How to ship from idea to revenue in 90 days
How to ship from idea to revenue in 90 days
You ship from idea to revenue in 90 days by splitting the quarter into six two-week sprints: weeks 1-2 are problem interviews and a landing page, 3-4 are the MVP build, 5-6 are a closed beta with 5 to 10 customers, 7-8 are paid conversion and a pricing test, and 9-12 are the onboarding flow that gets you to your first $5,000 of MRR. The plan only works if you skip native mobile, a real settings page, and polished docs until after revenue.
That is the whole playbook. The rest of this post is the version with milestones, the tools we'd actually use, and the four mistakes that turn 90 days into 180.
The setup: where 90-day plans go wrong
Most founders we talk to lose the first 30 days picking a stack. They debate Next.js vs Remix, Supabase vs Postgres on Fly, Stripe vs Paddle, and by the time they write a line of code, they're three weeks behind and the original wedge has gone soft.
The other failure mode is the opposite. You ship in two weeks, but you ship the wrong thing because you skipped the 20 customer conversations that would have killed the feature. Then weeks 3 through 12 are a series of pivots dressed up as iterations.
A real 90-day plan is two things: a sequencing of bets, and a list of what you explicitly will not do. The second list is harder to write than the first.
The 90-day shape (read this before anything else)
| Weeks | Phase | Output | Spend benchmark |
|---|---|---|---|
| 1-2 | Problem interviews + landing page | 25 conversations, signed-up waitlist, problem-statement doc | $0-500 |
| 3-4 | MVP build | Single-flow web app, hosted, 1 happy path, no auth polish | $2k-5k |
| 5-6 | Closed beta | 5-10 design-partner users, daily Loom feedback | $1k-3k |
| 7-8 | Paid conversion + pricing test | 2-3 paying customers, validated price point | $1k-3k |
| 9-12 | Onboarding flow + first $5k MRR | Self-serve signup, activation rate >40%, $5k MRR | $4k-8k |
Total realistic budget: $8k to $20k for the engineering work, plus your own time and roughly $300 to $800 a month in tooling. If you're targeting much higher, you're either over-building or hiring the wrong tier of help.
Weeks 1-2: Problem interviews and a landing page
The first two weeks have nothing to do with code. The goal is 25 conversations with people who have the problem you think you're solving, plus a landing page that captures the language those people used.
Use a Calendly link, a $50 Amazon gift card for non-network interviewees, and a Loom recording per call so you can re-listen. Take notes in Notion or a single Google Doc. After 15 calls you should be able to predict the next answer to every question.
Build the landing page in Framer or Webflow. Hero copy quotes a customer back to them. One email field, no marketing pages, no blog yet. Drive 200 visits with a personal Twitter or LinkedIn post and one targeted subreddit comment. If you can't get 25 emails from 200 targeted visits, the wedge is wrong and you should rework the page before writing code.
What to skip in weeks 1-2: a logo, a brand color system, a custom domain SSL config that takes you a day, any backend, any pricing page. Use a Vercel preview URL if you have to.
This is also when many founders try to find a technical advisor for their startup so they have a second opinion on the build decisions coming up. One hour a week from someone who has shipped is worth more than a full-time co-founder hunt at this stage.
Weeks 3-4: The MVP build
Two weeks. One happy path. No edge cases.
Pick a stack you (or your engineer) can ship in without thinking. Our default for a typical SaaS MVP:
- Next.js on Vercel for the app
- Supabase for Postgres, auth, and storage
- Clerk if Supabase auth feels awkward for your flow
- Stripe Checkout (not Stripe Elements) for billing later
- Resend for transactional email
- PostHog for product analytics from day 1
That stack is boring on purpose. Every part has good Cursor and Claude Code context built up, which matters because you're going to be vibe-coding at high speed.
The MVP must do one thing end to end. If you're building a meeting-notes tool, the one thing is: paste a transcript, get a structured summary, save it. Not: workspace management, billing portal, integrations with Slack and Notion and Linear. Those are weeks 12+, if at all.
This is the most common place to bring in outside engineering help. A mid-tier engineer at $1,000/week for two weeks shipping a focused MVP is the cleanest math we know. Senior at $1,500/week is overkill unless the wedge is genuinely architectural (real-time sync, multi-tenant from day 1, on-device ML). Junior at $500/week works if you're technical enough to spec tickets in detail.
What to skip in weeks 3-4: a settings page, multi-org support, password reset flows beyond what Clerk gives you for free, dark mode, a marketing site beyond the landing page, custom email templates, any native mobile work.
Weeks 5-6: Closed beta with 5 to 10 customers
You email the 25 people who signed up in weeks 1-2 and offer 5 to 10 of them a free 30-day beta in exchange for a 30-minute call each week and the right to record Loom feedback. Most will say yes. The ones who say no are telling you the problem wasn't urgent.
Watch usage in PostHog daily. The pattern you want: 3 to 5 of the 10 use the product 3 or more times in the first week. If nobody comes back, the MVP is wrong, not the marketing. If 8 out of 10 come back, you've found something and the rest of the 90-day plan is mostly execution.
Ship one improvement per day based on a specific user's feedback, and tell that user when their fix ships. This is the strongest retention loop you'll ever run. After two weeks of this, three of those beta users will be ready to pay.
Some founders also use this phase to start sketching what their first 5 customers should look like at scale: the persona, the trigger event, the average contract value you'll target in month 4.
Weeks 7-8: Paid conversion and a pricing test
In week 7 you switch the beta from free to paid. Stripe Checkout, one plan, one price. Don't build a pricing page yet; just put a button on the dashboard that says "Activate paid" and pipes to Checkout.
Try two prices, one week apart. Most B2B SaaS MVPs land somewhere between $29 and $99/month per seat. If you're targeting agencies or operations teams, $199 to $499/month per account is realistic. Pick the higher of the two prices you're considering. Founders almost always under-price out of fear, then spend six months trying to raise prices later, which is much harder than starting high.
Aim for 2-3 paying customers by end of week 8. That's not a vanity goal: paid usage tells you 4x more about retention than free usage, because beta users tolerate friction that paying users won't.
| Pricing approach | When to use | Risk |
|---|---|---|
| Flat monthly per user | Single-player tools, simple value prop | Caps revenue per account |
| Per-seat | Team workflows | Slows team adoption |
| Usage-based | API tools, infra, AI wrappers | Hard to forecast for buyer |
| Annual prepay with discount | Have a sales motion | Wrong for self-serve PLG |
If pricing genuinely confuses you at this stage, the right move is a 30-minute call with someone who has priced this exact category before. Not a generalist consultant; someone who shipped pricing for a comparable wedge.
Weeks 9-12: Onboarding flow and first $5k MRR
The last 30 days are about removing yourself from the activation loop. In weeks 5-8, you were probably hand-holding every new user. That doesn't scale and it also hides product problems.
The onboarding milestones, in order:
- Self-serve signup without your involvement. Email + password or magic link, no calendar booking.
- First-run experience that gets the user to value in under 5 minutes. Pre-populated sample data, a guided first action, a clear "you did it" moment.
- Empty states that sell the next action. Every empty list should say "Here's what to do next" with a button.
- One-day reactivation email if the user signed up but didn't complete the first action. Resend handles this fine.
- Activation rate target: 40% of signups complete the first valuable action within 7 days. Below 30% means the product still requires you in the loop.
By the end of week 12, the numbers we'd consider a successful 90-day run:
- $5,000 MRR (around 25 to 50 paying accounts depending on price)
- 40%+ activation rate
- <5% weekly churn
- 1 to 2 channels that produce signups predictably (paid ads, content, founder-led sales, referral)
- A clear "what to ship next" list, written by users, not you
If you hit these, you have a real business. If you miss most of them, the answer isn't "ship harder for another 90 days." It's: kill one feature, narrow the ICP, or restart problem interviews with a tighter wedge.
Comparison: 90-day plans by category
Not every category fits the same 90-day shape. Here's how the plan shifts.
| Category | MVP weeks | Beta size | Realistic 90-day MRR | Notes |
|---|---|---|---|---|
| SaaS productivity tool | 2 | 5-10 users | $3k-8k | Standard playbook |
| Vertical SaaS (e.g. for dentists) | 3-4 | 3-5 accounts | $5k-15k | Higher ACV, fewer logos needed |
| Developer tool / API | 2 | 10-20 devs | $1k-5k | Conversion is slower; OSS distro helps |
| Consumer subscription | 2 | 50-100 users | $2k-6k | Need paid acquisition channel by week 6 |
| Marketplace (two-sided) | 4 | 10 supply + 10 demand | $0-3k | 90 days is short; aim for liquidity, not revenue |
| AI wrapper on existing API | 1-2 | 10-30 users | $5k-20k | Faster builds, faster commoditization |
If you're building a marketplace, do not expect $5k MRR by day 90. The honest target is "10 transactions a week, both sides come back unprompted." Revenue follows liquidity, and liquidity takes 6 to 12 months even with a good wedge.
What to do this week
If you're at week 0 right now, do exactly three things this week:
- Write the one-sentence problem statement and the ICP. Stick it on your desk.
- Book 5 customer interviews for next week.
- Buy the domain and ship a Framer landing page with a single email capture.
Do not pick a stack yet. Do not hire anyone yet. Do not write code yet. The cheapest week of the 90 is the one where you didn't build the wrong thing.
If you're at week 3 and the MVP build is starting, the question is whether you build it yourself or bring in help. The math is straightforward: if you can ship the MVP yourself in 60 to 80 hours of focused work, do it. If not, a mid-tier engineer at $1,000/week for two weeks gets you there for $2,000, which is roughly the cost of one week of your runway if you've raised. Every engineer on Cadence is AI-native by default, vetted on Cursor, Claude Code, and Copilot fluency before they unlock bookings, which is why two weeks is a realistic timeline rather than an optimistic one.
The 48-hour free trial means you can test fit on the first two days before committing to the week. Try a mid-tier engineer for two days on a specific ticket from your MVP backlog and you'll know in 48 hours whether the rest of the 90-day plan is realistic.
You'll also want a way to manage scope creep as you go through week 5 onward, because beta users will ask for things that look small and aren't. Write the scope down before the week starts; protect it like cash.
Try this: book a mid-tier engineer for the MVP build week. Use the 48-hour trial on a real backlog ticket. If it doesn't work, you've lost 48 hours and zero dollars. If it does, you've compressed weeks 3 and 4 of your 90-day plan into 10 working days.
Common founder mistakes in the 90-day window
A short list of things that look reasonable and burn weeks:
- Building auth, billing, or email infrastructure from scratch instead of using Clerk, Stripe Checkout, and Resend.
- Hiring a senior engineer at $1,500/week for MVP work a mid handles fine.
- Skipping the closed beta and going straight to public launch in week 4.
- Polishing the marketing site before there's a product to market.
- Adding a settings page, an admin panel, and team management before there's a single paying team.
- Picking a niche stack because you read a blog post about it. Boring wins.
- Treating week 8 as "we're behind, push the paid conversion to week 12." It always becomes week 16.
FAQ
What's a realistic budget for a 90-day MVP build?
$8,000 to $20,000 for engineering, plus $300 to $800 a month in tooling. If you're doing it yourself, treat your time at $5,000 a month opportunity cost and you'll make smarter buy-vs-build calls.
Do I need a technical co-founder to hit 90-day revenue?
No. A technical co-founder is a 10-year decision; an MVP is a 4-week build. Use weekly-billed engineering help to ship the MVP, then revisit the co-founder question only if the product takes off and you genuinely need a CTO. Most successful 90-day plans don't involve a co-founder at all.
What if I miss the $5k MRR target at day 90?
Don't extend; diagnose. If you have 5+ paying customers who use it weekly, you have product-market signal and revenue will follow. If you have 50 signups and zero payers, the wedge is wrong and another 90 days won't fix it. Restart with tighter ICP and a sharper problem statement.
Should I raise money before or after the 90-day plan?
After, almost always. A pre-revenue raise costs you 15 to 25% of the company at a low valuation. A $5k MRR post-90-days raise is materially better terms because you're selling proof, not hope. The exception is deep-tech or hardware where 90 days isn't enough to ship.
When should I bring in a senior engineer instead of a mid?
When the architecture is genuinely hard: real-time multiplayer, multi-tenant from day one, on-device ML, low-latency video, complex billing. For a standard CRUD SaaS MVP, a mid at $1,000/week ships faster than a senior at $1,500/week because they're not over-engineering for problems you don't have yet. Senior makes sense at week 12+ when you're scaling, not building.
Can I do all 90 days without writing code myself?
Yes, if you have $15k to $20k of engineering budget and you're disciplined about specs. The risk is that you become a project manager for your own product and lose the customer feedback loop. The best founders we see write tickets, do customer interviews, and let the engineer ship; they stay in the loop on what users say, not what the database schema looks like.
If you're at the point where you're about to write your first ticket, the fastest way to start is to book a mid-tier engineer through the standard founder flow and use the 48-hour trial to test fit on a real piece of your MVP backlog. Worst case, you've lost two days. Best case, weeks 3 and 4 of the 90-day plan compress into the next 10 working days, and the revenue target at day 90 stops being theoretical.