A dessert worth making a plan for — turning a destination café into the thing people save for later and send to a friend, in front of 300–700K people across Meta and TikTok.

Dessert is not a habit, it is a plan. Nobody wanders into a soft-serve counter on the way home from work the way they wander into a coffee queue; somebody decides, usually on behalf of two or three other people, and usually with a day or two of notice. That single fact reorganises the whole marketing problem. The decision does not happen inside one person’s head where an advertisement can reach it — it happens in a group chat, between people, as a proposal that somebody has to make and somebody else has to agree to. A brand can be perfectly well liked and still never be the name that gets proposed, because being admired and being suggested are different behaviours and only one of them fills a room.

That leaves a destination dessert brand with an awkward gap to fund. The occasions it depends on are intermittent, so for most of any given month the audience has no live reason to think about it, and continuous presence in the feed has no obvious return — which is exactly why it tends to get underinvested in, and exactly why the brand goes missing at the moment the plan is being made. Meanwhile the category is easy to imitate and photographs identically everywhere: the pour, the swirl, the waffle, the marble counter. Dessert photography has converged so completely that a beautiful post is no longer evidence of anything, and it earns the same shrug whether the product behind it is exceptional or ordinary. There is also a ceiling that has to be respected rather than argued with. A brand people travel to has a real catchment, and pushing reach indiscriminately across the whole market buys impressions from people who were never going to make the trip. The honest ambition, then, was never the biggest possible number — it was to reach the people who actually plan this kind of outing, and to be worth passing on when they did.

Reach a meaningful share of the audience that genuinely plans dessert outings, and stay in front of them across the quiet stretches between occasions so the brand is already present when a plan starts forming. Reach was the primary measure of that presence. Views were the second, because reach on its own only proves a post was delivered: impressions accrue whether or not anybody registered them, and an awareness budget can always be spent into a large and hollow number. A view is the first piece of evidence that something held a person for a moment instead of sliding past them.

The third measure is the one that set the standard for the whole engagement, and it was chosen precisely because it is difficult. The target was saves and shares above 4% — not engagement in general, but the two specific actions that cost the audience something. A like is instant and free and commits the person to nothing. A save is somebody filing a thing away against a decision they have not made yet, which is the closest a feed gets to a stated intention. A share is somebody putting their own name behind a recommendation to a friend, which is the exact behaviour that precedes a visit in a category where one person proposes and a group agrees. Both are rarer than likes by a wide margin, and both get harder as reach expands into audiences with no relationship to the brand. Holding that floor while reach ran up was therefore a quality control on the reach itself: it ruled out the cheap version of the campaign, the one that buys volume from people who will never respond, and put the burden back on the work to be worth saving.

We took over social media management and changed what the feed was for. Instead of publishing more product photography into a category already saturated with it, we built the channels around the plan rather than the product: what is on this month and when it goes, which flavours are worth a special trip and which are about to disappear, what suits an after-dinner detour versus a birthday versus treating somebody. Each post was written to be self-contained enough to survive being forwarded — a person sending it to a friend should not have to explain what it is, where it is, or why they sent it, because a post that needs a covering message is a post that does not get sent. Video, photography and design were produced on a continuous cycle rather than in occasional shoots, and deliberately composed for the save and the share instead of the double-tap: the pour and the build rather than the finished plate, the first cut through a waffle, the craft and the process that a static flat-lay structurally cannot show. Vertical video carried most of the volume, cut to land inside the first second and to work with the sound off, and the design work was tested at thumbnail size on the assumption that most people would first meet it as a small image in somebody else’s message.

On top of that organic base we ran awareness campaigns across Meta and TikTok, promoting cuts that had already proved organically that they could earn a save, rather than commissioning separate creative built to look like advertising. That is the decision that made the 4% floor survivable once spend scaled, because what went behind budget was material with a demonstrated ability to provoke the expensive kind of response. Budget was weighted towards the catchments that realistically travel for a destination dessert rather than sprayed across the whole market, and it moved continuously: anything that held its save and share rate as reach expanded got more, and anything whose response collapsed the moment it left warm audiences was retired within days. Running alongside, influencer collaborations with local dessert and food creators put the brand into feeds it could not reach as itself. Those were framed as an outing rather than a product review and filmed in each creator’s own voice, because the metric we were chasing is fundamentally somebody recommending something to somebody else — and a creator doing that is the same behaviour operating at scale. The three parts compounded: what creators instinctively reached for told us what to film next, what earned saves organically became the ad, and the ads carried the strongest of it out to people who had never made the trip.

Over five to six months the work reached 300–700K people and delivered 800K–1.5M views, with saves and shares holding above 4%. The combination is the point rather than any single figure: a destination brand put in front of a substantial share of the audience that actually plans these outings, with the response weighted towards the two actions that indicate somebody intends to come back to it or has already recommended it to someone else — the behaviours broad awareness campaigns most often trade away in order to buy the reach in the first place.

See the rest of our client work on the case studies page.