Software as a business is not a theory class. It is a kitchen with hot pans, late invoices and release builds that go out five minutes before dinner. This is part five of the series, written after another day of bug reports, support emails and a small win that keeps the lights on.
Today Google said it will buy DoubleClick. Ads are getting tighter and pricier. Vista is on desks. The iPhone is coming. Twitter just had its big moment at SXSW. The ground keeps moving. Still, the rules of making and selling software feel surprisingly steady when you are in the trenches.
What problem do you solve that someone will pay for?
Start here. Not with features. Not with tech. With a problem that hurts and a group of people who already spend money to numb that pain. If your first landing page cannot explain the pain in one line, keep shaving. Make the promise crisp. Make the payoff obvious. People pay to stop hurting or to make more money. Preferably both.
Who exactly is your customer?
Pick a person you can sketch without guessing. A freelance designer with three clients and a messy invoicing habit. A sales manager with ten reps and a spreadsheet circus. Name their tools, their boss, their annoying deadline. Narrow beats vague. If your copy reads like it could fit anyone, it fits no one.
How do you price without painting yourself into a corner?
Charge for the value unit that grows with success. Seats, projects, tracked contacts, indexed files. Avoid a single flat fee that traps you as costs grow. Start with a simple plan that no one needs a calculator to understand. Test prices out loud. If you blush when you say the number, it is probably close to right.
What should you ship first?
Ship the smallest set that makes a full promise possible. Login, one core action, a way to pay, a way to say sorry when things break. Do not chase parity with giants. Pick a narrow job and crush it. Small launch, fast loop. Your first version is a question, not a trophy.
Web app or desktop app?
Ask where trust lives. Teams and shared data love the browser. Heavy offline work and media still like the desktop. Vista is noisy, but it is on millions of machines. The browser is getting stronger every month. Pick where your customer already lives and use the shortest path to them.
How do you get the first 100 customers?
Do things that do not scale. Talk in forums without spamming. Help in mailing lists. Show up in niche blogs with a quick case study. Offer a short free plan that solves a real slice. Ask every early user for one person you should talk to next. Warm paths beat cold ads at the start.
Where do you advertise without burning cash?
AdWords still prints money for Google and sometimes for you. Start with exact match only. Bid on the job to be done, not your clever brand. Track clicks to signups to paid, not just to visits. StumbleUpon can send a flood of bored traffic. You want intent, not boredom. Pay for results inside your funnel.
Is freemium a good idea?
Free is a promise and a cost. It can power word of mouth if your paid plan unlocks a clear step up. Do not give away the core benefit. Give away the try, not the buy. Watch support volume from free users. Free that hurts your paid plan is not marketing. It is a slow leak.
How do you support users without drowning?
Write the help docs you wish you had as a new user. Keep answers short and link to one page that tells the whole story. Use saved replies, but write like a person. Sign with your name. Fix the top three causes of tickets before adding features. Support is product research in plain clothes.
What should you track each week?
Watch signups, activation, paid conversions, refunds, churn. Define activation as the first real use of the promise. Not a login. A job done. Put the numbers in a tiny weekly email to yourself. Patterns beat single spikes. Celebrate boring upward lines. They are how rent gets paid.
Which tools actually help?
Use version control that your whole team trusts. Subversion is fine. Git and Mercurial are knocking, but pick one and move. FogBugz or Trac for issues. Basecamp to keep the chatter sane. Campfire or IRC for quick talk. Tools should lower blood pressure. If a tool adds ceremony, drop it.
How do you pick your stack without chasing shiny?
Rails is fast to build with. PHP still runs most of the web. Python and Django are steady. .NET is strong on Windows shops. Pick what your team can ship with today. Do not fight your own tools. Shipping beats perfect taste. You can rewrite a module. You cannot un-miss a market.
Should you host it yourself or rent?
EC2 is still early but already useful for bursts and experiments. S3 is great for assets and backups. A simple managed server from a boring provider can carry you far. Use a CDN for big downloads if you have the budget. Buy boring reliability until traffic proves you need fancy.
Do partnerships and affiliates help?
They can. Only if the partner touches your exact customer. Set a simple payout and pay on time. Give affiliates honest creatives and landing pages that convert. If they bring floods of junk traffic, pull the plug fast. One great partner beats twenty sleepy ones.
What about refunds and chargebacks?
Write a clear refund policy and stick to it. Be generous and quick. It lowers chargebacks. Track the reasons and fix what you can. Some users will try to game you. Do not design for cheaters. Design for honest buyers and keep the paper trail tidy.
How do you write copy that sells?
Lead with the outcome. Save the tech for later. Show one concrete result in the first screen. Use proof from real customers with real names. Replace three adjectives with one verb. Cut half the words. Then cut another quarter. Short beats clever when the credit card is out.
When should you hire?
Hire when a bottleneck blocks growth, not when you are tired. Start with support or a specialist that clears the founder to build or sell. Contract first if you can. Keep the team small enough to feed with a single pizza. Headcount is a subscription. Treat it with the same respect you expect from your users.
How do you choose features without bloat?
Say no by default. Keep a one page roadmap with problems, not features. When three paying customers describe the same pain in their own words, move it up. Kill features that no one uses. It is okay to disappoint a loud free user. Polish beats breadth.
What do you do when a competitor copies you?
Smile and keep going. Copying is not caring. They can copy features but not taste, support tone or pacing. Ship faster on the things that matter. Keep your promise sharper. Compete on clarity and care, not on checkbox races.
How do you handle security and trust?
Use SSL everywhere users expect privacy. Hash passwords. Rotate keys. Back up daily off site. Keep a short page that explains your approach in plain words. When mistakes happen, say what broke, what you did, and what you will do next. Trust is earned in bad hours.
What legal basics do you need?
Have a short terms page, a privacy page, and a fair license. Do not surprise your users. If you collect data, say why and for how long. Use real contact info. If you do email, honor unsubscribes fast. Clarity keeps you out of trouble more than clever clauses.
Should you take funding or bootstrap?
Money buys time and sometimes speed. It also buys new bosses. If your market is narrow and your sales are simple, customer money is the best kind. If you need a big push to get to any value at all, then a check might make sense. Match fuel to the trip.
How do you use press without going on a tour?
Reach out with a short note that shows the story, not just the release. Give a crisp hook and a real quote from a customer. Screenshots beat paragraphs. A niche blog that your buyers read is worth more than a glossy writeup that your mom reads. Press is a boost, not a plan.
What about email marketing?
Own your list. Ask for it with a clear promise and keep that promise. One helpful tip per week beats a promo blast once a quarter. Use tools like MailChimp or Campaign Monitor and keep your sender name human. Email is a quiet compounding channel.
Can you raise prices?
Yes, if you earn it. Add a new tier or increase value first. Announce it with time to switch and a thank you rate for current customers. Do not apologize for running a real business. Price is part of your message. Cheap can be scary.
How do you keep focus when shiny things pop up every week?
Make a weekly plan on one card. Three outcomes max. Close chat, code, ship, talk to customers, repeat. Twitter will still be there tomorrow. Your build will not ship itself. Protect the hours that produce money.
What is the real moat?
It is not code. It is not patents for most of us. The moat is your pace and your taste. It is how fast you turn feedback into fixes. It is your support tone. It is your steady hand when a big player swings in your lane. Culture is a quiet moat.
What does the DoubleClick deal mean for small software shops?
Clicks will get pricier. Tracking will get tighter. Do not bet the farm on cheap display inventory. Spend ad budget where intent is hot and post click pages work. Grow channels you own. SEO, email, product led invites. If you must rent attention, negotiate small and test everything.
Where do you find repeatable growth?
When one channel buys three new customers for every hundred dollars, do it again next week. When a channel buys one, cut it. Repeatable growth looks boring. Same ad group. Same anchor post. Same partner. Boring is the drumbeat of a healthy software business.
What mistakes did I earn the hard way?
I shipped a giant feature no one asked for. I wrote copy for my peers, not for buyers. I delayed a price increase for months and paid the rent in stress. I learned to release small, talk to customers weekly, and charge for real value. Tuition is paid in refunds and late nights. Learn and move.
What should you do next week?
Pick one metric to nudge. Improve onboarding by removing one step. Rewrite the first screen with a stronger promise. Call three customers and ask what almost made them quit. Ship one fix that cuts support tickets. Next week is where the business grows.
What sticks after the dust settles?
Solve a clear pain. Charge for the value unit. Ship small and often. Talk to customers. Keep expenses boring and your product sharp. Use ads with a leash. Own a list. Say no a lot. That is the game. Tools will change. These rules feel like they will not.