A product covers most of the process. The rest lives in spreadsheets beside it. Whether to keep paying for that gap or close it is a cost question, and it has a specific shape.
The usual comparison is licence cost against build cost. That favours buying every time, because the licence is a known monthly figure and the build is a quote.
It leaves out the expensive half. A product that fits eighty percent of the process carries the other twenty percent as human work: the spreadsheet maintained beside it, the manual step between two screens, the report rebuilt each month because the built-in one cannot express the question.
Count the hours. How many people touch the workaround, how often, and for how long. Multiply by a loaded salary rate, annualise it, and put it in the buy column where it belongs. That figure is often larger than the licence.
Then count the risk the workaround carries. A process that depends on one person and one spreadsheet has a failure mode that a product outage does not: it walks out when they resign.
Anything every business does the same way. Accounting, email, storage, and helpdesk ticketing are solved products, and building any of them is a decision to spend money reproducing something you can rent.
Anything regulated where the vendor carries the compliance burden and updates it for you. Anything where your volume is small enough that a per-seat licence is simply cheaper than any engineering.
The part of the operation that makes you different from your competitors. A general product cannot cover it, because a general product is built for the average of a market and your advantage is where you deviate from that average.
Also: where licence costs scale with something you are trying to grow. Per-seat pricing on a system every employee needs turns growth into a cost problem, and at some headcount the build pays for itself.
Most operations end up buying the commodity layer and building the connective tissue and the differentiated part. That is a normal, healthy architecture rather than a compromise.
What makes it work is deciding deliberately which side of the line each piece sits on, and integrating across the line properly rather than with a person and a spreadsheet.
Checklist
It has a maintenance cost that a licence hides rather than removes. The honest comparison puts a running cost on both sides, because a bought product also needs configuration, upgrades and somebody who understands it.
They sit between the two and carry traits of both: faster to a first version, and a licence that scales with use. They suit internal tools with modest logic. They struggle where the process is complex or the volume is large.
Own the code, keep it on a mainstream stack, and document it as it is built. Those three make it possible for another team to take it on, which is what lock-in actually means.
Describe what is slowing the operation down. We will come back with the approach, the scope, and what it costs.
Get in touch