Growth strategies for early-stage founders
24 strategies from Awiser's growth copilot, each written up with when it fits, how to run it, documented cases with their numbers, the evidence against it, and an evidence grade. Every claim links to its source. No account needed.
How to choose a strategy
Four stages, each with the question it has to answer first.
- Stage 1. No proven demandIs the problem real, and does anyone want this solved?10 strategies
- Stage 2. Users, no retention signalDo the first users get value, and do they come back?7 strategies
- Stage 3. Faint pull, earn the right to distributeHow do more of the right people find it?7 strategies
- Stage 4. A repeatable channelHow does the channel that works get cheaper and run without you?0 strategies
Start with the earliest stage whose question your own numbers cannot answer yet, which is also where Awiser's copilot starts. Then narrow by what you can spend: cash, hours, and whether you are willing to talk to people live.
All 24 strategies
Within a stage, the strategies to start with come first, then the ones with the strongest evidence.
All 24 strategies
Stage 1. No proven demand
Is the problem real, and does anyone want this solved?
- Set up the technical SEO basics
Google can index a page that is not blocked, returns HTTP 200 and has text. What the setup does, what it does not, and how long indexing takes.
Search and content1 to 2 daysNo cashEvidence BStart here
- Run 15 problem interviews before you build
Interview studies reached saturation at 9 to 17 interviews in empirical tests. How to run problem interviews without pitching, and what they cannot tell you.
Customer research1 to 2 weeksNo cashEvidence CStart here
- Sharpen your ideal customer profile into one sentence
An ICP names who cares most about what makes you different, and where to find them. Why a yes-or-no read-back test invites polite agreement, and what to ask.
Positioning2 to 3 hoursNo cashEvidence CStart here
- Test your one-liner on 10 strangers
Show your one-sentence pitch to 10 strangers in your audience and ask who it is for and what it does. Why 7 of 10 decides little and five seconds is too short.
Positioning2 to 4 hoursNo cashEvidence CStart here
- Onboard your first users by hand
Superhuman saw over 65% of customers fully switch after a human onboarding session, double self-serve, with selection bias. When doing it by hand pays off.
Demand tests2 to 4 weeksNo cashEvidence CStrong when it fits
- Ask target users to keep a one-week pain journal
Diary studies capture problems as they happen. Experience-sampling studies averaged 7 days and 79% compliance, mostly paid. How to run a pain journal.
Customer research1 to 2 weeksNo cashEvidence CSituational
- Build a free tool that captures emails
An agency saw leads rise from 138 to 606 a month with free tools, as close rates fell from 5.0% to 2.2%. How to gate a free tool without breaking EU rules.
Demand tests1 to 2 weeksNo cashEvidence CSituational
- Interview people who recently chose a competitor
Interview people who recently switched to or away from a competitor, not settled users. One practice needs 8 to 10 interviews. Who to recruit and what to ask.
Customer research1 to 2 weeksNo cashEvidence CSituational
- Observe your users in their communities
Watching the communities your users talk in surfaces pain quickly, but pain is not demand: 71 threads of complaints, none mentioning paying for a fix.
Customer research10 to 40 hours over a few weeksNo cashEvidence CSituational
- Test demand with a landing page before you build
The median landing page converts 6.6% of visitors, and most waitlist signups never pay. How to run a smoke test before building and read its result honestly.
Demand tests1 to 2 days, then a few weeks of trafficNo cashEvidence CSituational
Stage 2. Users, no retention signal
Do the first users get value, and do they come back?
- Launch on Show HN with a technical angle
Post something people can try on Hacker News. The median Show HN scores 2 points: the rules, when it fits, and what founders report.
Launches1 to 2 daysNo cashEvidence BStrong when it fits
- Ask every churned user what made them stop
One open question by email got about 19% replies from cancelled customers at Groove, against 1.3% for a multiple-choice survey. How to ask churned users.
Activation and retention2 to 4 hours, then a week of repliesNo cashEvidence CStrong when it fits
- Give your homepage one question per section
Visitors scan headings, and 71.6% of screen reader users turn to them first. How to structure a homepage so the headings alone say who it is for and why.
Positioning1 to 2 daysNo cashEvidence CStrong when it fits
- Offer a live onboarding call to every new signup
Superhuman reached 100% attendance only by requiring its onboarding call; offered as an option, about 15% came. When live onboarding calls pay off.
Activation and retention30 minutes per signupNo cashEvidence CStrong when it fits
- Run jobs-to-be-done interviews with recent adopters
JTBD practice interviews people who adopted recently, not long-time users. Commoncog found its customers came by a path it had not assumed. How to run them.
Customer research1 to 2 weeksNo cashEvidence CStrong when it fits
- Launch on Product Hunt
Product Hunt featured 6% of submissions in 2025, per Hunted.Space. What the rules forbid, when a launch fits, and what founders report.
Launches1 to 2 weeksNo cashEvidence BSituational
- Write an Indie Hackers post
Story posts outperform launch announcements on Indie Hackers, which retired Milestones and Groups in 2026. When a post fits and what founders report.
Launches2 to 3 hoursNo cashEvidence CSituational
Stage 3. Faint pull, earn the right to distribute
How do more of the right people find it?
- Make your pages readable by AI search engines
Google says its AI features need no special markup, and pages that added schema saw no major citation uplift. What to check so AI search can read you.
Search and content1 dayNo cashEvidence CSituational
- Publish one original data point
A two-hour table of public prices drew 12,000 visitors in a day, then the same founder's next post got no upvotes. When original data earns links.
Search and content1 day to 2 weeksNo cashEvidence CSituational
- Record one tutorial for a task buyers search
In a random sample of YouTube, 13% of videos had over 1,000 views, yet one founder got three customers from about 370 views. When a tutorial pays off.
Search and content1 dayNo cashEvidence CSituational
- Write an honest comparison with a competitor
Comparison posts convert well when they rank, says one agency that sells them, yet most competitor keywords converted under 4%. How to write one fairly.
Search and content1 to 2 daysNo cashEvidence CSituational
- Write one SEO post for a query your buyers search
Google has no preferred word count, and Ahrefs found 1.74% of new pages reach the top 10 within a year. How to pick a query that brings buyers.
Search and content4 to 8 hoursNo cashEvidence CSituational
- Submit to 10 well-chosen directories
One B2B SaaS submitted to about 160 directories: 171 of 3,300 sessions and 2 attributed signups. When directories fit, and how to track them.
Launches2 to 4 hoursNo cashEvidence BLow priority early
- Pay to boost a directory that already sends visitors
Uneed says its skip-the-line fee only moves your launch date, and one $99 BetaList feature brought 5 email signups. When paying a directory makes sense.
Paid acquisition2 to 4 hoursUnder $100Evidence CLow priority early
How the strategies are graded
What the grade and the verdict mean, and how sources are checked.
Every strategy carries two marks. The evidence grade says how well published sources support the result the strategy is meant to produce. The verdict is Awiser's call on how much an early-stage founder should care, given cost, risk and how fast a result shows.
- A. Strong evidence
- A dataset with a stated method and sample, or the platform's own rules or data, plus at least one independent documented case that agrees.
- B. Moderate evidence
- Two or more independent documented cases with numbers that agree, or one dataset with a stated method and no independent case.
- C. Limited evidence
- A single documented case, or a recognised method with no published outcome data.
- D. Unverified
- No verifiable source for the outcome was found. The strategy rests on reasoning alone.
- Start here
- Cheap, low risk and useful to most founders at its stage.
- Strong when it fits
- Reported results are good, but only when its conditions hold.
- Situational
- Useful for a narrow set of products or audiences.
- Low priority early
- Weak or contrary evidence, or costly for an early project. Try cheaper strategies first.
- Every quoted passage was read on its source page on the date shown, and a test refuses any quote that the checking script cannot find on that page.
- A sentence with a number names its source in the same sentence.
- A source that sells the channel, tool or service it measures is marked as vendor data.
- A single founder's report is one case, not a typical result, and the page says so.
- No result measured on Awiser's own users is published yet: too few projects have reported back on the same strategy for a number to mean anything.
A source says something else, or a better one exists? Send it and the page gets re-checked.
For agents and developers
The catalog is open to software too, with no key and no account. Responses carry the same sources and quotes as the pages, and each strategy's page URL to cite.
- Search
GET https://awiser.co/api/growth/strategies?q=first+users&stage=1returns the matching strategies, best match first. Filters:stage,category,no_cash=1,no_calls=1,limit.- One strategy
GET https://awiser.co/api/growth/strategies/<slug>returns the whole strategy: when it fits, steps, documented cases, evidence against, grade and every source with its quotes.
Awiser is the builder network where founders show their skills, find collaborators, and grow their projects with an AI copilot.