CRM you own vs seats you rent: what changes after year one

Two systems we run our own companies on

We are not describing this from a vendor demo. Two of our own businesses run on software we built, and we live with the consequences of both.

TailorManager runs Imperial Alterations across two locations. It replaced a desktop point-of-sale. Under it: 70 database tables, roughly 200 endpoints, 21 scheduled jobs, and 7 distinct interfaces — admin, CRM, production board, scanner, shop floor, kiosk, and customer tracking. Migrating off the old system meant moving 13,924 historical orders, which matched the new catalog at an 85.9% rate. The remaining 14% was hand work. Nobody tells you about the 14%.

BridalFlow runs Ira’s Bridal Studio: 123 database tables, 59 admin pages, 30 public pages, 27 scheduled jobs. Across both platforms we integrate Stripe, Twilio for SMS and voice, QuickBooks, Square, Microsoft Graph for email, thermal receipt printing, and web push. Holding it together: 3,268 automated tests.

Those numbers are not a brag. They are the honest size of the thing. If a vendor tells you a real operating system for a two-location business is small, they have not built one.

The difference nobody puts in the comparison table

A rented tool answers the questions its vendor anticipated. That is not a criticism — it is the business model. A vendor serves thousands of customers, so the reports are the intersection of what most of them want. Your specific question is, by definition, outside that intersection.

Here is what that costs, concretely. On our own bridal storefront we found that 139 product pages had earned zero search impressions in 90 days. Not low. Zero. Every page returned a healthy status code, the sitemap was accepted, and no dashboard flagged anything.

The only reason we found it is that we owned the site, the platform, and the analytics together, so we could join them and ask one unusual question: has this specific page ever been shown to a single human being? No rented stack we know of offers that as a report, because it is a question about the seam between three tools, and each vendor only owns their side of it.

Ownership is not about features. It is about the ability to look.

What ownership actually costs

We would rather you hear this from us than find out in month four.

• It costs more up front. Considerably. A seat-based tool bills you a little every month forever. A build bills you a lot once, then a little. The crossover is real but it is not in week three.
• You own the bugs. When a rented tool breaks, you file a ticket and complain. When yours breaks, it is yours. That is why 3,268 tests exist — not because we enjoy writing them, but because ownership without a test suite is just a liability with your name on it.
• Migration is uglier than the pitch. An 85.9% automatic match on 13,924 orders is a good outcome. It still left about 2,000 records for a human. Budget for that explicitly.
• Integrations do not stop moving. Stripe, Twilio, QuickBooks, Square, and Microsoft Graph all change on their own schedules. Owning the system means owning those upgrades too.

When you should just rent

If your process is genuinely standard, buy the standard thing. Off-the-shelf is the right answer more often than an agency that builds software will usually admit.

Standard means: your workflow looks like everyone else’s in your industry, the questions you ask of your data are the questions the vendor already answers, and you are not paying per-transaction fees on a volume that makes the percentage hurt. If that describes you, a build is expensive vanity.

The case flips when the seams are where your money is — when the thing you need to know spans booking and production and payment and marketing, and no single vendor sees all four.

What we would check in your stack

• Write down the last three questions you asked and could not answer. If all three cross a boundary between two tools, that is your case for owning. If all three were answerable and you just did not look, buy a report instead.
• Add up your per-seat and per-transaction fees over 36 months, not 12. Include the seats you will add as you grow. That is the honest number to compare a build against.
• Try to export everything. Not a report — the raw records with their IDs and timestamps. How hard that is tells you exactly how much of your business you currently control.
• Count the tools that hold one piece of a single customer journey. Every boundary is a question you cannot ask.
• Ask whether your process is actually unusual. Be honest. Most are not, and for those, renting wins.