Stop Bending Your Business to Fit Your Software

Think about how people wear a suit that almost fits. They don't return it. They learn to stand in it, a little differently, for years.

I see the same habit in operations. A company buys a platform built for the typical lender or property manager, and then its people learn to stand in it. They export, re-key, reconcile and chase. Nobody calls it a cost because nobody gets a bill for it.

The gaps are where the money is

After three decades in mortgage lending, fintech and real estate, I've stopped looking for savings inside the big systems. Most of them do their core job well enough. The waste lives between them.

It's the processor typing an application into a second system. It's the servicing analyst matching the same payment in two reports every month. It's the web lead that sat in an inbox over a long weekend because no one owned the step between marketing and leasing. And it's the exception file that eats an afternoon because the software only knows the normal case.

Each of these looks small on its own. Add them up across a year and a team, and you're looking at real margin.

What fitting the process looks like

The best example I have is TurnKey, a system I wrote for a public company client that operates manufactured housing communities. It began as our HomeBase platform at Millwright Logic.

The operator's sales process didn't fit any single product. Some prospects want to buy with a home loan, some only want to rent, and some haven't found the right home yet. So the application handles all three, including a rental-only path and a "don't see the right home" option. State-specific forms, such as Pennsylvania's, are built in.

Behind the application, the work that used to be manual is wired together. Loans go straight into MeridianLink. Each applicant gets a ShareFile folder and a no-login upload link, and that link is written back to the loan. Tenant screening runs through Settlement One, insurance quotes come from Assurant's API, and Rent Manager is connected on the property side. Social Security numbers are masked, and the whole application has been hardened for security.

What I want people to notice is what we didn't do. We didn't replace their loan system, their document platform or their property management software. We built the piece that was missing and let each system do what it's good at.

Buy what's common, build what's yours

I'm not against packaged software. I buy plenty of it. Accounting, payroll, a loan origination system, core servicing: these are mature, and rebuilding them is usually a mistake.

Custom earns its place where your process is different, where it touches the customer, or where your people are clearly working around a tool every day. It's also worth waiting if the process is still changing or nobody can describe it clearly yet. Write it down first.

Start with one step

You don't need a transformation program. Pick the workaround your team complains about most. Get a rough sense of what it costs in time and errors. Build the smallest fix, run it for a month, and look at the result. Then do the next one. Small wins build trust with the team, and the team is who decides whether a new tool actually gets used.

That's how a good suit gets made, too. Measure, cut a little, fit, adjust. The goal is something you stop noticing because it just works.

If you're running a lending, servicing, real estate or manufactured housing operation and you suspect your team is standing a little crooked to fit its tools, I'm always glad to compare notes.

By David Allen Wind, founder of Millwright Logic and DAW Partners.