In the room before launch.
Most agencies inherit a growth problem. We were brought in before there was anything to grow: pre-launch, pre-CMO, pre-funnel. That changes what an agency is for. Instead of forcing traffic through a journey someone else built, we shaped the acquisition strategy, the measurement architecture and the quote journey alongside the product itself.
The first months bought the unglamorous assets every direct engine needs and no dashboard shows: tracking that could prove what a direct priced quote really costs, against every other route in, a quote journey designed for how commercial buyers actually behave, and an acquisition model with numbers set from scratch.
The pivot that saved the budget.
The obvious playbook in insurance is bottom-funnel paid search. We advised against it, those auctions are priced by comparison sites with twenty years of head start, and a direct challenger bidding there is paying a premium to fight where the incumbents are strongest. But conviction isn’t evidence, so we tested it properly, with kill criteria agreed before a pound moved.
The auction confirmed it: CPCs no direct unit-economics case could carry. So it died on schedule, and the budget moved to where direct acquisition has the structural edge, social-first demand creation, reaching tradespeople and SME owners weeks before any comparison moment, with creative no aggregator can make about a brand it doesn’t own.
That’s the discipline this whole engagement runs on. Every hypothesis tested, every loser killed without ceremony, every winner scaled. The strategy wasn’t right because we said so. It was right because the market voted.
Creative that doesn’t look like ads.
Nobody scrolls TikTok hoping to hear about professional indemnity. The creative had to earn every second: real trade voices, vox-pop and hero content built with a London creative partner, UGC from creator partners, and a constant flow of stills and native assets from our in-house design team.
Underneath the volume sat the map: four commercial lines, each with distinct personas, pains and moments, Tradesman Insurance as the spearhead. The sole trader worried about a claim ending his business needs a different story from the SME owner comparing renewal terms, and each got their own angles, tested at volume across Meta, TikTok and YouTube, with X and Reddit run as structured experiments.
Compliance at the speed of testing.
Insurance creative dies in sign-off. Every claim needs compliance approval inside a global group’s governance, and the default outcome is a testing programme that moves at the speed of legal review.
We refused the trade-off and rebuilt the workflow with AI in the loop: claims frameworks agreed up front, AI-assisted pre-checks catching issues before they reach compliance, production tooling that turns an approved concept into channel-ready variants in hours. The time from campaign inception to campaign live collapsed, which in a testing-led engine isn’t admin efficiency. It’s learning speed, and learning speed is the whole competitive advantage.
A funnel that got better while it got bigger.
The quote journey didn’t exist when we started, so we treated it like creative: hypothesise, test, fix, again. Quote completion climbed from 59% to 85%. Quotable quotes, the share of starts the business can actually price, rose from 39% to 69%.
That second number is the commercial one. Traffic is easy to buy; quotable risk is the proof you’re attracting the right people, not the most people. And it wasn’t a one-line fluke: across all four lines, completion moved from 60% to 81% and quotable rate from 36% to 65%. The model repeats.
Meanwhile the engine’s costs fell as its volume grew: 1,135 quote starts through direct social, cost per quote start halved from £35 to £17, cost per priced quote down 72%, from £90 to £25.
