How to Calculate Wasted Ad Spend Honestly

Why "spend with no conversion" overstates waste

If you filter your search terms report to conversions = 0 and sum the cost column, you have not measured waste. You have measured the cost of everything except your winners, which in a healthy account is most of your budget.

Do the arithmetic. An account spends $20,000 in 90 days at a $2 average CPC. That is 10,000 clicks. At a 2% conversion rate it produces 200 conversions and a $100 CPA. If $100 is under your target, this account is working. And yet 9,800 clicks — 98% of them — produced nothing. Filter for zero conversions and the report will tell you that you wasted something close to $19,000.

That number is not wrong arithmetically. It is wrong as a description of the world. Those 9,800 clicks are not a leak. They are what you bought in order to get 200 conversions. Nobody looks at a store with a 2% purchase rate and says 98% of the footfall was wasted.

The same confusion shows up at term level. A search term spends $12 across four clicks and converts nothing. Your target CPA is $120. That term has not failed — it has barely been tested. Ten clicks on a good term at a 2% conversion rate will usually produce zero conversions. That is not evidence of anything.

This matters beyond bookkeeping. The naive number is the one quoted in free audits, and it justifies aggressive blocking: cut everything that did not convert, and you cut the long tail that was quietly producing your cheapest leads. If you are weighing up whether to run this yourself or hand it over, the honest maths on agency versus DIY is worth reading alongside this, because both decisions turn on getting the waste figure right in the first place.

The honest definition: intent, not outcome

Wasted ad spend is money spent on search terms that could not have converted no matter what happened after the click. You decide that by reading the term, not by looking at the conversions column.

This is the whole shift. Outcome-based waste is retrospective and noisy. Intent-based waste is readable and stable. A term like "plumber salary" was never going to convert for a plumbing company. Neither was "plumber apprenticeship near me", "how to unblock a drain yourself", "free plumbing quote template" or "plumbing jobs indeed". You do not need conversion data to know this. You need to read six words.

The categories repeat across almost every account we audit:

• Employment. "salary", "jobs", "hiring", "apprenticeship", "career", "cv". These are consistent enough to be worth their own treatment — see job seekers clicking your ads.
• DIY and instructional. "how to", "yourself", "tutorial", "diy", "guide". Someone learning to do it is not hiring you to do it.
• Free and cheap. "free", "template", "open source", "crack", "trial no credit card" when you sell a paid product with no free tier.
• Wrong product. You sell commercial espresso machines; the term is "nespresso pods".
• Wrong geography. You serve Texas; the term names a city in Ontario.
• Research and definitions. "what is", "meaning", "wikipedia", "reddit".

Some buckets are genuinely contested. Competitor names are the obvious one — for some accounts they are the best-performing terms in the account, for others they are a slow bleed. That is a judgement call with money on both sides, and we have laid out both sides of the competitor question separately. For the calculation below, put contested terms in their own line and decide deliberately rather than sweeping them into "waste".

The four-bucket calculation

Export 90 days of search terms, then sort every term into one of four buckets. Waste is bucket two plus part of bucket four. Buckets one and three are the cost of doing business.

Bucket 1 — converted. Any term with at least one conversion in the window. Not waste, full stop. Even a term at three times your target CPA is not waste; it is an expensive term that needs a bid or a negative at phrase level, which is a different problem.

Bucket 2 — could never convert. Zero conversions and the intent is wrong. This is real, immediate, unarguable waste. You can block it tonight.

Bucket 3 — right intent, not enough data. Zero conversions, but the term has spent less than one target CPA. A $30 spend against a $120 target CPA tells you nothing. Blocking these is how accounts shrink. Leave them.

Bucket 4 — right intent, spent past the point of doubt. Zero conversions and the term has spent two or more times your target CPA. "emergency plumber london" with $400 spent and nothing to show is a real problem — but it is a landing page, offer, hours or tracking problem at least as often as it is a keyword problem. Count it as waste in the number; do not automatically block it.

The threshold in bucket 4 is yours to set. Two times target CPA is a reasonable default. Below that, single-term data is too thin. Above three times, you are tolerating a lot of loss before you look.

A worked example

An account spends $20,000 over 90 days and records 200 conversions. Blended CPA is $100 against a target of $120. Sorted:

The naive figure — everything that did not convert — is $8,000, or 40% of spend. Presented in a sales deck, that is alarming.

The honest figure is $1,800 + $2,000 = $3,800, or 19%. And the two halves behave completely differently. The $1,800 is safe to remove with negative keywords this week and it will not cost you a single conversion. The $2,000 sits across maybe six or eight terms averaging $250 to $350 each, and each one deserves five minutes of thought before you touch it.

The $4,200 in bucket 3 is the number people get wrong most often. It is typically spread across hundreds of terms at $10 to $30 each. It looks like death by a thousand cuts. It is actually the tail that produces next quarter's conversions, and if you block it you will watch conversions fall while your "efficiency" metrics improve.

Across the accounts we audit, bucket 2 usually lands somewhere between 5% and 20% of search spend. Below 5% and the account is already well maintained or very tightly matched. Above 20% and there is usually a broad match keyword or a Performance Max campaign doing something nobody has looked at in months. If you want the mechanics of pulling those terms out, we covered how to find wasted ad spend in detail.

Four corrections that move the number

Before you trust your figure, apply four corrections. Each one tends to shrink the waste number, which is exactly why they get skipped.

1. Conversion lag

A term that converted last Tuesday may not show that conversion for another week. If your window ends today, the most recent two weeks of spend look worse than they are, and you will file live terms into bucket 4.

Check your own lag rather than guessing. In Google Ads you can segment conversions by days to conversion. If most of yours land within two days, end your window three days ago. If you sell something with a three-week consideration cycle, end it a month ago. For a 90-day analysis, a window running from 120 days ago to 14 days ago is usually safer than the last 90 days.

2. Conversions you never imported

This is the big one for lead generation. If phone calls under 60 seconds are not counted, if quote requests that close six weeks later never come back from your CRM, if in-store visits are invisible — then your conversions column is a partial record and every waste figure built on it is inflated.

We have seen accounts where the fix was not a single negative keyword. It was importing offline conversions and discovering that the "dead" terms were producing the business's best-value jobs. Before you calculate waste, confirm you are measuring revenue events, not just the ones that were easy to tag. If you are staring at a zero-conversion account right now, work through the order to check things in before you calculate anything.

3. The search terms report does not show you everything

You can only calculate waste on spend you can see. Google withholds terms that too few people searched, and the withheld portion still costs money. Your export's total cost will be lower than your campaign's total cost, sometimes by a wide margin.

So state your figure honestly: "19% of visible search term spend". If your export only accounts for 70% of campaign cost, say that too. The hidden slice is not automatically waste and it is not automatically fine. It is unknown. We go through what disappears and why in the search terms report and what it hides.

4. Campaign types that report no terms at all

campaign_search_term_view covers Search, Shopping and Performance Max. Display, Video and App report no search terms whatsoever. If a third of your budget is in Display, that third cannot appear in this calculation at all. Excluding it silently makes your waste percentage look better than reality.

Performance Max is a special case: it reports terms, but at a coarser grain than Search, and it takes campaign-level negative keywords only — up to 10,000 of them, with no shared lists. The practical consequences are covered in where Performance Max wasted spend hides and the 10,000 negative cap. Shopping behaves differently again; if your budget is mostly Shopping, read how Shopping negatives work before you act on your list.

What the honest number is actually worth

Blocking $1,800 of quarterly waste does not put $1,800 in your pocket. What happens next depends entirely on whether your campaigns are limited by budget.

If your campaigns are budget-limited — you hit your daily cap most days — blocking waste does not reduce spend. The same budget goes out, just to different auctions. In the example above, that $1,800 over 90 days ($600 a month) gets reallocated to terms performing at roughly the account's $100 CPA, so you would expect around six extra conversions a month. That is the real return, and it is a conversion gain, not a saving.

If your campaigns are not budget-limited — spend is set by how many qualifying auctions exist — blocking waste genuinely reduces spend. You save $600 a month and your conversion count stays flat. Your CPA falls.

Most accounts are somewhere between the two, and the honest thing to say is that you will get some of each. What you should not do is promise a client $600 a month back and then show them a flat spend chart.

There is a cost on the other side too. Every negative keyword carries a small chance of blocking something that would have converted. Broad match negatives carry a much larger one, because they behave unpredictably across match types — a broad negative on "free" can silently kill "free delivery over $50". Phrase and exact negatives are the safer default, and negative match types genuinely do not work like positives. A list that is too aggressive also gets unmanageable; we put some numbers on that in how many negative keywords are too many.

So the complete honest statement is something like: 19% of visible search spend is waste. 9% is safe to remove now and should return around six conversions a month at current CPA. The other 10% sits across eight terms that need individual review. Roughly 30% of campaign spend does not appear in this analysis at all.

That is a less impressive sentence than "you are wasting 40% of your budget". It is also one you can defend in a meeting.

Recalculating it every month without rebuilding the spreadsheet

The waste figure is not a one-time diagnosis. It regenerates. Broad match rewrites itself, Performance Max finds new audiences, seasonality drags in new phrasing, and a term you have never seen before spends $80 in a week.

The practical rhythm is: run the full four-bucket calculation once a quarter, and skim new search terms far more often than that. Anything from weekly to monthly works depending on spend — how often you should check search terms goes through the thresholds. Fold the waste figure into a wider review using the one-hour audit checklist so you are not looking at search terms in isolation.

If you want to know why the junk keeps arriving in the first place rather than just clearing it, why Google Ads shows irrelevant search terms explains the matching behaviour behind it.

This is the job PPCleaner does nightly. It reads the search terms your account paid for across Search, Shopping and Performance Max, judges intent term by term, and adds negatives for the bucket-two cases — the salary searches, the DIY guides, the wrong products. It never blocks a term that converted in the last 90 days, and every block can be undone in one click for 90 days. It does not touch budgets, bids, ads or keywords.

What it will not do is fix the parts of this article that are not keyword problems. It cannot import your offline conversions. It cannot see the search terms Google withholds. It cannot tell you whether "emergency plumber" failed because of your landing page. Bucket four is still yours. What it removes is the part that never needed a human in the first place — and, usefully, it keeps your waste number honest month to month, because a figure that gets recalculated every night has nowhere to hide.