AI startup growth marketing, the first 90 days, when the category is still moving.

The one problem 2022 playbooks do not cover

every startup growth playbook assumes the customer can name what they want. an AI startup usually cannot count on that. nobody searches “AI study mentor”. they search “how do i finish my homework faster”. nobody searches “AI companion”. they search for the feeling.

that single fact rewrites the method. search intent barely exists yet, so the channels that work are the ones where you can put the product in front of someone who was not looking for it: short form video, Reddit threads where the problem is already being discussed, and the founder’s own face. the creative has to teach before it sells. the landing page has to show the thing working, not list features.

at ZuAI the entire breakthrough was a thumbnail that said “under 18”. kids under 18 want content made for people under 18. that video did 2.1 million views and 120,000 users, and it is still the best performing ad in that account three years later. it worked because it taught the audience what the product was for in the first second.

The model is not the moat

a competitor with the same model can ship next week. often the same weekend. that is not a reason to panic; it is a reason to stop treating the model as the product.

what is defensible: distribution and retention. the channel you own, the creative engine that keeps producing winners, the community that recommends you, and the onboarding that gives a person a win before it asks for anything. an AI startup with an average model and a working distribution loop beats a startup with a better model and no users every time. i have watched it happen from both sides.

this is why AI startup growth marketing is a weekly loop, not a quarterly plan. the category moves monthly. the creative that worked in march is dead by may. analyze, execute, iterate, repeat, every week, until the number moves. the AI app growth marketer page lays out the whole loop.

Retention decides whether any of this is worth doing

before channels, before creative, before budget: does anyone come back?

a teen investing app i worked on scaled to 10,000 users on good creative and a good funnel. day 7 retention was 5 percent. we had a great product nobody returned to, which is the single worst position a consumer app can be in, because everything looks fine until it does not. we rebuilt the product into ZuAI, and only then did growth work.

so the honest first step of AI startup growth marketing is a retention read. under 20 percent day 7, spend on the product, not on acquisition. over 20 percent, you have something worth pouring users into, and the free calculators include one built to tell you which side of that line you are on.

The 90 day plan i actually run

days 1 to 30: production. the creative batch, the landing pages, the posting cadence, the subreddit map. nothing is live yet and nothing is guessed. this is the month founders want to skip and it is the month that decides the next two.

days 31 to 60: ship and baseline. everything goes live. organic first, because it is free research: what wins there becomes the paid ad. then paid runs until we know your real CAC, not the one in the deck. at ZuAI, 30 videos a month found two winners; the rest died on schedule.

days 61 to 90: drive the number down. one job from here. every sunday the new batch ships. every monday the winners move into the scaling campaign and everything else dies. ugly ads beat polished ads, and i have three years of spend saying so.

every friday, in writing: what we tested, what won, what died, what changes next week. raw numbers from the account.

Who this is for, and who it is not

consumer AI apps, post launch, with a product people come back to and a budget of about $5,000 a month or more for acquisition. US founders first, Europe and India too, all remote, inside your accounts.

not B2B AI. not enterprise SaaS with a six month sales cycle. not pre-product. those are real disciplines and they are not mine, and i will say so on the free breakdown call rather than take the engagement. the services page has the four engines and the honest fit in full.