Skip to content
Offlist

How to Start a SaaS Business: Demand, Pricing and Churn

A SaaS business is a bet that you can acquire a customer for less than they will pay you before they leave. Everything else is detail. Founders who understand that build differently: they sell first, price deliberately, and watch retention harder than signups. This is how to construct the model before you write the code.

Don't pay full price to build this.All 105 SaaS tools below come with a discount code, a startup programme or free credits.Claim these perks →

Sell the thing before you build the thing

Demand validation is not a survey. A survey measures politeness. The only signal that counts is someone agreeing to pay, in writing, for something that does not exist yet. Write a one-page description of the outcome, put a price on it, and take that to twenty people who have the problem. If none of them will commit, the problem is not painful enough.

Design partners are the strongest version of this. Three or four companies who agree to pay a reduced rate for early access, give you weekly feedback, and act as references at launch. They shape the roadmap in exchange for influence over it, and they convert into your first case studies. The reduced rate is the point: free pilots generate opinions, paid pilots generate requirements.

What goes wrong here is building for a hypothetical buyer who agrees with everything. If every conversation is encouraging and nobody has objections, you are talking to people who are not going to buy. Real prospects push back on price, integration and switching cost, and those objections are the specification.

Define an ICP narrow enough to be findable

An ideal customer profile is useful only if it tells you where to look. 'Small businesses' is not an ICP. 'Dental practices in the UK with two to six chairs that already use practice management software' is, because you can list them, find their email addresses, and name the competitor you are replacing.

Narrow beats broad early for a mechanical reason: your marketing copy, your onboarding and your integrations can all be specific, and specific converts. A wider market is available later by adding adjacent segments, and each expansion reuses the machinery you built. Starting wide means every asset is generic and nothing lands.

Write down who you are explicitly not selling to, and enforce it. Every early SaaS company gets pulled off course by one large customer in the wrong segment who pays well and demands a product you did not intend to build. Saying no to that deal is the highest-leverage decision of the first year.

Pricing is structure first, number second

Before you argue about the number, choose what you charge for. Per seat scales with the customer's team and is easy to forecast, but it punishes adoption and invites licence sharing. Usage-based scales with value delivered and grows without a sales conversation, but it makes revenue lumpy and budgets unpredictable for the buyer. Flat tiers are simplest and leave the most money uncollected.

Pick the axis that grows as the customer gets more value from you, then set the entry price higher than feels comfortable. Underpricing is not a growth strategy, it is a filter that attracts the customers who churn fastest and complain loudest. Raising prices later is straightforward for new signups and awkward for the existing base, which is a good reason to start high.

Annual billing is worth engineering for from the start. It collects a year of cash up front, removes eleven monthly opportunities to cancel, and materially improves reported retention. Offer a discount for it, make it the default on the pricing page, and build the billing system so switching between terms does not require a support ticket.

The four numbers that decide whether this works

Gross margin comes first. Subtract hosting, third-party APIs, payment fees and support cost from revenue. Software businesses are valued on the assumption this is high, and an AI-heavy product with inference costs per request often is not. If your margin is thin, you have a services business with a login screen, and the growth maths does not apply.

Then customer acquisition cost and payback period. Total everything spent on getting customers, including your own time, and divide by customers won. Payback is that figure divided by monthly gross profit per customer. Under twelve months means you can grow on your own cash. Beyond eighteen, you are funding growth from a balance sheet you do not have.

The fourth number is net revenue retention, which combines churn with expansion. If existing customers spend more each year than the ones who leave took with them, the business grows even with zero new signups. That single property is what separates a SaaS company from a subscription treadmill, and it is a product decision more than a marketing one.

Churn is the thing that actually kills you

Growth hides churn for as long as new signups outnumber cancellations. While that holds, the chart goes up and nobody investigates. Then acquisition plateaus, the leak becomes visible, and by then the cohort data you needed to diagnose it was never collected. Track cancellations by signup month from the first month, not from the month you start worrying.

Most early churn is not a product failure, it is an activation failure. The customer signed up, never reached the moment where the product became useful, and cancelled without ever forming an opinion. Find the action that correlates with staying, then rebuild onboarding so reaching it is the only path. Everything else in the interface is secondary to that one step.

Involuntary churn is the cheapest to fix and the most ignored. Cards expire, banks decline, and a meaningful share of monthly cancellations are simply failed payments nobody retried. Dunning sequences, card updater services and a pre-expiry email recover a chunk of that revenue with no product work at all.

The first ten customers, and where they come from

Not from content, and not from ads. The first ten come from people you contact individually, from the community where your ICP already argues about this problem, and from the design partners who agreed to pay before launch. Content and paid acquisition are scaling channels, and they need a proven conversion path to scale.

Put every one of those conversations in a CRM from the first email, even if the CRM is free and you are the only user. The value is not pipeline management at ten deals, it is the record of what objection came up, what you said, and whether it worked. That record becomes your sales copy and your onboarding script.

The stack for all of this, CRM, product analytics, email, billing and payments, is genuinely expensive at list price and largely free in year one if you apply to the founder programmes. Our SaaS directory keeps the live terms for each one, including the revenue and headcount ceilings that decide eligibility. Apply before you need the tool, not after the trial expires and the invoice lands.

The offers mentioned above

Moderate to redeem

Pulled live from the directory, so what you see here is what the listing says.

  • HubSpot

    Marketing & Growth

    Up to 90% off year one

    Percentage discountLink on plan
  • Pipedrive

    Sales & CRM

    14-day free trial; save up to 42% billed annually

    Free periodDirect link
  • PostHog

    Data & Analytics

    $50,000 in credits for 12 months

    CreditsLink on plan
  • Mixpanel

    Data & Analytics

    1 year free

    Free periodLink on plan
  • Stripe

    Finance & Accounting

    $25,000

    CreditsPartnerLink on plan
  • Chargebee

    Finance & Accounting

    Free billing until $1M in cumulative revenue

    Free periodLink on plan
  • Mailchimp

    Marketing & Growth

    Free plan, 14-day Standard trial, 15% off for 12 months at 10,000+ contacts

    Free planDirect link

SaaS codes, all in one place

105 listings in SaaS, with the redemption codes and programme links behind one subscription. Browsing is free.

Common questions

How do I validate a SaaS idea without building it?

Put a price on a written description of the outcome and ask people to commit. A signed design partner agreement or a prepayment is validation. Interest, feedback and waiting list signups are not.

How much money do I need to start a SaaS business?

Less than the tooling list suggests, because almost every layer of the stack runs a founder programme covering the first year. The real cost is your time between the first line of code and the first paying customer.

Should I charge per seat or per usage?

Charge on the axis that grows as the customer gets more value. Per seat suits collaboration tools, usage suits infrastructure and anything with a variable delivery cost. Mixing both confuses buyers and slows deals.

When should I hire my first salesperson?

After you personally close enough deals to know the objections and the winning sequence. Hiring sales before a repeatable script exists means paying someone to discover the process for you, badly.

More on SaaS

Guides, comparisons and code round-ups covering the same ground.

Browse SaaS perks · See what access costs