Custom software guide
Is it cheaper to build or buy software over five years?
Whether building or buying is cheaper over five years depends on how many users you pay for, how fast the subscription price rises, and what the build costs to make and run. Enter your own numbers below and the calculator shows both running totals year by year and where the lines cross, if they do.
The platform you would buy
The system you would build
Five years of cost, side by side
The platform numbers the calculator asks for
Users today is the number of people you pay a seat for now. Yearly user growth is the percent you expect that number to grow each year, and it compounds, so ten percent a year takes 100 users to about 146 by year five.
Price per user per month is the list or contract price for one seat. Yearly price increase is the percent the vendor raises that price at each renewal, and it compounds the same way.
Add-ons and extra modules per year covers features sold on top of the seat price. Admin staff time per year is what you pay people to configure the platform, manage users and work around its limits. Integration and connector costs per year covers paid connectors, middleware and the work that keeps them running. One time implementation cost is the setup, configuration and data load in year one.
The build numbers the calculator asks for
One time build cost is the quote you have, or the figure you expect, to design and build the system. One time migration cost is moving your data and people off the current tool. Both land in year one.
Yearly hosting and infrastructure covers servers, databases, storage and the services the build runs on. Yearly support and maintenance covers fixes, security updates and keeping it running. Yearly change and improvement budget is what you plan to spend adding to it as the business changes.
How the calculator adds up five years of cost
For each year the platform cost is that year's user count, times that year's seat price, times twelve, plus add-ons, admin time and integration costs. The user count and the seat price each grow by their own yearly percent from year two. Implementation is added once, in year one.
For each year the build cost is hosting plus support plus the change budget. The build quote and the migration are added once, in year one.
Each running total adds the year to the years before it. The build becomes cheaper in the first year its running total falls below the platform running total. If that never happens in five years, the platform is the cheaper route on these numbers.
Amounts are in whatever currency you enter, and the math runs in your browser. Nothing you type here is sent anywhere.
What this build or buy model leaves out
Staff training on a new system, and the hours people spend learning it, are not in either column. Neither is downtime or slower work during a migration, or the opportunity cost of the money and attention a build takes up front.
The model also leaves out what a process that fits the software is worth, and the cost of working around a platform that does not fit. Either can outweigh every figure here, so treat the result as one input to the decision.
Questions buyers ask
How do yearly price increases change the build or buy math?
A yearly price increase compounds with user growth, so the platform bill in year five can sit well above year one with no new features. A build has no seat price to rise, so its yearly cost moves only with hosting, support and the change budget you set.
What costs does a five year build or buy comparison usually miss?
Staff training, slower work during a migration and the opportunity cost of paying for a build up front are the usual gaps. Admin time spent working around a platform is often missed as well, which is why the calculator asks for it as its own line.
Can a platform stay cheaper than a custom build for five years?
Yes, and often for a small team. A low seat count, a modest seat price and slow price increases can keep the platform running total below the build for all five years, and the calculator says so when that happens.
Should a company build software only because the numbers say it is cheaper?
Cost is one reason among several. Fit to process, ownership of the code and data, and who keeps the system running matter as much, and a build that wins on cost but fails on fit is not a saving.