The “My Bar” Pattern Is the Most Underrated Retention Mechanic in Consumer Apps — and Kooler Just Proved It
Most social platforms are built to maximize time-on-screen. The best ones are built to minimize it. That inversion is the entire thesis behind Kooler, a cocktail app relaunched by designer Eduardo Duccigne after a fifteen-year gap, and it’s the reason I’m writing about a drink recipe app on a blog that normally dissects scheduling tools and algorithm shifts. The mechanic at the center of Kooler — inventory matching, not content browsing — is the same mechanic that separates a social account people open daily from one they mute after two weeks. If you run accounts for a brand, a creator business, or a niche community, the design lesson here is worth more than the app itself.
What Kooler actually is, minus the press-release framing
The source material is a Product Hunt launch page, so let me separate what’s sourced from what’s my read. Per the maker’s own post, Kooler is a free iOS app with no ads. It has four stated features: a “My Bar” module where you log what’s in your cabinet and see every drink you can make; a library of 100+ curated recipes spanning classics and Brazilian drinks that most cocktail apps skip, named specifically as Rabo de Galo, Macunaíma, and Caju Amigo; clean recipe pages with consistent measurements and photography; and check-ins that let you rate a drink you had at a bar. Community recipe submissions are described as “next,” meaning they don’t exist yet. Pricing is free with no ads. User counts, retention numbers, funding, and revenue are all not disclosed.
The backstory matters because it explains the design instincts. Duccigne says he and a friend built a first version in 2010 as a small social network for cocktail recipes, shut it down in 2011 when he moved to São Paulo for a job, and rebuilt it fifteen years later — this time as a designer using AI tools, from the database through to the iOS app. That’s a real data point in the ongoing “non-engineers shipping software” story, and I’ll come back to it, because it has direct implications for creators who keep telling themselves they can’t build the tool their audience needs.
The problem Kooler solves is not “finding recipes”
There are roughly a thousand cocktail recipe apps, and almost all of them solve the same problem: discovery. You browse, you save, you screenshot, you never make anything. The failure mode is well documented in the recipe-app category broadly — the same dynamic that made Pinterest a graveyard of saved-but-never-cooked meals. Saving is cheap. Making is expensive. Apps that optimize for saving win engagement metrics and lose to their own archive.
Kooler’s bet is that the real bottleneck is decision fatigue at the moment of use. You’re standing in your kitchen at 8pm. You have gin, sweet vermouth, and a half-empty bottle of Campari. You do not want to browse 100 recipes. You want the app to tell you that you can make a Negroni and that you’re one ingredient away from a Boulevardier. That’s an inventory-first, browse-second interface, and it’s meaningfully different from what Difford’s Guide or the IBA’s official cocktail list offer, both of which are reference libraries rather than decision engines.
Why this matters more than the recipe count
A 100+ recipe library is table stakes in 2025 — Punch, Liquor.com, and every YouTube cocktail channel have effectively infinite recipes. The count is not the moat. The moat is the matching logic: given a constrained set of inputs, what’s the highest-value output? That’s the same question a social media manager answers every Monday morning when they look at their content calendar and ask “what can I actually ship this week with the assets, time, and team I have?”
The Product Hunt comment thread surfaces the interesting edge case before the maker does. Gal Dayan, who ships a product called Dial, asks the sharpest question on the page: when a recipe needs something close to what’s in your bar but not exact — a different vermouth, gin instead of a specific bitters — does Kooler surface a “one swap away” match, or is matching strict on exact ingredients? His framing is the whole ballgame: “that’s the difference between ‘here’s what’s technically in your bar’ and ‘here’s what you’re actually drinking tonight.’” The maker does not answer that question in the thread. So as of this writing, the matching tolerance is not disclosed. That’s the single most important open question about the product, and I’d bet it’s also the hardest engineering problem in the app — fuzzy ingredient substitution is a genuinely difficult matching problem, and it’s where most inventory-based apps quietly give up and go strict.
The “one swap away” pattern is everywhere in creator tooling
Here’s where I connect this to the work you actually do. The “one swap away” logic is the same logic that makes a repurposing tool useful or useless. If I upload a 12-minute YouTube video and the tool tells me “here are 10 clips,” that’s a strict-match library — technically correct, practically overwhelming. If it tells me “this 40-second segment is the one that will work on TikTok because it has a hook in the first two seconds and a clean emotional beat at the end,” that’s a one-swap-away match. Same input, radically different output value. The inventory in Kooler’s case is bottles. The inventory in your case is footage, hooks, formats, and platform-native constraints.
What creators and social teams can borrow from Kooler
1. Constraint-first content planning beats idea-first content planning
Most content calendars start with “what should we post?” Kooler starts with “what do we have?” For a social team, the equivalent is an asset audit: what footage do we already have, what’s already performed, what can be reshot cheaply, what’s evergreen. When I’ve run accounts with a real cadence — 30 posts across 5 platforms in a month, say — the weeks that fell apart were never idea-poor. They were asset-poor in a way the calendar didn’t anticipate. A constraint-first planning pass, done before the ideation session, would have caught that.
The Kooler version of this is “My Bar.” The creator version is a living inventory doc: raw footage folders, approved brand assets, evergreen posts with proven engagement, and a short list of formats the team can execute without a shoot. If your inventory doc is empty, no amount of brainstorming will save the calendar.
2. Curated depth beats encyclopedic breadth
Kooler ships 100+ recipes and explicitly calls out Brazilian drinks “most cocktail apps skip” — Rabo de Galo, Macunaíma, Caju Amigo. That’s a positioning decision, not a content decision. The app is not trying to out-Difford Difford’s. It’s claiming a lane: Brazilian cocktails, done properly, in an app built by a Brazilian designer.
For creators, this is the single most repeatable strategic move available. A LinkedIn account that covers “B2B marketing” competes with everyone. An account that covers “B2B marketing for climate-tech founders” competes with almost no one and gets recommended more aggressively because the engagement signals are cleaner. The algorithm doesn’t reward breadth. It rewards the ability to predict who will engage, and narrow lanes make prediction easier. This is true on YouTube’s recommendation system, on TikTok’s For You feed, and on Threads — the platforms differ in mechanics but converge on the same preference for legible audience signals.
3. Check-ins are a sleeper feature — and a warning
Kooler’s check-in feature lets you rate a drink you had at a bar. On the surface it’s a lightweight social hook. Structurally, it’s a data collection play: it tells the app what people actually drink, not what they say they want to drink. That’s the same distinction between stated preference and revealed preference that separates useful analytics from vanity dashboards.
Here’s the warning, and it’s my take rather than a sourced claim: check-ins are also the feature most likely to die from cold-start. A check-in feature with no friends on the platform is a diary. A diary is fine, but it doesn’t compound. If Kooler’s next move is community recipe submissions, the check-in data becomes the ranking signal for which submissions surface — which is smart — but only if enough people check in to make the signal meaningful. That’s a chicken-and-egg problem the maker hasn’t publicly addressed.
Why TikTok creators should care more about this than LinkedIn ones
If you’re a TikTok creator, your entire distribution model already runs on the “one swap away” principle. The For You feed doesn’t show your video to your followers first; it shows it to a small test cohort and decides based on watch time, rewatches, shares, and completion rate whether to expand. That’s a matching engine with fuzzy tolerance — it doesn’t require an exact audience match, just a plausible one. Kooler’s inventory matching is the same shape of problem at a much smaller scale.
If you’re a LinkedIn creator, your distribution is closer to a strict-match library: your existing network sees your post first, and expansion depends on engagement within that network before the algorithm pushes it wider. The “one swap away” logic applies less directly, but the constraint-first planning lesson applies more, because LinkedIn rewards consistency and specificity in a way that punishes scattershot posting.
Where I think Kooler falls short — or at least where the questions are
I want to be careful here, because the source is a launch page and the maker is a solo designer who rebuilt a dead product after fifteen years. That’s an admirable story and it deserves a fair read. But fairness includes naming the gaps.
Matching tolerance is undisclosed. As covered above, the “one swap away” question went unanswered in the thread. If matching is strict, the app is a lookup table. If it’s fuzzy, it’s a decision engine. Those are different products with different retention curves. I’d bet the answer is somewhere in between — exact match on base spirits, loose match on modifiers — but that’s a guess, not a sourced fact.
Community submissions are “next,” not shipped. User-generated recipe content is the hardest moderation problem in any food app. Alcohol recipes add a layer: you’re not just moderating spam, you’re moderating potentially harmful consumption patterns. The maker hasn’t said how submissions will be reviewed, who owns submitted recipes, or what happens when a submission is wrong. Not disclosed.
No Android, no web. The app is iOS-only per the maker’s post. For a creator or social operator thinking about building an audience tool, that’s a meaningful limitation — roughly half your potential audience is on Android, and a web version is table stakes for shareability. A recipe you can’t link to is a recipe that doesn’t travel.
No social graph, no sharing mechanics described. The original 2010 version was “a little social network.” The 2025 version reads as a personal utility with a check-in feature bolted on. Whether it’s actually social — can you see friends’ bars, follow their check-ins, share a recipe to Instagram stories — is not stated. If it’s not social, it’s competing with a notes app, and notes apps don’t need a Product Hunt launch.
The AI-built angle is a story, not a moat. Duccigne built the database and iOS app with AI tools as a non-developer. That’s genuinely impressive and it’s a real signal about where solo product development is heading. But it also means the barrier to entry for a competitor is now roughly one motivated designer and a weekend. The moat, if there is one, is curation taste and the Brazilian cocktail lane — not the code.
Where the math breaks for creators thinking of building something similar
If you’re a creator reading this and thinking “I could build the Kooler for my niche,” run the math first. The app is free with no ads. There’s no disclosed monetization. A free app with no ads and no subscription needs either massive scale (which requires marketing spend) or a paid tier that doesn’t exist yet. For a creator with an audience, the better play is usually the inverse: build the tool as a lead magnet for your existing audience, not as a standalone product. Your audience is the distribution. The app is the retention hook. That’s a fundamentally different business than what Kooler is attempting, and it’s the one I’d recommend to most creators reading this.
What I’d watch / test next
Three concrete things to do this week, whether or not you ever download Kooler.
First, run a constraint-first audit on your own content inventory. Before your next planning session, list what you actually have: raw footage, approved assets, evergreen posts with proven engagement, formats you can execute without a shoot. Then plan the week from that list instead of from a blank page. My prediction: you’ll ship more and stress less, and you’ll notice which gaps are structural rather than situational.
Second, test the “one swap away” question on your own repurposing workflow. Take one long-form asset and ask your tool — or yourself — not “what clips can I make?” but “what’s the single closest-to-publishable clip, and what’s the one edit that would make it work on a specific platform?” If your tool can’t answer that, it’s a library, not a decision engine, and you should treat it accordingly.
Third, watch Kooler’s next move on community submissions. If the maker ships a submission flow with a clear moderation model and a ranking signal from check-ins, that’s a genuinely interesting template for niche community apps — and a case study worth studying for anyone building audience-facing tools. If submissions ship without a moderation model, that’s a different lesson, and not a good one.
The broader takeaway: the most valuable consumer apps of the next few years won’t be the ones that give you more to browse. They’ll be the ones that tell you what you can make right now with what you already have. That’s true for cocktails. It’s true for content. And it’s the design principle I’d bet on over almost any feature roadmap.
