same unit.
2. Your targets
This is the whole point. A score against someone else's benchmark is not a measurement.
After fulfilment and overheads, before ads.
Conversions divided by sessions.
The share of spend you accept on terms that never convert.
Headline
CalculatingThe component breakdown, the reasoning behind every score and the prioritised action list unlock below.
Unlock the full audit
FreeGet the score for all five components, the formula and tolerance used for each, the weighted score table, the weakest component ranked by cost, and a prioritised action list with the reasoning shown. Plus CSV and print export.
-
-
Component scores
-Each component is scored out of 100 against the target you supplied. A component whose inputs are missing or impossible is excluded from the average and says so, rather than being scored zero.
What to do about it
Ranked by weightExport
How the score is actually built
Short answer
Five components, each scored out of 100 against a target you supply, then averaged using weights that are printed on this page. The score measures attainment of your plan. It does not measure quality, and it cannot tell you why a number is low.
Each component reduces to a ratio. For a metric where lower is better, such as cost per conversion, the ratio is your target divided by your actual. For a metric where higher is better, such as return against break-even, the ratio is your actual divided by your target. A ratio of 1.00 or better scores 100. Below that, the score falls at a fixed rate.
Once you have the five component scores, they are combined as a weighted average. The result is reported out of 100 and mapped to a grade. Nothing else is added, and no component is scored from a benchmark this page invented.
The formula, in full
ratio = 1.00 → score = 100
ratio < 1.00 → score = 100 × (1 − 0.5 × (1 − ratio))
score is clamped to a floor of 0
overall = Σ(component score × weight) ÷ Σ(scored weights)
The penalty rate is 0.5
At a ratio of 1.5 — that is, 50% worse than target — the component scores 75. At a ratio of 2.00 it scores 50. At 3.00 it reaches the floor of 0. The rate is a judgement call, it is a named constant, and changing it is a deliberate decision rather than a tweak.
The weights, and why they are what they are
A weighted average is only as honest as the weights, so they are stated rather than implied. If you disagree with one, change it in the explanation below and apply the same weighting to your own accounts.
| Component | Weight | Direction | Why it carries this weight |
|---|---|---|---|
| Cost per conversion | 25 | Lower is better | An account can survive a mediocre click-through rate. It cannot survive paying more per order than the order is worth. |
| Return against break-even | 25 | Higher is better | Return is the only component that decides whether the account is a business or an expense. |
| Waste share | 20 | Lower is better | Waste is recoverable this week. Poor return usually is not. |
| Budget utilisation | 15 | Inside a band | Under-spending means the plan was never tested. Over-spending means it was never agreed. |
| Conversion rate | 15 | Higher is better | A lever that does not cost money, and the slowest of the five to move. |
The weights sum to 100 so they can be read as percentages of influence. Budget utilisation is scored differently from the other four because it has a good and a bad direction, and is explained under thresholds.
Thresholds, tolerance and the five components
Every threshold in this tool is either something you supplied or a named constant printed here. None of them is a benchmark lifted from a competitor's marketing page.
Cost per conversion
actual = spend ÷ conversions against your target. If conversions is zero there is
no cost per conversion, so the component is not scored and says so. A dash, not an infinite number.
Return against break-even
actual = revenue ÷ spend against 100 ÷ margin. Break-even
return is the ratio at which advertising exactly consumes the gross margin, leaving nothing for
overheads. It is a function of your own margin, not a round number.
Waste share
actual = waste ÷ spend, as a percentage, against the share you said you would
tolerate. A negative tolerance is refused rather than clamped.
Budget utilisation
actual = spend ÷ planned × 100. This is the one component with a good and a
bad direction, so it is scored on distance from 100%: full marks inside 95–100%, then falling
at the same 0.5 rate as everything else per percentage point of distance.
Conversion rate
actual = conversions ÷ sessions × 100 against your target percentage. If
sessions is zero the component is not scored.
What the score cannot tell you
A score is a compression. It is genuinely useful for tracking a trend and genuinely dangerous as a diagnosis, because every component that was excluded is invisible in the total.
- It does not know your competitors' cost per conversion.
- It does not know whether a high cost per conversion is a bid problem or a margin problem.
- It does not see seasonality, so a bad month and a bad quarter score the same.
- It does not check that the revenue figure is real revenue rather than an attribution artefact.
- It does not look at one campaign carrying the account, which is a structural risk the score will not catch.
How to use it honestly
Use the score to notice that something moved, then use the component table to find what moved, then go and look at the actual data for that component. Skip the middle step and the score is worse than no score, because it will feel like an answer.
What the grades mean
| Score | Grade | What it actually means | What to do |
|---|---|---|---|
| 85–100 | Strong | Performing to plan on the components that could be scored. | Stop editing. Note the score and its date, and check again next period. |
| 70–84 | Healthy | One or two components are off target; nothing is broken. | Work the weakest component only. Changing two things at once makes the trend unreadable. |
| 50–69 | Needs work | At least one component is materially off plan. | Fix the highest-weight failing component before touching anything else. |
| Below 50 | Urgent | The account is not performing to its own plan. | Treat it as a decision, not an optimisation: cut, restructure, or change the target and say why. |
The band edges are judgement calls. They are named constants in the page source, and the reasoning is recorded in the tool's own research file.
How to run this every month without fooling yourself
- Fix the window first. One full billing period, same dates every month. A rolling seven days will swing on a single weekend.
- Get the actuals from the platform, not from memory. The score is only as good as the revenue figure, and attributed revenue is the easiest number in the account to overstate.
- Take waste from the search-term miner rather than estimating it. An estimated waste figure will make the waste component agree with whatever you hoped for.
- Do not move your targets mid-month to make the score look better. If the target genuinely changed, note the date and start a new series.
- Keep the scores. The trend is the output. A single reading tells you where you are today, which you already knew.
Where the numbers come from
- Spend, conversions and sessions: campaign and account reports, same date range.
- Revenue: the transaction or conversion value export, not a summary tile.
- Planned spend: whatever you actually wrote in the plan, not a rounded-up aspiration.
- Waste: the non-converting spend figure from the search-term miner.
- Margin: finance, not marketing. Fully loaded, after fulfilment.
The most common way this is misused
Setting the target to whatever the last result was, so the score lands near 100 and the meeting moves on. A target that is derived from the outcome measures nothing. Set the target from what the unit economics can bear, then let the score tell you how far away you are.
Taking the score somewhere useful
A number is only worth having if it changes a decision. Three things reliably follow from a scorecard like this one.
1. Setting the next target
If cost per conversion is the failing component and the target was ambitious, the useful question is whether the target was wrong, not whether the account is. If the target was set from real unit economics, the account is wrong. This tool cannot tell you which, because only you know how the target was set. That question is worth more than the score.
2. Prioritising the work
The component table is ordered by weight, so the top failing row is where the leverage is. An operator who fixes waste share and conversion rate while ignoring return on break-even is doing the easy work and leaving the expensive work.
3. Deciding to change the plan
If budget utilisation has been below the band for three periods, the plan is not being followed and the plan is probably wrong. Either raise the planned figure to something you will actually commit to, or cut the budget and stop pretending.
What pairs well with this
- The ROAS calculator to work out what a return target has to be before you set it.
- The search term miner for the waste figure this scorecard consumes.
- The budget pacer when utilisation is low because the daily budget is set below what the monthly plan requires.
When not to use it
- On a brand-new account with no conversion history. Most components will refuse to score and you will be reading a score built from one or two numbers.
- When the reporting window is too short to contain a full buying cycle.
- As a comparison between accounts, unless they share the same targets and margin.
- As evidence in a pitch. It measures attainment of a plan, not competitiveness.
What happens to your numbers
Short answer
They stay in your browser. The calculation runs in the page, the figures are never transmitted, and there is no analytics script on this site to collect them. Unlocking the audit sends your name and email so the copy can be emailed to you — not your account data.
You can disconnect and use the whole tool with JavaScript running and no network at all. The score, the component table and the export all work offline. That is worth knowing precisely because the numbers in this tool are commercially sensitive: they describe what you are paying and what you get back.
Details worth stating plainly
- No account figures are sent to any server, at any point, in either mode.
- No cookies are set by this tool and no third-party script is loaded.
- The unlock form stores a short-lived local token so you are not asked twice on the same device.
- Clearing your browser storage removes it immediately.
- The copy you request is emailed from our server, so the emailed copy is the one place your numbers exist outside your machine. Ask and we will delete it.
Last reviewed 2026.
Account audit questions, answered properly
What is a PPC account audit scorecard?
A scorecard turns several account measures into one number you can track month to month. It is useful because a single figure is hard to argue with in a meeting, but it is dangerous if the figure is produced by a benchmark nobody agreed to. This scorecard scores five components against targets you supply yourself, so the number measures performance against your plan rather than against somebody else's.
Where do the weights come from?
They are a judgement call, and they are printed on the page. Cost per conversion and return against break-even carry the most weight at twenty-five points each, because an account can survive a mediocre click-through rate but not a cost per conversion above the margin. Waste share carries twenty, budget utilisation fifteen, and conversion rate fifteen. Change a weight and the score changes, which is why the reasoning is shown rather than buried.
Why does my score change when I change a target?
Because the score measures attainment, and attainment is measured against a target. A higher target makes every component harder to satisfy, so the score falls. That is correct behaviour, not a bug: it means your plan has become more ambitious. What would be wrong is a score that stayed the same when the standard changed, because then the number would be measuring nothing at all.
Can this tool tell me what is wrong with my account?
No, and any tool that claims to is guessing. This one names the weakest component and gives you the next measurement to take. Deciding whether to cut a bid, rewrite an advertisement or pause a keyword needs the search terms, the ad copy and the search intent, none of which this page has. It is a scoreboard, not a diagnosis.
How often should I run this?
Once a month, on the same day, against a full billing period rather than a rolling week. A score computed from four days of data will swing on a single bad day, and comparing a strong week to a weak one tells you about the weather. Keep the previous score so the trend is the real output; a single reading only tells you where you are today.
Sources and method
What is measured, and where the formula comes from
Cost per conversion, return on ad spend and conversion rate are the standard ratio definitions used across advertising: spend divided by conversions, revenue divided by spend, and conversions divided by sessions. Break-even return is one divided by the gross margin fraction, which is the ratio at which advertising consumes the entire contribution.
Every other number in this tool — the five weights, the 0.5 penalty rate, the utilisation band, the grade edges — is a judgement call made for this tool. Each one is a named constant in the page source, is printed in the results, and is recorded with its reasoning in the tool's research file.
Deliberately not modelled
- Seasonality or a year-on-year comparison. It needs at least two full years of history.
- Impression share. It is only visible inside the platform, and guessing it would invent the single most-cited number in the trade.
- Any forecast of next month. A projection from one period of data is a confident-looking fiction.
- A competitive benchmark. There is no reliable free source of one, so there is no benchmark here.
Last reviewed 2026. If your own data contradicts a default here, change the default rather than the conclusion.