Software as a Business Part 4: Lessons from the Trenches
Dialogue
Founder: I shipped the beta. People like it. My downloads look decent. How do I turn this into a real business without burning the cash and my sanity.
Mentor: Good. Now forget the launch buzz. Treat this like a shop. Stock, customers, receipts, returns. The web two point oh crowd will cheer your Ajax tabs. Your bank will only cheer deposits.
Founder: So what should I do on Monday.
Mentor: Monday you measure. Tuesday you charge. Wednesday you support. Thursday you prune features. Friday you raise your price. Then repeat.
Evidence from the trenches
Some numbers and scars from building and selling software the past few years across desktop and web apps. Names removed to protect the guilty and the lucky.
- Trial to paid conversion: For paid downloads, a clean site with a single focused product tends to land between 2 and 5 percent. If you ask for an email to unlock trials, plan for fewer trials but warmer leads. For a web app with a credit card after trial, expect 10 to 25 percent of trial signups to pay if the setup takes less than ten minutes.
- Refunds: If you offer a no questions asked refund window, you usually see 1 to 3 percent of orders refunded. Clear feature pages and a real screenshot tour lower returns more than legal text does.
- Support load: Paid desktop apps without online sync see roughly one support email per 100 customers per week. Add sync or import from Outlook and watch that climb. Web apps with a forum and decent help see fewer tickets if the welcome email nudges new customers to a first win within the first day.
- Pricing: The most common mistake is going too low. Moving from 19 to 39 to 59 usually raises revenue while barely moving conversion. Anchors work. A visible higher plan makes the middle plan look safe.
- AdWords spend: For niche business terms you can often buy a customer for 20 to 80 dollars. Consumer terms swing wildly. If your product price is under 30 you will fight to break even on ads alone. Pair ads with a good email sequence and a reminder inside the product.
- Churn on subscriptions: For small team web apps billed monthly, expect 3 to 7 percent of paying accounts to cancel each month until your product becomes part of their daily habit. A weekly value email tied to usage lowers cancels.
- Piracy: Simple keys get shared. Phoning home every start creates angry threads. A friendly license with light checks and clear value from updates keeps paying customers happy and takes the air out of warez drama.
- Trust signals: A physical address, real names, and a clear privacy page beat badges. If you are small, show small. People buy from people.
This lines up with what we see publicly too. 37signals talks about charging real money and keeping products tiny. Joel Spolsky and Eric Sink keep reminding us that pricing is a guess until you test it. On the tool side, Google Analytics is finally usable for free tracking. Website Optimizer started testing doors and buttons without a PhD. On the hosting side, Amazon S3 arrived for cheap storage that does not wake you at 3 am. All of this removes excuses.
How to put it to work
Here is a simple plan to move a product from hobby to business. No silver bullets. Just steady steps.
- Pick a money page: Your pricing page is the cash register. Three plans max. A bold middle plan. A short pitch with one clear benefit. Real customer quotes. Monthly for web apps. One time plus one year of updates for desktop.
- Track the basics: Install Analytics. Create goals for trial starts, downloads, and purchases. Set up daily email alerts when conversion drops. You are blind without this.
- Run one simple test: Use Website Optimizer or your own rotation to test just the headline on the money page. Keep the winner. Move on. Do not chase tiny wins forever.
- Tighten the first five minutes: The first run is your golden window. Preload a sample project. Guide the user to a small win in one minute. Send a short welcome mail with a tip and a link back into the product.
- Charge enough: Start at a price that makes you a little nervous. Add one plan above your target with a team feature. Use yearly pricing for a break that equals about two free months. That locks in cash and lowers churn.
- Make paying easy: For web apps use credit cards with recurring billing through a gateway like Authorize.net or PayPal. For desktop offer PayPal and a card processor like ShareIt or SWREG. Keep the form short. No account required to buy.
- Choose one channel: Do not try to be everywhere. Pick AdWords or content with a blog and guest posts or affiliates. Give it a month of focus. Measure spend to paid users not just clicks.
- Keep a tiny backlog: Say no a lot. Ship small. A weekly release with a public change log builds trust. New customers see motion. Old customers see care.
- Invite real feedback: A link inside the product that captures a screenshot and a note gets gold. Reply fast and short. People pay for software and for the feeling that someone is there.
Risks that can bite you
- Platform shifts: Windows is baking a new security model that will nag apps that write into Program Files. If your installer or updater needs admin rights, fix that now. On the Mac side, the Intel switch is still shaking out. Ship universal builds and test on real hardware.
- Browser changes: IE7 and new Firefox builds break sloppy CSS and some Ajax tricks. Test your app with no toolbars and odd zoom settings. Customers will blame you, not the browser.
- Payment freezes: Payment providers can lock funds without warning if chargebacks spike. Keep a second processor ready and keep cash in your bank not only in their wallet.
- Search swings: Google shifts can bury a site for weeks. Do not bet the farm on one keyword. Build an email list and a feed. Write helpful posts that keep bringing people long after a stumble in search.
- Single point of failure: If you built around one developer, one server, or one vendor, plan a spare. Backups tested by a restore are the only backups that matter.
- Feature creep: Every request sounds reasonable. Ten of them kill the product. Keep your core promise on a sticky note. If a feature does not make that promise stronger, park it.
Graceful exit
Building software as a business is less about grand ideas and more about small steady moves. Charge money. Watch the numbers. Talk to customers. Ship tiny improvements. The web is full of noise right now and that is fine. Real buyers still show up for clear tools that solve one real job.
If you are on the fence about raising prices or trimming features, pick one and try it for two weeks. Your next lesson will not come from a blog post. It will come from your own checkout page.
See you next week. Bring numbers.