The Buy Versus Build Spectrum
"How can a build versus buy decision be a spectrum?" one might ask. An individual decision to build or to buy is, after all, just one decision. However, as organizations grow and change, these buy versus build decisions happen more often. What's usually true with each decision, is that any given system, either bought or built, needs to integrate with some other system somewhere. That "integration" may be as unsophisticated as manual data entry by a clerk, retyping all the information from one system into another. It may be a nightly ETL job. It may be a near-realtime API integration. For two systems, the upkeep and management of the "integration" is almost trivial, and requires very little oversight and management. For a dozen systems, this can be troublesome.
ESB vendors will gladly jump into this space and declare, "we have a solution for this!" as though it's just a technology problem to solve. This isn't just a technology problem. Yes, the technology that underpins the communications between each of these systems is vital, but there's a core issue that must also be addressed. The content of each integration — the data — is the lifeblood of your organization. The nature of these integrations is called enterprise architecture. The enterprise architecture represents, at a high level, the bloodstream and nervous system of your organization. How do vital bits of information get from one place to another? How are errors handled? What are the key metrics that we need to observe and respond to?
Every buy vs build decision influences the enterprise architecture. Many organizations make too many of these decisions without considering enterprise architecture, and discover that their IT systems interfere with their operations instead of empowering them. The more systems that are bought, the more complex the enterprise architecture becomes. The more systems that are built, the greater the risk of creating a monoculture where innovation can't enter. Buy versus build is a spectrum that weighs complexity, rigidity, quality, time to market, and dozens of other factors. Like many spectrums, it's important to find a balance, and the right balance is different for everyone.