The 90-day startup marketing plan, built for one person

This is a 90-day marketing plan for a founder who is the entire team: one measurable goal, twelve workstreams, and about ten hours a week, sequenced so each phase builds the thing the next phase needs. Days 1–30 build a base, days 31–60 bank authority and proof, days 61–90 spend both on a launch that lands.

Every other 90-day marketing plan I read while writing this one was built for somebody else. Some are handover documents for a marketing director joining a funded company — useless when there is no team to get to know. Others are phase-shaped essays that say "build your content engine" in month two and never explain how. This is the capstone of a month of writing on this blog, so it does the opposite: every task below links to the complete playbook for that task. It's a plan you can execute, not a plan you have to research.

What should a 90-day marketing plan include?

A 90-day marketing plan should include one measurable goal, a short list of channels you'll genuinely work, a week-by-week schedule of named tasks, an honest weekly hour budget, and a review point at the end of each 30-day phase. That's the whole specification. Everything else is decoration.

The hour budget is the part almost every template skips, and it's the one that decides whether your quarter survives contact with reality. Corporate plans budget dollars because dollars are the scarce input at a company with staff. You are not that company. Your scarce input is the evening after the code is written. So the unit of planning here is hours per week, and the honest number is eight to twelve — one working day, plus change — rising to about twenty in launch week. If the plan needs more than that, the plan is wrong.

Why the plan runs foundations → authority → launch

The order is not arbitrary — it's dictated by lead times. Some marketing assets pay out in days, some take months, and a launch is worth what the base underneath it is worth. Put them in the wrong order and each one arrives too early to help.

Content has the longest lead time, so it starts on day one and never stops. Reviews, testimonials, and backlinks have a medium lead time and depend on having something for people to look at, so they occupy the middle. A launch is instant and non-repeatable: it converts whatever credibility already exists into attention, which means it has to come last. Founders reverse this constantly — they launch in week two, get their forty-eight hours, and discover that a spike landing on an empty page leaves nothing behind. The full argument for sequencing this way is in the GTM strategy for solo founders pillar; this post is the execution schedule for it.

There's also a 2026 reason the base matters more than it used to. AI referrals to the top 1,000 websites reached 1.13 billion visits in a single month, up 357% year over year with ChatGPT driving more than 80% of them (Similarweb, reported by TechCrunch) — yet those engines still send publishers under 1% of their referral traffic even when they cite them. Increasingly the prize is being named in the answer, not clicked in a list. Answer engines don't cite launch-day excitement. They cite pages, reviews, community discussion, and attributed articles: exactly the assets phases one and two produce.

The whole 90 days on one screen

Here is the plan. Twelve workstreams, three phases, and the hours each one costs in a week when it's the active push. A solid bar means it's your main work that phase; a faded bar means it's ticking over in maintenance at roughly half the hours. Every workstream links to the full playbook behind it.

Read it column by column and the quarter plans itself. Phase one is five workstreams and about ten hours a week. Phase two is the heaviest at roughly twelve. Phase three thins back out to ten until launch week, which eats about twenty and then hands you the rest of the month to convert what it produced.

Days 1–30: build the base

The goal of the first thirty days is to have something worth pointing people at. Nothing in this phase produces users, and that's expected — you're building the surfaces every later phase spends.

Week 1 — one sentence and one page. Write the sentence that says who your product is for and what it replaces, then put it at the top of your site. Create your public project page in the apps directory so your product has a second home that accumulates authority while you work. Budget two hours; the sentence will take longer than you think and it silently determines whether everything after it lands.

Weeks 2–3 — publish and show up. Start the content base with three articles that answer questions your buyers actually type, each leading with the answer in the first paragraph. In parallel, pick exactly one community and be useful in it — the indie hacker communities map covers which rooms suit which job. One room, worked properly, beats five you lurk in.

Week 4 — feedback before features. Run two rounds of structured pre-launch feedback with people who ship — not friends, who will tell you it looks great. Then submit to a first handful of launch directories to seed the earliest links. End of phase one, you should be able to say: my positioning is written, three pages are live, I'm known in one room, and I've heard two hard truths about my product.

Days 31–60: bank authority and proof

The middle thirty days exist to produce the assets that take weeks to mature, so they're ready when the launch needs them. This is the phase founders skip, and skipping it is why so many launches convert nothing.

Weeks 5–6 — proof. Go after your first honest reviews and two or three publishable testimonials. Both do double duty now: they're social proof on your page and they're citation fodder for answer engines. 5W's research found that Wikipedia and Reddit alone drive over 25% of US ChatGPT citations — community discussion and user-generated opinion outrank most newsrooms in what these engines quote.

Weeks 7–8 — authority. Now the slow assets: editorial backlinks, a build-in-public cadence you can sustain, and the first pitches for podcast appearances and newsletter swaps. Pitch in week seven and the episodes land in month three, right when you want them. This is also the phase where the favor queue earns its keep — reviews, feedback, and featured articles are exactly the things other founders can give you and you can give back.

Days 61–90: launch and amplify

The last thirty days spend everything the first sixty built. You are not launching a product to strangers; you're launching a product that already has pages, reviews, and people who know your name.

Weeks 9–10 — set the date and assemble the crowd. Put your date on the launch calendar and work backward through the launch-day checklist. The single highest-leverage thing you can do in these two weeks is line up the humans who'll show up — a cold launch and a supported launch look identical on the calendar and nothing alike at 9am.

Week 11 — launch. Run the day itself against the Product Hunt playbook, then sequence your other directories over the following days rather than firing them all at once. Twenty hours, one day of nerves.

Week 12 — convert and consolidate. The spike fades in forty-eight hours; the work of turning attention into users takes the rest of the month, which is the whole subject of getting your first 100 users. Email everyone who signed up, ask the new users for reviews while they're warm, and fold what you learned back into the content base. Then the next quarter starts from a higher floor — which is the entire point of building in this order.

The engine running through all 90 days

One thing appears in every phase of this plan and in almost no other marketing plan you'll find: reciprocity. Look back at the Gantt and notice how many rows depend on another human doing something for you. A review needs a reviewer. A testimonial needs someone who used the product. A backlink needs an editor. A podcast needs a host. A supported launch needs a crowd.

That's the quiet dependency inside every 90-day plan, and the generic templates wave at it — "reach out to your network" — as if the network were a given. If you had one, you wouldn't need the plan. The workable version is to earn access by giving first: other founders launching this quarter need exactly what you need, and each of them arrives with a real account, a real opinion, and a real website. Trading with them is the one channel that produces distribution on day one with no audience and no budget. I've made the full case for that in reciprocity marketing for founders, and it's why Favors.dev exists: a founder marketing co-op where a points economy enforces the trade, so you can't spend help you haven't earned and the givers never end up carrying the takers.

There's a timing argument for doing this now rather than next year. In GoodFirms' early-2026 survey of marketing and SEO practitioners, 65% named adapting to AI-driven search as their top challenge while only 14% track AI citations at all. The signals that get a brand cited — reviews, community mentions, attributed articles, links — are the ones a founder can earn by trading, and most of the market isn't even measuring them yet. That gap won't stay open for four quarters.

What this plan deliberately leaves out

A plan is defined as much by what it refuses as by what it schedules. Ninety days at ten hours a week is roughly 120 hours — enough to do five things properly or fifteen things badly. Here's what didn't make the cut, and why:

The paid-ads omission is the one founders push back on hardest, so the honest exceptions are laid out in why paid ads rarely work for bootstrappers. And if you want to see the full channel menu with the reasoning for which ones earned a row in the Gantt, that's the indie hacker channel scorecard. The rest of the zero-budget reasoning lives in SaaS marketing on no budget.

How to tell it's working (and what to do if it isn't)

Judge each phase by whether its assets exist, not by revenue. Revenue is a lagging indicator with a lag longer than this plan. At the end of phase one, count published pages, community conversations, and feedback rounds. At the end of phase two, count reviews, testimonials, and referring domains. At the end of phase three, count signups, activated users, and how many of them came from somewhere you can name.

If a phase produced nothing, don't start the next one — the sequencing is the whole mechanism, and a launch on an empty base is the most expensive way to learn that. And if every asset shipped and the users still didn't come, the problem is upstream of marketing. That's a positioning or product answer, and no amount of channel work will paper over it. Watching where your reputation actually accumulates helps here too: the leaderboard and your founder profile make the compounding visible in a way a spreadsheet doesn't.

Frequently asked questions