A credit union I looked at could not offer a loan product it wanted to offer. Not because of risk appetite, pricing or regulation. Because of what the loan origination system could represent.
In 2022 I spent a few months researching why more credit unions were not doing solar lending. The market was growing, the balance sheet logic worked, and credit unions had structural advantages: local footprints that matched installer territories, member relationships and payment histories that supported lending to people a FICO cutoff would miss, and decades of experience pricing risk across autos, mortgages, cards and personal loans.
Several were doing it well and building real portfolios. Most were not. And when I got past the surface answers, the reasons had very little to do with strategy.
The specific numbers from that research are dated now, and the incentive structure has changed more than once since. The constraint I found has not changed at all, and it is not specific to solar or to credit unions.
Four barriers, and only one was about appetite.
The product was structurally complex.
Solar loans had evolved into instruments with long durations, dealer discounts, tax credit assumptions and a payment that changed at a set point depending on whether the borrower applied a rebate to principal. Credit union lending, by contrast, is mostly vanilla fixed and variable products with a clear APR, which is a good thing. The mismatch was real, and it was not the lender being unsophisticated.
The origination system could not represent it.
This is the one I keep coming back to. The loan itself was financeable. The system of record could not express it without custom development: a scheduled principal event, an optional recast, a different amortization path depending on borrower behavior. That work sat behind a full slate of other digital projects, competing for the same constrained technical resources.
So the product died in a queue. Not in a credit committee.
The risk type did not fit either bucket.
Secured loans are backed by the asset. Unsecured loans are backed by creditworthiness. These sat in between under a UCC-1 filing, which carries real risk mitigation but requires state and local regulatory expertise to administer. Smaller institutions did not have the legal bandwidth to build that playbook.
The category carried political weight.
Credit unions generally stay on a non-political line, which is easy for autos, homes and cards. Anything energy-related has been politicized from several directions, and for some boards that alone was enough to table the conversation regardless of the economics.
The pattern underneath.
Here is the part that applies whether or not you have ever thought about solar.
Every institution believes its product roadmap is set by strategy. In practice it is set by what the systems of record can represent without a project. If your core system cannot express a product structure, that product does not get launched, delayed, or debated. It quietly never reaches the agenda, because the people who would propose it already know what the build request looks like.
I have watched the same thing in companies with no connection to lending. A services business that could not sell a subscription because billing could not handle it. A manufacturer that could not offer a usage-based contract because the order system assumed a unit sale. A distributor that could not price a bundle because the product catalog was built one SKU at a time.
In each case leadership described the constraint as a market question. It was a data model question.
What to do about it.
Ask what your system cannot represent.
Not what it does badly. What it cannot express at all. That list is your real product roadmap constraint, and almost nobody has written it down. It is usually a short list and it is usually surprising.
Separate the pilot from the core.
The credit unions that moved ran the new product on a platform beside the core system rather than waiting for the core to be extended. A second stack is a real cost, and it beats a three-year wait. It is the same logic I described in how legacy infrastructure kills new ventures: protect the new thing from the system that was built for the old thing.
Share the build when the economics are marginal alone.
Credit unions have a structure for this, the CUSO, which lets several institutions jointly fund a capability none of them would build alone. Most industries have an equivalent, whether a consortium, a shared services arrangement or a partner who has already built it. The question to ask is whether you need to own the plumbing or only the customer relationship.
Count the queue as a strategic cost.
If a viable product is waiting behind twelve other projects, the delay has a price. Put a number on it. That is the argument that gets technical capacity reallocated, and it is more persuasive than any roadmap slide.
The takeaway.
The research question I started with was why credit unions were not lending to a growing market. The answer was mostly that their systems could not hold the product, and the work to change that never won an internal priority fight.
That is a recurring, expensive and fixable pattern. Financial services is where I first saw it clearly, and it is a large part of what our systems work is for: making the platform able to represent the business you want to run next, rather than the one you were running when it was configured.
What can your system not represent?
Book a Call
If you like what you see, we think you’re gonna love what you hear. Book a first consultation with us, and together we’ll figure out how to make your life a little better.
