CO2e↓CO2 Media

Client Work · Circular-fashion marketplace · Week one · Aug 2026

One number: £23,400 removed

Callum Hodge · 13 Aug 2026 · figures representative, brand anonymized

Before growing anything, stop paying for nothing. A circular-fashion marketplace spending £14,000 a month arrived with a search-terms report nobody had opened since launch. Week one of the engagement removed £23,400 of annualized waste: queries that could never produce a buyer or a seller, priced and fenced off. This write-up is the anatomy of that single week, and the argument for why waste is where every engagement should start.

£23,400

annualized waste removed

13.9%

share of spend it represented

5 days

from access to fenced

The account we walked into

A two-sided marketplace, buyers and sellers of second-hand fashion, eighteen months old and growing fast enough that nobody had looked backward. Paid search had been set up competently at launch: sound structure, sensible keywords, broad match given room to explore. Then it ran unattended while the team shipped product. Broad match kept exploring, as it does, and by the time we opened the account its search-terms report was a museum of queries no one had ever reviewed. The setup was never wrong; it was abandoned, which in paid search amounts to the same thing at a delay.

Where does £23,400 hide in a £14,000 monthly budget?

In the long tail, in plain sight. Reading 90 days of search terms line by line produced four waste families, each priced from its own clicks and the account's conversion data:

Waste familyExample queriesAnnualized cost
Donation seekers"donate old clothes", "clothes bank near me"£8,900
Repair & alteration"mend jumper hole", "tailor prices"£5,700
Fast-fashion bargains"cheap dresses under £10"£5,200
Jobs & press"sustainable fashion jobs", journalist queries£3,600

None of these people will ever list or buy a garment on a resale marketplace this quarter. Every click was real money: 13.9% of total spend, in the same territory as the 8.51% of paid traffic that industry measurement flagged as outright invalid in 2025, a leak the industry prices at $63 billion a year. The difference is that invalid traffic needs platforms and fraud vendors to fight it; this waste needed someone to read a report. Waste you can fence off yourself is the cheapest performance you will ever buy.

What did the five days look like?

  • Days 1–2: full 90-day search-term read; every wasted query tagged to a family and priced from its own click and conversion data
  • Day 3: 176 negatives built into four themed shared lists, phrase and exact match, plurals and misspellings covered, because negatives do not expand the way positive matching does
  • Day 4: launch checklist QA, the same one every Launch Sprint runs; two tracking gaps flagged for the following week
  • Day 5: lists live, pacing caps set, the £23,400 figure written into the first report with its family-by-family breakdown

The pricing discipline is the part worth stealing. "We have irrelevant queries" wins no budget and changes no behavior; "donation seekers cost you £8,900 a year" ends the meeting. Every failed query family carried its own annualized figure, which is also why the marketplace's finance lead, not its marketer, became the engagement's loudest internal advocate by Friday.

What happened to the freed budget?

Nothing dramatic, which is the honest answer week one deserves. The freed spend stayed in the account and flowed to the queries that already converted; sign-up volume rose 9% in the following month with budget flat. The growth work, feed and Shopping structure of the kind described in the refill write-up, started in week three on cleaner data. Judged against the 2026 cross-industry medians ($5.42 CPC, 8.18% CVR), the account went from paying above-median prices for below-median outcomes to a fair fight, and the negatives library it now owns is the cheapest asset in its marketing stack.

Why lead a write-up with waste instead of growth?

Because for a sustainability marketplace the two are the same claim. A brand built on keeping garments out of landfill should not be paying Google for clicks from people looking for a clothes bank; the waste column and the carbon column both exist to make the invisible cost visible, then remove it. Every fenced-off impression is also grams that never get emitted, which made this week the account's largest single emissions cut as well as its largest single saving. Week one is also the cheapest week we will ever bill: £850 as a fixed Launch Sprint, against £23,400 of yearly leak found in this account.