Google Ads: 300 exact keywords at $175 CPL were paused; replacements ran at $442–$689
The first thing most people do on inheriting a Google Ads account is look at what's running. The second thing, if they're tidy, is delete what isn't. Both are mistakes, and the second is the expensive one.
Here is the shape of one account.
Count | |
|---|---|
Campaigns that have ever spent money | 507 |
Campaigns enabled today | 16 |
Ratio | 32 to 1 |
Four hundred and ninety-one campaigns sitting paused or removed. To someone doing a tidy-up, that's clutter. To someone doing an audit, it's nine years of experiments with the results attached.
What was in the paused campaigns
We pulled every keyword in every paused campaign with a conversion history, filtered out brand terms, and asked which ones had actually worked.
Count | Spend | Conversions | Cost per lead | |
|---|---|---|---|---|
Non-brand keywords in paused campaigns with ≥1 conversion | 526 | n/a | n/a | n/a |
Of those, exact match with proven performance | 300 | n/a | 1,323 | $175 |
Three hundred exact-match keywords, each with a real conversion record, each cheaper per lead than the account's current non-brand average of $199, none of them serving. They had been paused in a restructure, not because they'd stopped working, but because the campaigns they lived in were being replaced.
Meanwhile, eleven new campaigns had just been built from scratch, on manual bidding, with keywords chosen fresh. Nine days later, those new campaigns were running at $442 to $689 per lead.
Source of keywords | Cost per lead |
|---|---|
The 300 paused exact keywords, historical | $175 |
The 11 new campaigns, first nine days | $442–$689 |
The account had already done the research. Someone paused it and started over.
Where the value actually is
A live campaign tells you what's happening now. A paused campaign with history tells you what happened when someone tried something, and the account had tried nearly everything.
Paused asset | What it contains |
|---|---|
300 converting exact keywords | A ready-made keyword list with attached bids and CPLs |
Dozens of paused Target CPA campaigns | Evidence of what automated bidding did, per category, per state |
44 paused brand ad groups | Every brand term and misspelling the business ever matched |
Landing page test campaigns from 2019 and 2021 | Which pages converted, and which didn't |
Paused match-type variants of the same theme | A direct exact-vs-phrase comparison on identical keywords |
That last one is worth pausing on. Because the account had built separate campaigns per match type for years, it contained a natural experiment: the same keyword theme run as exact and as phrase, side by side, with conversion data on both. Nobody would design that today. It exists in the archive, and it answers a question the current team was about to spend money re-asking.
The 43 that never worked
The archive also contains the failures, which are just as useful.
Count | Spend | |
|---|---|---|
Campaigns that spent money and never recorded a conversion | 43 | $70,557 |
Forty-three ideas, $70,557, zero leads. Each one is a documented dead end: a state, a category, a targeting approach that didn't work. Delete them and the next person will rebuild at least one of them.
The case against deleting anything
Reason to delete | Why it's wrong |
|---|---|
"It's clutter" | Filters exist. Enabled-only view is one click. |
"It's confusing" | The confusion is the naming convention, which deletion doesn't fix |
"Old data isn't relevant" | CPC has doubled since 2018; conversion rate patterns haven't changed much. The relative performance of keywords is remarkably stable |
"We're restructuring anyway" | Which is exactly when you need the history most |
Deletion is irreversible. Google Ads removes performance history for removed keywords from most reports. A paused keyword keeps everything; a removed one is gone.
What to do with an inherited account
Step | Action |
|---|---|
1 | Pull every keyword across every status with lifetime conversions > 0 |
2 | Filter out brand. Sort by cost per lead. |
3 | Everything under your target CPL is a candidate, regardless of what campaign it's in or whether that campaign is paused |
4 | Map each candidate to the live campaign it belongs in by intent and geography |
5 | Move them, exact match, bid set from historical conversion rate × target CPL |
6 | Only then look at what's missing and research new terms |
This takes an afternoon. It replaces weeks of keyword research with the account's own record of what has already converted, and it starts the rebuilt structure at $175 per lead rather than $442.
The general point
Nine years of paid search produces something more valuable than the current campaign set: a record of what a specific audience in a specific market actually clicked and converted on, at a scale no keyword tool can reproduce. Every paused campaign is a page of it.
The live campaigns are a snapshot. The archive is the book. Read it before you rewrite it.