The organization that learns
A few years ago I wrote two articles about digital transformation. The first borrowed an image from MIT’s George Westerman that I have never managed to put down: done right, transformation turns a caterpillar into a butterfly; done wrong, all you have is a really fast caterpillar. The second argued that the place to start is your processes, because processes are the connecting tissue of an organization — they bring together people, data, technology, and policy in a flow enveloped by time.
In the years since, most institutions took that journey seriously. Processes were digitalized. Onboarding that took twenty days takes minutes. Customers moved to the glass, and organizations reorganized behind it.
And in doing so, they built themselves a new silo. This one is harder to see, because it doesn’t look like a legacy system or a paper form. It looks like success.
The new silo is analysis
Walk through any well-run financial institution today and count the places where analysis happens. The lending team has its reports. The deposit side has its dashboards. Finance has its models, marketing has its segments, risk has its scorecards. Each one is competent. Each one answers the question it was asked.
Now ask a question that crosses them — the kind a board actually asks. Why did deposit growth slow in that branch network, and what should we do about it? The data exists. The analysts are capable. And yet the answer takes weeks, arrives as a document, answers exactly what was asked and nothing around it, and is out of date by the time it shapes a decision.
In one 2025 industry survey, 76% of organizations admitted making business decisions without consulting data they already had — because it was too hard to access (Sisense). Read that carefully. Not data they lacked. Data they had.
We did not fail to invest in analytics. We invested magnificently. What we built, without meaning to, is a hundred small rooms where insight is produced — and no corridor between them. Analysis became the new silo, and it happened while we were celebrating how much analysis we finally had.
Knowing is not learning
Here is the distinction I keep returning to. An organization knows something when the fact exists somewhere inside it. An organization learns something when the next answer is better because of the last one.
By that standard, most institutions do not learn. They re-discover.
The question asked in March gets answered in March, presented in April, and forgotten by May. When a related question arrives in September, the work starts again — often from scratch, sometimes by a different analyst, occasionally reaching a different answer from the same data. And when the analyst who held the context moves on, the context moves out with them. The institution knew. The institution did not learn.
Peter Senge described the learning organization more than thirty years ago: an organization that continually expands its capacity to create its future. What struck me rereading him recently is how much of that vision assumed the learning would live in people. It largely still does. Which means institutional memory has the shelf life of a tenure, and institutional reasoning has the durability of a meeting.
Three questions to test your institution
If you want to know whether your organization learns or merely knows, three questions will tell you:
When your team produces a good analysis, does it persist anywhere a future question can find it — or does it live on as a PDF in someone’s sent folder?
When you decided something significant last year, did that decision — its reasoning, its evidence, its outcome — make any subsequent decision easier? Or did it simply happen?
Pick an important number your leadership acted on last quarter. Could someone today show where it came from, how it was computed, and what it touched?
Most institutions, honestly assessed, fail all three. Not because their people are careless — because no layer of the institution has ever had that job. We built systems to record transactions, systems to produce reports, and nothing whose purpose is to make the institution’s answers accumulate.
Why this matters most in member institutions
For a credit union, this is not an efficiency question. Member relationships are measured in decades. Decisions about pricing, risk, and member experience compound — each one shapes the conditions of the next. An institution whose understanding accumulates will make each year’s decisions on a richer foundation than the year before. An institution that re-discovers will spend that same year paying, again, for knowledge it already bought.
The first wave of transformation connected our processes. It was necessary, and it worked. But it optimized what the institution does, and left untouched how the institution decides. That is the next frontier, and it will separate institutions in the coming decade the way digitalization separated them in the last one.
The caterpillar became a faster caterpillar, then finally a butterfly. The butterfly’s next act is not to fly faster. It is to remember where the flowers were.