30 Minutes to Fix 2027 Marketing: Inside HubSpot's LOOP Playbook

It all clicked when HubSpot woke up in May 2024 to find 140 million visits gone in a year — an 80% collapse that broke the old math of more visits = more customers. The brutal chart hides a more interesting tail: demand generation rose 20% while traffic fell. The traffic that vanished was noise; the traffic that stayed carried intent. My read is straightforward — AI flipped marketing from a world where visit volume wins to a world where visit value wins. Chasing count is now the nostalgic game; capturing intent is the winning one.

Remember the old funnel: straight, linear, pour visits in the top, get customers out the bottom. That funnel has turned into an hourglass. The top — awareness — ballooned and fragmented. Creators, AI search, an endless stream of new platforms all puff up the top at once, but each platform wants to keep people inside, so earning a click to your site is harder than ever. The bottom, meanwhile, tightened and accelerated: AI made personalization, finding the right buyer and speaking their language, and lifting conversion easier than ever. Getting visits is hard; converting them is easier. The waist of the hourglass thinned, both ends got huge.

Kieran Flanagan and HubSpot CMO Kip Bodnar turned that shock into a playbook: LOOP. First unveiled live on HubSpot's Unbound stage, the book is titled Loop: Outlearn. Outmarket. Outgrow. — hundreds of hours of work distilled from 897 long-form interviews with sharp operators. Four steps you can run for any tactic or campaign: 1) Express your taste and judgment to AI, 2) Tailor content to where your audience actually lives, 3) Amplify the message across the modern buying channels, and 4) Evolve by feeding real-time lessons into the next loop. Run one loop at a time and let results compound like interest.

The first loop, Express, is my favorite because everyone now has the same tools, same prompts, same models, and the output is a sea of sameness. For a hundred years the hardest part of marketing was making things — time, budget, people to produce. Making is no longer the problem; choosing is. When production cost drops to zero, the only thing with a price is judgment. AI hands you the average of everything it was trained on — where all your competitors already live. You have two options: outsource your taste to AI or instill your taste into AI so it creates something uniquely yours. The book is adamant: never outsource taste; teach it.

What does taste mean? The book nails it with four stories. Red Bull refused to fix a drink that tasted awful because people buy identity, not flavor — a $10 billion company. Liquid Death put water in a beer can and sold it to punk kids embarrassed by their drink — now worth over $1.4 billion. A 50-cent eye patch and the line the man in the Hathaway shirt sold shirts for years from a single image. The inverse is Jaguar: ignoring 70 years of heritage with an award-show ad that showed no car, sales fell 97% in six months. Taste is the ability to sense what will make your customer pause before they ever do; AI can only mimic the average, you teach the live signal.

The tool for teaching is the taste profile. Not a dusty brand guide in a forgotten folder, but a living document updated every time you do marketing, and AI reads it before it creates anything. It rests on three questions: who is your customer and how do you earn trust (how do they speak?), what are your stories — brand, product, why you, why now — and most critically, why would someone stop scrolling for you? Why would they give you a slice of their scarce day? In the book the taste profile has two blocks: customer taste and brand taste. The demo is blunt: a page built without a taste profile is a generic SaaS template you could resell by swapping the logo; a page built with your taste profile is unmistakably you.

The second loop, Tailor, confronts the content deluge. AI-assisted content is up 4x in two years, but no one's attention is up 4x. The second shock is half-life: Titanic played 287 days in theaters, the average film now lasts 5 weeks; a LinkedIn post lives less than 24 hours, half the audience for that lovingly crafted email arrives in the first two hours. Everyone is drowning. The book places this on a simple 2x2: bottom-left dead zone — the slot factory churning volume with low quality; bottom-right invisible excellence — brilliant but rare campaigns; top-right is the goal: remarkable and relentless — great content at scale, which AI finally makes possible. Remarkable means original data, customer stories, and examples you could not get by asking ChatGPT.

The living proof for that matrix is Sabrina Romanov: 3.4 million followers in 28 months as a solo creator, no large team, no millions. Her secret is two-sided. To be remarkable she runs two tests: start long — make a long YouTube video or newsletter then slice it into pieces for every channel, and start short — post 30-second clips or a few-paragraph LinkedIn posts, test 20 to 30 ideas in a week or two and double down on the two or three that truly take off into richer long-form. That is the content strategy for the AI era. To be relentless she takes one strong piece and runs an AI system that rewrites scripts for video and reformats hooks for LinkedIn based on what that audience cares about. The system, not a 24-hour grind, creates relentless scale.

The third loop, Amplify, is distribution's new playbook, and the book gives you one channel you can act on Monday: paid. At the June 2025 NBA Finals, creator PJ Assetero made an AI ad for a platform called Kauchi for a couple thousand dollars that outperformed brand films costing hundreds of thousands, even millions. The lesson: the most expensive part of paid used to be filming the ad; it is not anymore, AI does that for you — the question is whether you are producing enough variations. Same brand, same budget, same platforms: with 10 creatives a month you get a given ROAS; with 20 you lift it by 65%. The book's math is vivid: if 10 creatives return $3 for $1, 20 returns $6, 100 returns $30 — leverage lives in creative volume.

HubSpot backs the creative is the new targeting thesis with its own tests. For years we tuned audiences: titles, interests, lookalikes. Now Meta and Google know your audience better than you do. Manual versus Meta Advantage gave a 22% ROAS improvement; Google Performance Max cut customer acquisition cost by 33%. Your job is not to hand-craft audiences but to give the algorithms enough creative variations around distinct jobs-to-be-done so they can choose who sees what. The new benchmark is 20 creatives a month, with no ceiling. The second lever is optimize for quality, not quantity: optimize for form fills and you get low-intent demand; for qualified leads a bit better; for booked meetings a bit better; but the goal is a closed dollar. You get there by enriching platforms with first-party data and your CRM, telling them what good looks like.

The fourth loop, Evolve, is about learning speed, told through two flight stories. In 1903 Samuel Langley had $1.5 million and the U.S. government, perfecting the perfect plane for years, twice crashing into the Potomac. The Wright brothers had $1,000, no backing, but cheap gliders they flew, fixed, and flew again, iterating fast — nine days after Langley's second crash they flew at Kitty Hawk. That's the power of learning. In marketing, quarterly campaigns are Langley; two- and six-week sprints are the Wright brothers. A two-week sprint gives 26 chances to learn per year versus four. Sprinting used to be hard because pulling data, checking it, building decks was manual and forgettable; AI now automates that admin, and HubSpot's sprint template can be photographed, dropped into your AI tool with recent campaign data, and turned into a first draft plus a next-sprint proposal on the spot.

The talk closes with the mental barrier story of May 6, 1954. For the first half of the century runners believed a sub-four-minute mile was impossible — the body would break. Roger Bannister ran 3:59.4, and within twelve months more than a dozen runners broke the barrier too. LOOP feels daunting the same way, the authors say; the traffic drop left them no choice, but you don't need a crisis to start. The book wraps 897 videos into prompts, templates, and tools to make the first loop tangible: a taste-profile QR to build tonight, a photographed sprint sheet to run next week, and a weekly practice of remarkable and relentless as the new habit you can start on Monday.

Watching this, my biggest takeaway is that marketing advantage in the AI era is not producing more but deciding what not to produce — and turning that decision into a consistent language the model can repeat. HubSpot's 80% loss with 20% growth is reassuring for anyone ready to let vanity metrics go and double down on intent; when energy moves to creative volume and feeding data back to platforms, ROAS actually moves. A reminder, though: LOOP is not a campaign template you frame, it is a habit — without updating the taste profile each loop, increasing variation, and shortening the sprint, the loop stays a nice poster.