Same plan, two rulebooks
When a vendor sets up your program, it uses its rules, not yours. The gaps show up on day one.
Every vendor implementation starts with the same promise: we’ll configure it to match your plan. And every one ends with small differences nobody signed off on.
Not because anyone was careless. A vendor’s platform has defaults, and a plan document is written in a different language than a configuration screen. “Receipts due within 60 days” becomes a setting called claim_window that ships at 90. “Cards for dependents” becomes spouse-only, because that’s how the vendor’s card program works.
Each gap looks minor in a spreadsheet. Each one lands on a member. A limited-purpose FSA set up as general-purpose can cost an HSA member their eligibility. A replacement card fee the plan never approved becomes a member complaint.
The same gap shows up in every program that runs through a vendor:
An operations leader with nearly 40 years on the payer side put it plainly: “Both [rules] exist… the system is built with this definition of the rule, and this system’s built with this definition of the rule.” Her favorite example is a bed day. A claim system counts midnights; a clinical system counts hours. “It’s those silly little differences between the data dictionaries in different applications that make the integration so hard.”
The root problem is that the plan’s rules and the vendor’s settings live in two places, owned by two companies, and nothing compares them. The comparison happens in someone’s head during a testing call, if it happens at all.