Reach people while they are still deciding.

It fits best where people plan visually before they buy: home and interiors, weddings and events, food, beauty, fashion, garden, renovation, travel and gifting. It fits worst where demand is instant and unplanned, or where the product is impossible to show. The reach exists either way — ads reached 340 million users in January 2025 — but reach is not fit. We look at your categories, margins and asset inventory first, and if the honest answer is that your money belongs in another channel this quarter, that is what we will tell you.

Intent and timing. Elsewhere you interrupt someone mid-scroll; here they are already assembling a project, often weeks or months before they buy, and they save what they like for later. That changes three things: the artwork has to be worth saving rather than merely stopping the thumb, the measurement window has to be longer than a week, and the channel behaves more like demand creation than harvesting. Run it beside paid search and the search terms it creates show up later in the funnel — which is exactly why we report both together.

The floor is set by learning, not by ambition: one campaign needs enough conversions each week to optimise, so for most brands that means a few thousand a month before production. We charge a monthly management fee scoped to the work — strategy, build, production and reporting — rather than a percentage of your media, because a percentage rewards us for spending your money. Creative production is quoted per set so you can approve a number. If the money cannot support a fair test, we will tell you before you sign anything.

Plan on eight to twelve weeks for a fair read, and expect the first two to look quiet. These shoppers are early, so conversions arrive after a longer consideration window than performance channels train you to expect, and judging week two on last-click data is how promising accounts get killed. What you should see quickly is early signal — which concepts get saved and clicked — and clean measurement. We report both from week one, with the caveats attached, so nobody has to pretend early numbers mean more than they do.

It changes the work rather than removing it. Automation decides bids, placements and sometimes asset combinations well, and Digiday reported a 9.1% ROAS increase and 8.4% CPA decrease for advertisers in the Performance+ creative alpha. What it cannot do is decide which products deserve spend, write a concept worth saving, keep a product feed honest, or design a test that proves incrementality. Our job moved upstream to exactly those things — and to checking that automation is optimising toward the outcome you actually sell.

Yes, and often the best arrangement is that they own the brand and we own the adaptation. We write the briefs, specify the formats this feed needs, and produce the variants that testing calls for, all inside their guidelines. Where you would rather we handled production end to end, our ad creative and product photography teams shoot to the same shot list, so the media plan and the assets arrive together instead of waiting on each other.

With three views and one test. The platform view shows in-flight signal; your own revenue data shows what actually closed; blended efficiency shows what the business earned per pound spent across everything. When a spend decision is big enough to matter, we settle it with a geographic holdout or a controlled on/off window, which is the only method that answers what would have happened anyway. Every report states the method and its uncertainty, because a range you can reproduce is worth more than a confident single figure.

Usually three things, and they are the same three every time: creative that was resized from another channel instead of built for this one, spend split across too many small ad groups to ever leave the learning stage, and a conversion setup nobody verified before judging the results. We start by proving the tracking, consolidating the structure and producing purpose-made asset sets, then hold the plan for a full consideration window. If it still does not pay after a fair test, we will say so and help you move the money.

You do, in every case. We work inside your own ad account and business manager, your catalogue and tags stay in your name, and the artwork files, briefs and testing history are handed over as we go rather than at the end. The reporting is built on your data sources, so it keeps working whoever runs the media next. We would rather earn the next quarter on results than hold anything hostage — and it keeps us honest about what is really working.

That is how we prefer it. This channel creates demand that other channels harvest, so reading it in isolation flatters or punishes it unfairly. We plan it with your paid social, paid search and organic search work, report everything against the same definitions, and watch branded search and direct traffic move when the flight starts. It is equally fine to hire us for this channel alone while other agencies run the rest — we will share the measurement design with them.