On the floor of every Toyota plant runs a length of yellow cord, suspended within arm's reach of every line worker. It is called the andon cord. Pull it, and the entire production line stops. Not a section and not a station, the whole line.
The economics of the cord look insane on paper. Stopping the line costs thousands of dollars per minute in lost output. Toyota built its global reputation on running that line with relentless efficiency, and yet, by deliberate design, the company hands every worker the authority to halt production the moment they see a defect.
Toyota figured out something the rest of manufacturing took decades to accept. A defect caught at station 12 costs roughly $50 to correct. The same defect caught after the car ships costs $5,000, plus the recall, plus the brand damage, plus the customer who never buys a Toyota again. The cord is not a cost; the cord is the cheapest insurance policy in the building.
A new member opens a checking account. Day 14 arrives, and there is no direct deposit. Day 21 arrives, and the debit card has not been swiped. Day 30 arrives, and the mobile app has been opened twice and never since. Every one of those signals is a defect on the chassis, and each is showing up at a workstation where someone, in theory, could pull a cord.
Nobody pulls the cord.
The line keeps moving. The account sits on the books, counted as a member, reported to the board, and celebrated in the growth column. Nine months later, it either closes or lingers as a share account that runs at negative margin while incurring statement printing, fraud monitoring, and compliance overhead. The defect is now in the customer's driveway. The cost to fix it is no longer $50.
2025 NCUA data shows that credit union membership grew by nearly 2.9 million accounts over a 12-month period. (NCUA, 2025) Industry benchmarks indicate dormancy rates of 44-50% for new checking accounts within the first year. (Cornerstone Advisors and Alkami, 2025) With member acquisition costs exceeding $400 per account, the math is straightforward and unflattering: the industry is spending roughly half of its acquisition budget on accounts it will lose before the relationship yields any return. That is not a marketing problem; it is a system that does not stop when the chassis comes off the line crooked.
In most credit unions, the signals that would trigger an intervention live in different systems than the people who could act on them. Direct deposit data sits in the core, app activity sits in the digital banking platform, card usage sits with the processor, and marketing automation lives somewhere else entirely. The information is technically present somewhere in the building, but it is not assembled into a single view that says, "This member is going dormant, and here is the action that needs to happen this week.”
Toyota's cord works because it is physical, visible, and tied to a defined response that activates the moment it is pulled. The worker pulls the cord, and within sixty seconds, a team lead is standing at the station, diagnosing the defect, deciding whether the line resumes or the part gets replaced, all without a committee forming, a quarterly review convening, or an analytics request landing in a shared inbox that may or may not be opened by Friday.
The translation for a credit union is not a literal cord but a small set of early-warning signals tied to a defined response for each one, and three signals are enough to start:
- No direct deposit by day 14
- No card transaction by day 21
- No mobile or online banking login by day 30
Each signal should trigger a specific, pre-built intervention, whether that is a text, a call, or a digital prompt waiting for the next time the member logs in, and the action does not have to be elaborate, so long as it is automatic and reaches the member while they are still on the line.
The objection most credit union leaders raise when they hear this is that intervention sounds expensive, manual, and impossible at scale. That is the same objection Toyota's competitors raised in the 1970s. The answer is the same now as it was then: the cost of catching the defect early is always lower than the cost of catching it late, and the work of catching it early can be designed into the system rather than bolted onto someone's existing job.
Matt Fehrmann, Chief Information Officer at Kohler Credit Union and Swaystack client, put it directly when his team described their own engagement approach to American Banker earlier this year: "People want engagement, and they want simple engagement." The phrase matters because of the word that sits between the two clauses. Members are not asking for less engagement. They are asking for engagement that does not require them to do the institution's work. (American Banker, 2025)
That is the core of what Toyota understood. The line worker does not pull the cord because they want to slow the factory down; they pull it because the system has made the right action the easy action. The institution absorbs the operational discipline, so the worker can absorb the simplicity. For a credit union, that means engagement is not a series of campaigns aimed at members. It is an internal system that detects the right moment, decides the right intervention, and acts before the member has to ask for help they would never think to request.
The institutions that have begun building this kind of system share a few characteristics: the signals are defined in advance rather than interpreted after the fact, the response to each signal is pre-built rather than assembled on demand, the data flows are connected so that a card processor signal can trigger a marketing automation action without three meetings and a vendor ticket, and the metric that matters is not how many touches went out but how many defects were caught before they left the line.
This is the part of the work that doesn't photograph well because there is no logo, no campaign theme, no tagline, and no headline moment for a board deck. Internally, it looks like a list of conditions paired with a set of automated responses, governed by a small cross-functional team that meets weekly to review what tripped, what got resolved, and what slipped through, which is dispatch logic rather than marketing.
It is also the part of the work that compounds, because every defect caught at day 21 becomes a member who funds the account, sets up the deposit, keeps the card in their wallet, refinances the auto loan, and refers a family member, while every defect missed at day 21 becomes a member who shows up nine months later as a closure code and a write-off of the original acquisition spend.
Toyota did not invent the andon cord because they were sentimental about quality. They invented it because they ran the numbers and concluded that the cost of stopping the line was a rounding error compared to the cost of shipping a defect. The cord paid for itself in the first month it was installed.
Credit unions have the same calculation in front of them. Every dormant account that closes in month nine started giving off signals in week three. Every member who has not set up direct deposit was told this on day 14. The information is already in the building. The only question is whether anyone has the authority, the system, and the institutional permission to pull the cord while the chassis is still on the line. If your institution is ready to build that system, let's talk.
