Questions
Build, buy, or assemble a third path?
Buy when the capability is something every business needs and a product already does most of it. Build when it is the reason customers choose you, or when nothing on the market meets a hard requirement. When both are partly true, buy the common core and build the edge that sets you apart, then check the choice against five years of cost rather than one.
The question beneath the question
Ask first whether this capability is where you compete. Payroll, ticketing, and bookkeeping rarely win a customer, so a well-run product almost always beats a homemade one. A pricing engine, a matching method, or a workflow your customers praise may be the business itself, and handing it to a vendor hands away what makes you different.
Ask second whether something on the market does most of the job. If the best product covers 80 percent of your needs and the missing part is cosmetic, buy. If the missing part is the part that matters, buying means bending your business to fit the software, and that cost appears in every workflow for years.
Where the estimates go wrong
Build estimates usually describe the first version and stop there. Software must be maintained, secured, and changed as the business changes, and a common planning assumption is that upkeep runs 15 to 20 percent of the original build cost every year. The engineers doing that upkeep are also not building the product your customers pay for.
Buy estimates fail differently. Licence prices rise at renewal, seat counts grow with the company, and integration takes longer than the sales call suggested. The quietest cost is leaving: if your data can only be exported in a format nobody else reads, the vendor has priced your exit for you.
A worked example
A 60-person company needs support ticketing for 20 support staff. A product costs $50 per user per month, $12,000 a year, plus $15,000 to set up and connect. With a 5 percent rise at each renewal, five years cost about $81,000.
Building it would take two engineers four months, roughly $100,000 at a loaded cost of $150,000 each a year. Upkeep at 20 percent adds $20,000 in each of the next four years, so five years cost about $180,000, more than twice the purchase. Ticketing wins this company no customers, so buying is the clear answer.
Change one fact and the answer turns. If the same company sold scheduling software and its routing logic were the reason customers renewed, building that logic would be an investment in the product rather than a cost of running the office.
How the room weighs build and buy
Build, buy, and an assembled path each get their own Champion, and the five-year sums, renewal rises and upkeep included, are worked exactly from your figures. Craftsman tests whether your team can truly build and run the thing with the people and hours it actually has.
Gray Wolf reads for what the quote leaves out: seat minimums, renewal terms, and fees for data or extra environments. Grandfather asks what each path would cost to reverse in year three. The Verdict states which path leads and the figure that would change it, and the Human Decision is recorded in your own name.
Questions
- Is building ever cheaper over time?
- Yes, at scale. When licence costs grow with seats or usage while the build cost stays largely fixed, the lines can cross; find the year they cross with your own numbers, and ask whether you will still want the same tool by then.
- What about free open-source software?
- The licence may cost nothing, but hosting, security updates, and the people who understand the code all cost money. Treat it as a build that starts further along, and price its upkeep the same way.
- How long a horizon should we use?
- Five years suits most business software, because it captures at least one renewal and the true cost of upkeep. Shorter horizons flatter whichever option has the lower first-year bill.
Bring Argumont a Decision worth challenging.
Start with a paid 60 day pilot on your own real Decisions.