How to Find a SaaS Idea Without an Audience: Borrow Distribution, Mine Communities, and Verify With Free Tools

The short version: Katie's first product came from a forum thread she didn't write — she picked the idea with the most votes and built the plugin for it. Today that habit runs a portfolio of 19 WordPress plugins making around $150,000 a month. Ivan reached $10,000 a month with two partners and no audience of his own. Evan paid a creator $800 and got roughly a million views. None of them built an audience before they had customers.

I went through the 100 founder interviews published by Starter Story and pulled every play where someone started with zero audience. The strongest ones share a pattern, and it isn't "find your niche." It's: go where your buyers already are, and let someone else's audience be your first test market. This article is that playbook — the three routes ranked, the four ways founders borrowed distribution, two amplifiers (AI to find the pain, free tools to verify it), and the ways borrowing backfires.

Why does borrowed distribution rank first?

Borrowing distribution means one specific move: using someone else's audience — a platform's feature-request list, a coach's students, an agency's client roster, a creator's followers — as the first place your product meets buyers.

I ranked the three main routes by what they cost, how fast they produce a signal, and how much of the problem they solve on their own. The ranking is from this dataset of 100 interviews, not from the industry at large.

Routes 2 and 3 have their own guides — testing an idea from your own pain point and finding gaps inside products that already make money. Both are solid routes, and both leave you standing at the finish line with nobody watching. That is why the first route leads here: it starts with buyers, not with an idea. These three are also a slice of a bigger map — the full set of patterns and angles from the same 100 interviews lives in the main guide.

Mal, who built a niche app called Prayer Lock, learned the difference the expensive way:

"I wasted years building stuff nobody wanted... Instead improve an already successful app. Add your personality. Copy others distribution strategies."

Din, who built Supergrow, gives indie hackers the same warning:

"Don't try to create new market. Just work on the validated market... go for markets which are actually generating revenue."

Nobody in this dataset built an audience first. They borrowed one.

How do you borrow distribution?

Four lanes show up again and again. Each one is a different answer to "where does an audience that isn't mine already exist?"

1. Platform request lists. Katie's first plugin came from a site called the WooCommerce ideas forum. The move was simple, and the words matter: "We chose an idea that lots of people were voting for." The plugin password-protected store categories — and because it was first to serve that request, it ranked on Google almost immediately: "We got our first sales within a few days." The same instinct — find a platform's public wish list and build from it — powers the Data Fetcher playbook covered in the competitor guide.

*First step: open the official feature-request forum or marketplace wish list for the platform you already use, sort by votes, and read the top 20 requests.*

2. Coaches and experts. Ivan's buyers are Upwork freelancers, so he went to the people who teach Upwork: coaches with students of their own. The first coach came through a beta user's introduction. The second came from cold outreach, with an unusual opener: "I straight up just offer him $1,000 to jump on a call with me — essentially pay him to be our user." The deal structure: "If they completely sign a client, they onboard them and set them up... they get 30% commission for a lifetime. And if they just refer them, then that's 20%." Ivan calls 20–30% lifetime the standard for this kind of partner. Result: $10,000 a month in the third or fourth month after launch, on two partnerships. Thomas, who built Packager, ran the same lane with experts instead of coaches: "Partnering with people who specialize in the product. So for us that was Microsoft MVPs. They would create demos of the product, which was a highly targeted audience." Packager now makes over $60,000 a month.

*First step: list the consultants, coaches, and course creators whose students are your buyers — then work out what a lifetime commission is worth to each of them.*

3. Agencies. Zach and Brendan sell to agencies instead of brands — the firms that manage marketing for 20–30 clients each. Finding them is a browsing exercise: "You can go to the Shopify agency partner page directory, the Klaviyo partner page directory, and they basically give you a force-ranked list of agencies by size and influence." Then comes the key assumption, in his words: "Because the agencies are all [on] the same platform, you can kind of safely assume they have the same problem." One sale lands an entire client roster, which is why a 19-month-old product sits at $96.2K in monthly revenue.

*First step: open the partner directory for the platform your buyers use and look just below the top tier — the agencies with 10 to 30 clients.*

4. Creators and communities. Evan's outreach to creators had one rule: "Send each creator the same message every time, which should always start with 'paid promo?'" Then a call, then a fixed structure: "We pay them something like $800 with a $600,000 minimum view clause." One of those videos produced roughly a million views, 1,800 downloads, and about $3,000 in revenue; his app Locked reached $14,000 a month. Tor ran the stronger version of this deal in Israel with his localized calorie app: he deliberately waited until the app made $20K a month before approaching big fitness influencers. His reasoning: "I can negotiate a better deal... so they get 50/50" — a full revenue share that gives the creator "skin in the game." The app went from $20K to over $80K a month. And the free version of this lane is community itself: Sam grew Algrow inside a niche Discord server where he would "mute my mic and literally just share my screen" and help people find their own niches. His first 400 users came from there; the product now makes around $14,000 a month.

*First step: list 20 creators or communities where your buyers already hang out — then send every one of them the same two-line message.*

And once a borrowed audience is watching, the next move is to get paid before you build: that is selling it before you build it, covered in its own guide.

How do you use AI to find ideas with no audience?

Five workflows from the dataset, in ascending order of effort:

1. Mine the chat logs. Sam's method: copy days' worth of chat history from a community your buyers use, paste it into an AI, and ask a specific question: "List me all of the pain points that these people have talked about over the last couple days." He looks for pain points that repeat — "the more struggle there is, the more likely the product is going to do better."

2. Scan an entire review category. Nick built BlockToPin by studying what people hated about every existing tool in his space: "Research their G2 reviews, maybe [AppSumo] reviews, [Capterra]... watch every single one of them and understand what are the pain points." His MVP bar: "[Claude Code]... it can be done in one, two weeks maximum." BlockToPin makes over $16K in monthly revenue — after his earlier AI apps earned him zero.

3. Turn your existing spreadsheet into the product. Jack had been building complicated P&L spreadsheets by hand for clients. The product version was almost literal: "Downloading the spreadsheet as a CSV, uploading it to Cursor and saying 'build it' — then identifying all the ways that Cursor got it wrong." Profit AI now makes $30,000 a month.

4. Swipe the best onboarding, skip the mockups. Mal's advice from building Prayer Lock: "Do not design any mock-ups. Instead study the best apps in your niche." He built a Canva board of "all of the best onboarding screens of the apps I liked and of my competitors" and used them directly — then shipped the code in days.

5. Make 2,000 phone calls. Gorav and Jock did 20 customer calls a week — about 2,000 in total — ran every call through an AI notetaker into a database, and connected it to Claude through MCP. A non-technical founder built the whole analysis in natural language. What it revealed: their tool had been trying to do everything, and only one feature mattered. They cut the rest.

Now the honest part. In 100 interviews, nobody shared a story about AI sending them the wrong way. Not one founder described a direction that failed because an AI's analysis was bad. So this article can't give you an "AI misfire" checklist — it doesn't exist in this dataset. What the data does show is a narrower role: AI compresses material that already exists — chat logs, reviews, calls, screenshots. It is a reading machine, not a decision-maker.

Which free tools can verify an idea before you build?

The founders in this dataset verify with public data before writing code. Six tools carried most of that work:

Then there's the part the tool pages leave out: a green light is not a sale.

When the data says go — and it's wrong. Aayush's team ranked number one on Google for six to eight months on a keyword with high volume and low competition. The traffic came; the buyers didn't: "Of course, like, this was not a very high intent audience." Ankit got 120,000 views on a video and, by his own count, "nine or eight trials" — because the comments were memes, not questions: "They're not asking about the product... that's the biggest sign that it's not going to lead to high conversions."

When the data says no — and it's wrong. Steve built for Android when the conventional wisdom said iOS: "That's where all the money is... But we went against that." The reason wasn't contrarianism — it was the numbers: a 2025 subscription-apps report showed iOS users cost around four times more to advertise to and converted only about 20% better. Result: "around 80% of our users and around 80% of our revenue comes from Android," on the way to over $100,000 a month.

A green light from a tool means people are looking. It doesn't mean they're buying. Traffic is just people watching. A sale is proof. Don't mistake attention for intent to buy.

And one more honest note: no founder in this dataset accused a tool of lying to them. The tools measure attention. Attention is not a payment — that gap is yours to test, not theirs.

What goes wrong when you borrow?

Every borrowed channel can break in a different way. Five failures from the dataset:

The questions those failures point to: What does this channel actually charge — and did you read the pricing page before building on it? Does it deliver buyers, or just views? And if access gets revoked, what's left?

Everything you borrow can be taken back. Your users, your product, and your data are the assets that actually stay in your hands.

FAQ

Is borrowing distribution the same as affiliate marketing?

No. Affiliate programs pay for clicks or sales that come through a link. The founders here recruited named partners — coaches, agencies, creators — one at a time, and gave them a stake in the outcome: 20–30% lifetime commissions, 50/50 revenue shares, minimum-view contracts. The recruiting was manual and personal.

What if I can't pay partners upfront?

Most deals in this dataset are performance-based. Ivan paid 20–30% of revenue, for life; his one upfront payment was $1,000 — for a single phone call, just to get the meeting. Tor's influencer deal was a pure 50/50 revenue share. Evan's creator contract was a small fee ($800) plus a minimum-view guarantee, so the creator carried the performance risk too.

Can AI find an idea if I don't know where my buyers hang out?

The workflows in this dataset all start with a place that already contains your buyers' words: a Discord server, a review category, a spreadsheet from your own work, a stack of sales calls. AI compresses that material. Finding the place is still your job — Sam started his search on a Discord directory, Nick started on review sites.

Are these tools actually free?

The free ones: TikTok search, app-store reviews and autocomplete, Reddit plus Map of Reddit and F5bot, and Google Trends. The cheap-but-paid ones: one founder bought a single month of a keyword tool for $129 to do concentrated research; another mentioned his app-store keyword tool costs about $10 a month. Sensor Tower's free tier wasn't described in the interviews, so treat the cost as unverified — check it before you plan around it.

What happens if the platform changes the rules?

It happened twice in this dataset: a billing-rule change that made a channel too expensive (Lots and WhatsApp) and a community ban (Dennis and a travel subreddit). Both founders rebuilt elsewhere — one on his own infrastructure, one in communities that allow promotion. Nobody described being compensated for what they lost. Borrowing means the lender keeps the keys.

---

One more thing. The ranked routes, the four borrowing lanes, and the tool list are in a one-page checklist I put together from the interviews. Free, just an email.

Get the checklist →

---

Related reading

• How to Find a SaaS Idea That Actually Makes Money
• How to Find SaaS Ideas From Competitors
• How to Pre-Sell a SaaS Product

*Source note: every case in this article comes from Starter Story's founder interviews. I went through 100 of them and checked every load-bearing number against the original transcripts. All results are self-reported.*