Most mornings I check on my business from my phone before I’m anywhere near a desk. Cash flow, project statuses, a summary of yesterday’s calls, whether a client’s payment went through overnight. My developers check the same app from wherever they are. The app we’re all checking is one my company owns outright, and it runs just about everything.
Allen App Tools runs on a system we built for ourselves. Billing, CRM, proposals, e-signatures, booking, time tracking, the client portal, sales-tax filing, the marketing site, the email sequences. Twenty-two-plus products’ worth of jobs, five core business functions, one codebase, one login. I made a map of everything it stands in for, and this post is me opening the books on it.
The top of the actual map: what the app does in-house, what those jobs rent for. Figures are conservative single-tool retail estimates.
The bill, added up
Add the whole map up and the subscriptions it replaces come to about a thousand dollars a month, priced at a single seat of each tool. That’s the conservative read. The honest read is bigger, because most of those tools charge per person, so the moment a developer, a PM, and a marketer log in, the same pile runs two to three thousand a month. Call it somewhere between twelve and thirty-six thousand dollars a year that stays in the business instead of leaving it.
Run that out five years, the horizon you’d judge any business investment on, and the rent that never gets paid lands somewhere between sixty and a hundred and eighty thousand dollars. The app’s own running bill is the hosting, storage, and AI usage on the rails I’ll show you below, and it is a small fraction of either column.
The part I did not see coming
The app doesn’t just replace tools I was paying for. Most of the map is things I never bought, because I could never justify them. Dunning, the polite chasing of failed payments, rents for about $150 a month, so small shops skip it and eat the failed payments instead. Proposal software that tells you when the client opened the quote. E-signatures with a real audit trail. An AI copilot that drafts from my own project data. A full audit log with scheduled backups. At rental prices, every one of those was a luxury line item I said no to.
In a system you own, each of those was an add-on to something already standing, a room added to a house that was already framed. So my small agency now runs automation that the subscription math usually reserves for companies ten times its size. That was the real surprise: owning didn’t just cut the bill, it widened what the business could afford to do.
What it did to the week
The money is the measurable part. The other change is harder to put on a chart: mental load. Twenty tools means twenty logins, twenty places data can disagree, and a low hum of wondering which one holds the truth. One system means the truth lives in one place. Billing runs weekly on its own. Reminders, installment schedules, and status updates fire from scheduled jobs inside the app, the same jobs that would otherwise be a Zapier bill and a box of duct tape. Nobody got hired to run any of this, and none of it waits for me to remember it.
And because the whole thing rides along in a clean mobile app, the business answers questions wherever I am. My developers pull up their projects, hours, and payouts the same way. Checking on the company stopped being a desk task.
What we kept renting, on purpose
Owning everything was never the goal. The map has a section I’m just as glad about, the rails we rent deliberately: Stripe moves the money, Gmail and Google Workspace carry the mail, Google Cloud hosts it all, the AI vendors sell us inference by the sip, and Slack, Meta, and Upwork feed data in. The app owns all the logic and all the data on top; the vendors just move the bits. That’s the same rule I wrote about in the history of businesses owning their software: rent the commodity, own the operation. This is what the rule looks like implemented.
The rails it rents on purpose, and the whole map totaled at retail prices.
The honest caveats
Here’s the part a fair reader is already thinking: I build software for a living. Building my own stack is my trade, the way a cabinet maker’s kitchen tends to have good cabinets. You should not vibe-code your billing system next weekend, and I wouldn’t hand this app’s job to a prototype either. It also wasn’t free. It cost real building time, and it carries a real care plan, mine, the same kind I sell.
But that’s exactly why I’m showing the books. This is the thing I sell, demonstrated on the business I care most about. Your version would be smaller, because you don’t need twenty-two replacements. Most businesses need the three or four spots where the tools don’t talk and somebody’s afternoons disappear. I laid out the sizes these builds actually come in, real examples from the books, in custom software comes in sizes. What custom software actually costs now has the current price ranges for exactly that. This post is what the top of the menu looks like from the inside.
If you run a business on a pile of subscriptions, you’re normal, and most of the pile is probably fine. But add yours up sometime, multiply by sixty months, and sit with that number for a minute. Then, if you want, book a call and we’ll figure out which part of the pile is worth owning. There’s no rush. The math will still be true next quarter.