Every finding below follows the same shape: what the job should have been, what it was actually doing instead, and what closing that gap was worth. These are real findings from live commercial work — not illustrative examples.
A Performance Max campaign went live without brand-term exclusions during an agency handover. Broad-match, all-inventory bidding began capturing branded search traffic that was previously converting organically at near-zero marginal cost — and re-buying it at auction price.
Organic revenue ran roughly 40% down year-on-year for five consecutive months following launch, with no corresponding drop in branded search volume — the traffic didn't disappear, it moved channels. Timing aligned precisely with the campaign launch date, not any seasonal or ranking factor.
Brand-term negative keyword lists applied at account level. Branded and non-branded performance reporting separated so cannibalisation stays visible going forward. Incremental ROAS re-baselined once brand traffic was excluded.
The free-shipping threshold was raised from £15 to £25 as a margin-protection measure. The change went live untested and coincided with a measurable site-wide conversion rate decline, concentrated in first-time visitors and lower-AOV sessions.
The CVR decline began on the exact date of the change. It was disproportionately concentrated in sessions with cart values between £15-£25 — precisely the segment now falling short of free shipping. Cart abandonment at the shipping step rose in the same window.
Reintroduced a £15-£20 threshold, with a tiered incentive tested (free shipping at £20, gift-with-purchase at £30) rather than a blanket increase — and a rule that any future threshold change gets A/B tested before full rollout.
On revenue alone, email looked healthy — around £8,000/month gross. Once VAT, cost of goods, platform fees, and dedicated people time were properly allocated, the channel was actually losing ≈£1,000/month. The ESP plan was sized for ~500,000 subscribers, while regular sending only reached ~80,000, with no flow strategy to engage or monetise the rest.
Full cost allocation (VAT, 25% COGS, £4,000 platform cost, £2,000 people cost) against £8,000 gross revenue produced a net loss — invisible on a standard revenue-only report.
ESP plan right-sized to the actively-engaged base — saved £1,500/month immediately. Campaign build automated, reducing hands-on time to roughly 20% of a role. Rather than a single-role cut, capacity was mapped holistically across the team, and the reduced workload absorbed into an existing role.
A senior marketing role became vacant. The default move — re-hire at full cost — was treated as a decision to make deliberately rather than automatically, once it was clear parts of the role's job could be absorbed elsewhere or automated.
A capacity review across the wider team identified where the departing role's responsibilities could be redistributed without overloading any one person, and where AI-supported tooling could take on the more repeatable parts of the job.
The role was not re-bought at full headcount cost. Responsibilities were redistributed and upskilled into existing roles, with tooling covering the repeatable work — no redundancy, no backfill.
A new retail site was under consideration — a six-figure annual commitment. The default approach in most growing retailers is to trust instinct and comparable footfall data from the agent; that's not enough to justify a job as large as Scale.
The same store-economics model used on a previous successful site opening was applied here: rent against realistic (not agent-quoted) footfall, category-adjusted unit economics, and a contribution-margin breakeven timeline.
Three scenarios modelled — conservative, expected, and optimistic footfall — before signing, rather than after. The commitment only proceeds once the model shows it can actually earn the Scale job, not just fill a gap on the high street.
15 minutes. No pitch, just a look.
Book a free chat →