TL;DR
- SaaS wins on LTV and compounding. Templates win on time-to-cash, zero churn, and margin that stays near 100% as you scale.
- These are different jobs, not different price models. SaaS is a service business. Templates are a publishing business.
- Build one thing and market it forever, build SaaS. Build many things and let each earn independently, sell templates.
- Numbers below, including the ones that do not flatter templates.
Every few months someone posts the same question on Indie Hackers: SaaS or templates? We spent the last twelve months running a catalog of React Native, Expo and Supabase templates priced $49 to $99. Here is the actual math instead of the slogan.
Two different jobs
A SaaS is one product someone pays monthly to keep using. Unit economics are ARPU times months-to-churn. Distribution is a funnel you optimize forever. Costs are hosting, support, and permanent feature-parity pressure from competitors who never sleep.
A template is a one-time purchase. Someone pays once, downloads the source, and never touches your infrastructure again. Unit economics are price times volume. Distribution is SEO, GitHub, and word of mouth. Costs at rest are close to nothing, because the buyer's laptop hosts the code.
That difference is not cosmetic. SaaS is a service business dressed as software. Templates are a publishing business dressed as software. If you hate status pages and support queues, do not build SaaS. If you hate SEO and writing, do not sell templates.
What twelve months actually taught us
The surprising numbers were structural, not revenue.
Refund rate has stayed under 4% against a seven-day no-questions guarantee. Support volume per sale dropped noticeably after we invested in README.md files and video walkthroughs, though several things shipped in that same window so we cannot cleanly credit any one of them. And churn is zero, because there is nothing to churn from. That is not a growth hack. It is a structural property of one-time payment, and it cuts both ways.
For context on the other side: ScrapingFish's analysis of 937 Stripe-verified Indie Hackers products found more than half generate no revenue at all, and only around 5% clear roughly $8,333 per month, which is the $100K-per-year line.
Template distribution is a different shape but the top of the funnel is just as unforgiving. Most templates on Gumroad and Lemon Squeezy also make nothing. The difference is that a template that finds a niche turns on cash flow without adding operational overhead.
The unit economics side by side
| Dimension | SaaS | Templates |
|---|---|---|
| Payment model | Recurring | One-time |
| Churn risk | High, every month is a re-decision | Zero, payment is terminal |
| Time-to-first-dollar | Weeks to months | Days |
| Ongoing infra cost | Hosting, DB, monitoring, support | Close to nothing |
| Support burden | Continuous, unbounded | Bounded to setup, about a week |
| Compounding | Yes, MRR stacks | Weak, sales do not stack |
| Marginal cost per customer | Real | Near zero |
| Failure mode | Death by churn or CAC blowout | Death by silence |
| Optimal team size | 1 to 3 | 1, and it stays there |
| Exit multiple | Typically 3 to 6x ARR | Typically 1 to 3x annual revenue |
That last row is the honest one. Templates carry lower multiples because buyers price the absence of recurring revenue. But a template business is also cheaper to run indefinitely, which changes the exit calculus for a solo founder. You do not have to sell if it keeps paying without you.
Steelmanning SaaS
Compounding. Every subscriber you keep adds to next month's baseline. Several years in, that baseline is a business. Templates start near zero every month.
Pricing power. Someone paying $99/month for a year has paid $1,188. That is twelve times our most expensive template. If you can build something a business pays $99/month for and hold them two years, you are building something structurally more valuable per user than any template can be.
A data flywheel. SaaS gives you telemetry, cohorts, usage patterns. You learn what to build next from live behavior. Template buyers vanish. You get email replies, the occasional GitHub issue, and silence. Iterating on templates is like iterating on a book: slow, blind, one shot per edition.
Enterprise doors. A SaaS has a login, a dashboard, an admin panel, an invoice generator. It can slot into a company. A template cannot.
Where templates actually win
Support ends at week two. Once the template runs, the buyer leaves your world. Their Supabase, their App Store account, their EAS project. Our inbox on a given Monday holds a handful of tickets, almost all setup, almost all one reply.
Time-to-cash is days. Publish, write one post, first sale inside a week is realistic. Getting a SaaS to the point where someone pays $19/month takes months of onboarding polish, billing plumbing, and email infrastructure.
Range is a moat. With SaaS you have one product forever. Our catalog now runs to nineteen templates. Each is an independent SEO surface, an independent launch, an independent asset. If one flops the others are unaffected.
No infrastructure hostage situation. Supabase raises prices, the buyer runs their own project. Expo changes EAS pricing, the buyer pays. Gross margin near 100% that stays there as you scale is genuinely rare.
Solo founders scale further. A one-person SaaS at $10K MRR probably carries a few hundred customers and a support load that eats weekends. A template business at similar monthly revenue can be a hundred-odd sales, most of whom never write to you.
The hidden cost of SaaS
Read enough post-mortems and the theme is not product failure, it is operational drag. Incidents at 2am. Stripe webhooks that silently stop firing. Trial users who never convert but still file tickets. The billing page failing an upgrade for the third time this month.
Templates dodge most of that. Ship code, buyer owns code, done. When something changes upstream we update the template, push, and email license holders. No downtime, no incident report, no war room. The tradeoff is that documentation has to carry weight a support team would otherwise carry.
The hidden cost of templates
You become a content marketer. There is no freemium hook, no viral loop, no "my friend used this so I should too." You get traffic by writing. Every post is an asset and every skipped week is a missed month of compounding search. If you hate writing, do not sell templates.
The second cost is decay. React Native shipped a new architecture. Expo keeps shipping SDK releases. Every cycle is a maintenance sprint across the whole catalog. If you sell lifetime updates, and we do, you signed up for that permanently. Price accordingly.
The pricing math
We landed on $49 to $99 because it is the credit-card-decision zone. Under $50 nobody asks their finance team. Under $100 nobody asks their spouse. Above $100 you are selling to companies and you need invoices, POs, and a sales motion. Fine business, different business.
At $79, a template prices at roughly one to two hours of a US developer's time. A React Native and Supabase auth stack takes days to build properly from scratch, even with an agent writing most of the lines.
Now compare that $79 to a $19/month SaaS. The SaaS hits parity in month five and passes it forever after. At eighteen months of retention, strong for SMB SaaS, that is $342 LTV. More than four times the template price.
That is why SaaS wins on LTV and it is not close. But reaching $19/month and holding users eighteen months is a specific business with specific requirements: onboarding, dunning, a status page, a support policy, a churn problem by year two. Templates skip all of it and earn on day one.
What we would do differently
Ship the bundle sooner. Every individual sale is evidence the buyer wants your work, not just one file of it. We priced individually for most of the first year. The bundle should have existed months earlier.
Test a licensing tier. Free-with-attribution alongside a paid commercial license. Open-source companies worked this out years ago and template sellers largely have not.
Lead with the ecosystem, not the file. Templates that ship with agent configuration, pre-installed skills, and slash commands are worth more than a plain repo. We build that way now. We were slow to make it the headline. The buyer is not paying for code, they are paying for a foundation an agent can extend without breaking it.
The decision tree
SaaS if you have deep domain expertise in one vertical, you enjoy customer conversations, you tolerate operational overhead, and you want a compounding asset you can eventually sell.
Templates if you have range across stacks, you are a strong writer or SEO operator, you want zero ongoing infrastructure, and you can live with sales that do not stack.
Both if you have a template that could plausibly become a hosted service. Ship the template, watch which buyers ask for a hosted version, and let them validate the SaaS for you.
Most indie hackers pick SaaS because that is what the timeline celebrates. A modest template catalog run by one person is one of the least stressful, highest-margin businesses in software. It will not get you on a podcast. It might get you out of your job.
If you have sold either, I want the number that surprised you. Not MRR. The structural one: refund rate, support load per sale, how long a customer actually stayed. Those are the numbers nobody screenshots and they are the ones that decide whether the business is livable. Drop yours in the comments.
