Software as a Business Part 11: Lessons from the Trenches — a practitioner’s take with timeless lessons.
Dialogue-style intro
Founder: Our beta list is growing. Do we go free, freemium, or paid from day one?
Me: Charge unless you have a clear reason not to. Free is a strategy, not a default. What is your cost to acquire a paying customer?
Founder: We are still figuring pricing. Maybe nine dollars per user. Or twenty. Hard to tell.
Me: Do the math backwards. What is your target LTV? What churn can you live with? How fast do you want payback on CAC?
Founder: Can we worry about that after launch?
Me: That is like building a boat and thinking about water later.
Evidence section
Right now AWS is on stage at re:Invent shipping new toys and shaving cents off bills. Microsoft just surprised with Surface Book. Slack keeps growing teams like crazy. The signal is clear. Software that solves a sharp, specific pain and is easy to try wins. The rest is overhead.
Some hard truths from the past few years of building and advising products that charge monthly:
- Pricing is part of the product. Teams that talk to ten buyers and test three price points close more revenue than teams that debate on a whiteboard for weeks. You can always move up if you lead with value. You almost never move up if you start cheap.
- Onboarding drives MRR more than features. A guided first session, smart defaults, and one clear aha moment within five minutes will beat a long feature list every time.
- Hours matter. Replies in under two minutes on chat during trial week double close rates for many tools. Not fancy automation. Fast humans with short answers.
- Churn is a product metric. If you lose more than five out of a hundred customers each month, you do not have a sales problem. You have a fit problem. Fix the core job, not the email copy.
- Expansion is cheaper than acquisition. Growing current accounts by solving more of their day beats buying new clicks. Think add ons, usage tiers, and team seats.
Numbers to keep on the wall:
- MRR and new MRR split into new logos, expansions, and reactivations
- Net revenue churn, not just logo churn
- Payback period on paid channels in months
- Activation rate within 7 days of sign up
- Support first response time during trial
Benchmarks are noisy, so treat them as hints. Still, a simple target that keeps teams honest is this: LTV to CAC near three to one with payback under a year. If you cannot see a line to that, adjust price or change the channel mix.
Implementation notes
Pick one ICP and write copy only for that person. Not for every company on the planet. One job, one title, one use case. You can add more later.
Choose a trial that matches time to value. If your aha moment is in day one, go with a 14 day trial. If data import takes a week, go 30 days. Free plans are fine if they lead to a clear paid path inside 60 to 90 days.
Put pricing on the site with limits that map to outcomes, not arbitrary rows. Number of projects, monthly messages, reports per month. Avoid price by seat unless the product gets better with more people.
Set up one source of truth for revenue. Stripe or Braintree into a simple warehouse, then a clean dashboard. Track MRR movements daily. No spreadsheet gymnastics.
Make onboarding a checklist and drive it by email and in product nudges. Three to five steps. Celebrate each one. Ask for zero meetings unless the buyer asks first.
Sell with content that teaches. Short posts, real screenshots, a two minute demo video with voice, and one customer story that shows the before and after. You do not need production value. You need clarity.
Keep support close to product. Rotate engineers in the support queue a few hours each week. Bugs get fixed faster and feature ideas get sharper names.
Review cohorts monthly. Group signups by month, chart activation and paid conversion. Ask why the best cohorts did better. Copy that pattern forward.
Risks
Platform risk. If your value leans on someone else’s API, have a plan for rate limits, pricing changes, or lockout. Keep a light mode that still works when the hose gets tight.
Concentration risk. One big customer can fund your month and also own your roadmap. Cap their share or split the work so you do not become a custom shop by accident.
Channel risk. Paid search can be a slot machine. Set a hard daily cap and demand payback inside 90 days. Own an email list and an audience you can reach without bidding.
Price risk. Being the cheap option attracts the wrong buyer and makes support heavy. Being the premium option forces taste and quality. Pick one with intent.
People risk. If only one person can ship or sell, you are fragile. Write runbooks, record demos, and pair on calls. Make it boring to be away for a week.
Graceful exit
Today feels like a turning point for builders. Clouds get cheaper, devices get faster, and buyers expect clarity. The teams that win pick a sharp problem, price with courage, and watch a few boring numbers every week. Software as a business is simple but not easy. Talk to customers, charge real money, and keep the first session magical. The rest follows.
If you are choosing between free, freemium, and paid today, start paid with a generous trial. Give people a clear path to success and a fair price for the time you save them. Then show up fast when they have a question. That is the whole trick.