I have built three BA practices across three major Canadian insurers. Each one survived. That is less common than you would think.
In my experience, most BA practices are set up to have a good first year. The hiring is done well, the intake process gets documented, the first few deliveries go smoothly. Leadership is satisfied. Then year two arrives and things start to drift. By year three, what started as a structured practice is just a collection of individuals doing their own thing.
The failure is rarely about talent. It is almost always about structure.
The most common mistake is treating the BA practice as a staffing layer rather than a governance function. When a practice is built around headcount and assignments rather than standards and accountability, it will not hold. The moment business pressure increases and resource constraints hit, consistency disappears.
The second mistake is building an intake process with no teeth. An intake model that exists in a Word document but is not tied to actual gates or decisions is decoration. I have seen practices where the BA intake process was technically in place but ignored for every urgent request, which is to say, most requests. After six months, no one used it at all.
The third mistake is investing in methodology without investing in how that methodology gets passed on. A practice that relies on one or two senior people to carry the knowledge is fragile. When those people leave or move into management, the knowledge goes with them.
Year two is when the first generation of processes meets the second generation of work. New projects come in that do not fit the original templates. New stakeholders show up who were not part of the initial setup. New team members join who were not trained by the founding group.
If the practice was built on documentation alone, it buckles under this pressure. If it was built on shared standards with a mechanism for updating them, it tends to hold.
The practices I built that survived all had one thing in common: they had a governance model that was separate from delivery. Someone was accountable for the practice itself, not just the projects it served. That person spent time reviewing how work was being done, not just what was being delivered.
An intake model has to do two things. It has to filter incoming work, and it has to shape how that work gets started. Most intake models do the first thing. Very few do the second.
When an intake model only filters, it becomes a queue manager. The BA gets assigned, the BA does what they do, the output depends on that individual. When an intake model also shapes how work starts, it ensures that every BA begins an engagement with the same orientation to the problem. Stakeholders are identified the same way. Business objectives are articulated with the same rigor. Scope boundaries are established with the same consistency.
The intake model is not the front door of the BA practice. It is the foundation. If it does not set standards, nothing else will.
At Equitable Life, I built an intake model that included a mandatory requirements orientation session with the business lead before any BA work formally started. That session covered scope, timeline, decision rights, and what the business was actually trying to change. It took forty-five minutes. It prevented weeks of rework.
Most organizations treat capability development as a training budget line. Send people to a course, check a box, done. That is not capability development. That is continuing education. The two are not the same.
Capability development means building shared language, shared tools, shared ways of approaching problems. It means creating situations where BAs learn from each other, not just from external sources. At Manulife, we ran monthly knowledge-sharing sessions where BAs presented how they had approached a specific challenge. Not training, not methodology updates. Just practitioners showing other practitioners how they thought through a problem.
That practice outlasted my time there. I know because people who came to work with me years later had been shaped by it.
Three things.
First, ownership that is not tied to delivery. The BA Manager role cannot be a senior BA who also carries a full project load. If the person responsible for the practice is buried in delivery, the practice will be managed in whatever time is left over, which is none.
Second, a feedback loop. The practice needs a regular mechanism for reviewing whether the standards are working. Not a retrospective on projects, but a review of how the practice itself is functioning. Are intake templates being used? Are elicitation approaches consistent? Are newer BAs getting the guidance they need?
Third, executive acknowledgment that the practice is a function, not a resource pool. This is the hardest one to get. When leadership sees BA as a capability to tap rather than a function to invest in, the practice will always be under-resourced and over-demanded. Changing that perception usually requires demonstrating business value through delivery, which is a long game.
Building a BA practice that lasts is not complicated. But it is disciplined. The organizations I have seen do it well all made the same choice: they treated the practice as something worth maintaining, not just building. That decision, made early and consistently, is what separates the practices that are still running from the ones that quietly dissolved into something unrecognizable.