Obviously there is a cost to creating enterprise architecture (EA) blueprints and models of data, rules, processes, events/schedules, strategy/goals/objectives, policies/procedures, rules, decisions, people, places, things, etc.
Doing EA is expensive. The good news is that EA is optional.
The bad news is that not doing EA is going to be even more expensive later!
Your company must decide which path it will take. Executives need to decide whether or not to do EA. It's basically a simple decision: Think long-term or short-term? Pay me now or pay me later?
The question of whether to do EA boils down to these three questions:
1. Does the company want to build products that fit, connect, work as intended, last, are shared, are reused, and change?
2. Does the company truly want or need interoperability, integration, quality, alignment, reusability, and flexibility?
3. Does the company want reliability, compliance, reduced time-to-market, and reduced costs?
Here are some “rules of thumb” for deciding whether to do EA:
If the system you are trying to build or maintain is not complex, then you really don't need to do EA.
If your company wants to build complex products and services so they work, change, and last, then you must do EA. Do architecture (analysis*), then engineering (synthesis*/design), and finally manufacturing (building/coding). By relying on engineering models and architectural blueprints to build "instances" of solutions, you could say that your company is in the engineering business.
If your company dives in, starts building systems, and decides to skip architecture and engineering, then it is already decided; you are not doing EA. What you're really doing is building solutions that can't be instantiated, replicated, shared, or reused. If you're lucky these systems may work, they may change, and they may last. But we all know they will take too long to build, they won't work very well, they won't work as planned, no one will really know how they work, and they will take forever to change. By day two, the systems will already be legacy solution. Sooner or later the systems will fall apart like a house of cards*. Your company is not in the engineering business; you are in the manufacturing business building "one off" or point-in-time solutions.
Enterprise architecture is optional. If you want to stay in business, do EA.
* analysis - taking things apart
* synthesis - putting things togeter
* like Lehman Brothers, Sept. 25, 2008