The integration is the product
The screens are the easy fifth of the work. The rest is making your systems agree with each other — and that part decides everything.
Nobody wants an app. What they want is for their systems to agree with each other: the order in the store, the shipment with the carrier, the payment in the ledger, the forecast in the plan — all telling the same story, without a person retyping parts of it in between.
Software that makes that happen looks, in a demo, like it is mostly screens. It isn’t. The screens are the easy fifth.
Count conversations, not screens
Before believing any estimate, count how many systems the new thing must talk to, and how many systems those must talk to in turn. Each conversation has a cost that outlives the build: the vendor’s interface has limits, schedules, and outages. The accountant’s export format is what it is. Authentication with the customer portal is negotiated with someone who answers email weekly.
None of that shows in a screenshot. All of it shows in the bill and the timeline.
Speaking languages you don’t own
Your software’s neighbours each have their own dialect, their own opening hours, and their own opinions about change. The work is not in the screen — it is in the handshake: the retry when the warehouse system is down at month-end, the record that arrives twice and must be recognised, the field that means dollars in one system and cents in another.
This is also where the fragility concentrates. A screen that is wrong is an annoyance. A handshake that is wrong is a wrong number, in a system people trust, discovered late.
Why it’s worth it anyway
Integration is the difference between software that helps one person and software that runs an operation. A tool that stands alone must be fed by hand, watched constantly, and trusted provisionally. One that talks to its neighbours feeds itself: data arrives, records stay in step, and the person who used to do the retyping does something only a person can do.
So when a proposal arrives, count the conversations before counting the screens. The screens are what you see in the demo. The conversations are what you get in production.
Working on something this applies to? Tell us about it — thirty minutes, no pitch.