Intro.
#Why You Should Start With Things That Don't Scale
In his 2013 essay "Do Things That Don't Scale," Paul Graham calls out a trap early founders fall into constantly: the belief that if you build a good product, users will show up on their own. Reality runs the other way. Early on, you have to bring users in by hand, one at a time.
The core idea is: do now what you won't be able to do later. Things you could never do for a million users are exactly what you should do for your first 100 — track people down and persuade them individually, install the product for them yourself, watch over their shoulder as they use it. That's not inefficiency. It's the standard playbook for an early stage.
Two well-known examples: Airbnb's founders went door-to-door in New York personally recruiting hosts, and when listing photos were weak, they'd go take professional photos themselves. Stripe's founders, instead of emailing a link to anyone willing to try their payment integration, would say "just hand me your laptop" and install it on the spot. Both were labor-intensive and about as far from scalable as it gets.
TIP
Things that don't scale aren't a lifelong habit — they're something you do only in the early days. The point is twofold: land your first users, and learn as fast as possible why the product isn't sticking.
02
#What Counts as Traction vs. What's a Vanity Metric
The traction that grant reviewers and investors actually credit isn't "how many people saw this" — it's "how many of the people who saw it actually acted." The core is action, repetition, and payment, not impressions or reach. Even with the same number, the two columns below carry very different weight.
| Category | Vanity metric (reference only) | Behavioral metric (credited) |
|---|
| Interest | App downloads, page impressions | Completion rate of first core action after signup |
| Engagement | Social followers, likes | Weekly / monthly return visits (retention) |
| Demand | Waitlist signups | Paid conversion rate, number of transactions |
| Persistence | Cumulative signups | Share of users still active after 30 days |
| Customer response | A compliment — "I like it" | Paid, or actually referred a friend |
Vanity metrics only ever move up and to the right, which feels great, but tells you almost nothing about whether the business actually works. "30 people who pay and come back every week" beats 10,000 downloads in a review — because the former is proof that the demand is real.
주의
If your application states "5,000 cumulative signups" with no return-visit or payment data next to it, a reviewer may read that as "acquired a lot, but nobody sticks." Even a small number, pair it with behavioral metrics.
03
#Pick 1–2 Channels and Go Deep
Gabriel Weinberg and Justin Mares' book "Traction" maps out 19 channels for acquiring users — content, search ads, SEO, word of mouth, sales, partnerships, offline events, and more. What matters isn't the count, it's the choice. Early on, don't touch all 19 — concentrate your firepower on the 1–2 with the best odds.
For an early-stage founder, the channels realistically within reach in the 0-to-100 stage usually fall into five buckets. Paid channels like ads aren't a priority at this stage.
| Channel bucket | What it is | Fit for 0→100 | Watch out for |
|---|
| Cold outreach | Direct recruiting via DM, email, or phone | High | Personalize every message — no spammy mass sends |
| Community | Participating in the group chats, Discords, or forums where your target hangs out | High | Don't spam promotional posts — contribute and help first |
| Offline touch | Face-to-face pitches at events, meetups, or stores | Medium | Especially effective when your audience already gathers in one place |
| Content | Articles, reviews, or videos on the problem that drive search traffic | Medium | Slow to pay off — better to double down after you hit 100 |
| Partnerships | Connecting or teaming up with someone who already has your customers | Situational | Over-reliance becomes a channel risk — use for validation only |
The two fastest options in the 0-to-100 range are usually cold outreach and community. Both cost almost nothing and let you hear directly, one conversation at a time, why the product isn't landing. Content and SEO compound over time, so it's more efficient to go deep on them once you already have some users.
04
#A 0→100 Execution Playbook
Below is a sample built around a 4-week cycle. It's not a fixed answer, just a framework showing the sequence. The core is repeating "recruit → onboard personally → listen → double down on one channel" as fast as possible.
- Week 1 — Manual recruiting: build a list of 30–50 target people by hand, pick one channel, and reach out individually with a personalized message. No automation, no scheduled sends — one at a time.
- Week 2 — Hands-on onboarding: don't just hand new signups a link — install it for them or walk through the screen together until they hit their first core action. Watch closely for where they get stuck.
- Week 3 — Collect feedback: talk to the people who used it, even briefly. "Why didn't you come back" is the single most important question. Write down reasons for churn, not compliments.
- Week 4 — Double down on one channel: pour next cycle's effort into whichever channel got the best response. Cut channels that didn't work without hesitation.
While running the cycle, tracking just two numbers a week is enough: (1) how many people completed the core action for the first time this week, and (2) what share of last week's users came back this week. Cumulative totals like total signups aren't worth chasing at this stage.
체크
The point to flip on automation and paid ads is usually after you've hand-built roughly 100 users and confirmed return visits aren't zero. Pouring ad spend into a channel where nobody sticks is pouring water into a leaky bucket. Plug the leak first, then scale.
How much value you get back per dollar spent (value relative to acquisition cost) is a topic to dig into seriously after you pass 100 users. See a separate piece on unit economics (customer acquisition cost and lifetime value). In the 0-to-100 range, the priority is hands-on recruiting, not number-crunching.
05
#A Pre-Submission Traction Self-Check
Before submitting a grant application or IR materials, use the list below to check whether your traction reads as "evidence of action." If more than half is blank, you're better off running one more hands-on cycle than polishing the document.
- You can explain in one sentence which channel you used to get your first 100 users, and how.
- You can cite a number for "core actions completed," not downloads or impressions.
- You know your 30-day retention rate.
- If you have paid conversions, you can state the count and the amount.
- You can list 2–3 reasons you heard directly for why users churned.
- You focused on 1–2 channels, and you can explain why those channels.
- You can back up "we have return visits" with evidence — screenshots, logs, and the like.
TIP
Reviewers care less about the size of the number than about whether this team is learning by meeting customers directly. Even 100 users, if you can check all 7 boxes above, is more persuasive than a vague 10,000 downloads.
Summary.
#Frequently Asked Questions (FAQ)
Q. I still have zero users. What should I do first?
A. Build a list of 30–50 target people by hand and start reaching out one at a time through a single channel. Person-to-person conversation — not ads, not automated sends — is what gets you from 0 to 1.
Q. Manual recruiting is too slow. Can't I just fill the funnel with ads?
A. If you turn on ads while return visits are at zero, whoever you acquire just leaks right back out. Ads are a tool for scaling up a leak you've already confirmed doesn't exist. Recruit by hand first and fix why people aren't sticking.
Q. Is it okay to apply for a grant with almost no traction?
A. Yes. Just show "evidence of action" and "what you learned" instead of big numbers. Even with a small sample, laying out your core action, retention, and reasons for churn signals a team that can execute.
Q. Wouldn't running several channels at once be faster?
A. Usually slower early on, because your firepower gets spread thin and no single channel gets properly validated. Focus on 1–2, confirm at least one channel actually works, then expand.
CTA
Experience an OpenSeed AI review yourself. Multiple AI agents collaborate to analyze your business plan, delivering not just a score but a capability-by-capability diagnosis, supporting evidence, and concrete recommendations for improvement.
Is Your Traction Actually Reading the Way Reviewers Read It?
OpenSeed's AI review has multiple agents diagnose the traction and execution evidence in your business plan, capability by capability, and prescribe exactly what to strengthen and how. Check it before you submit.
🔒 Free during beta · your submission isn't saved
Start Free AI Feedback →