The benefits table is where a medical insurance sale is won, misunderstood, or quietly stored up as a future complaint. Design principles for the least glamorous, highest-stakes screen in health insurance.
Every misunderstood benefits table becomes a claim dispute later. That is the whole argument for taking this screen seriously: the table of limits, sub-limits, networks and exclusions is the moment a member forms their belief about what they bought, and every gap between that belief and the policy is a support ticket, a complaint, or a story told to colleagues at exactly the wrong time. In embedded distribution — where cover is chosen in minutes inside someone else's product — the table does the job a broker's explanation used to do. It has to work alone.
Most benefits tables are not designed at all. They are the insurer's filing document, exported to the screen. Fixing that is a product task, and it has learnable rules.
Rule one: structure by member question, not by policy clause
Policy documents are organised for regulators and lawyers: definitions, then benefits, then exclusions, in clause order. Members arrive with situations. Will my daughter's asthma medication be covered? What happens if I need surgery? Can I see a doctor without an appointment letter? A readable table is organised around those situations — outpatient care, hospital stays, medication, maternity, dental, emergencies — with the clause-level detail nested beneath, not leading.
The test is simple: a member with a concrete worry should find their answer by scanning section titles, without knowing any insurance vocabulary.
Rule two: every number needs its consequence
A table row that says outpatient consultation, 20% co-payment, max SAR limit per visit is a fact. It is not yet information. Information is what happens to the member: you pay a fifth of each clinic visit yourself, up to a stated cap. The single highest-leverage edit in benefits design is pairing every limit with a worked consequence in plain language — one line, concrete, in second person. Sub-limits deserve this most of all, because a member who sees a large annual limit will assume it applies to everything, and sub-limits are precisely where that assumption fails.
The same discipline applies to the words themselves. Deductible, co-insurance, reasonable and customary, pre-authorisation — each term either gets a one-line plain explanation at the point of use or it is a comprehension failure shipped on purpose.
Rule three: design the comparison, not just the table
Embedded medical journeys usually present two to four plan tiers. Members do not read four tables; they compare them, and the design should do the comparative work for them. That means aligning identical rows across tiers so the eye can travel horizontally; showing differences prominently and shared features quietly, because the differences are the decision; and resisting the marketing urge to give every tier a superlative. If the honest summary of a tier is cheaper, narrower network, higher co-payments, saying exactly that builds the kind of trust that survives a claim.
One caution from insurance practice: never let the comparison collapse into a row of ticks. A tick that means fully covered sitting beside a tick that means covered to a low sub-limit is the most common single source of benefits-table complaints. If two ticks mean different things, they should not look the same.
Rule four: exclusions are content, not fine print
The instinct is to minimise exclusions visually — they are, after all, the product's bad news. The claims data argues the opposite. Disputes cluster on a predictable shortlist: pre-existing conditions and their waiting periods, maternity waiting periods, dental and optical boundaries, and treatments that need pre-authorisation the member never sought. Surfacing that shortlist plainly at purchase costs some conversion at the margin and saves it many times over in disputes, complaints to the regulator, and renewal churn. An exclusion discovered at purchase is a decision; the same exclusion discovered at claim is a betrayal.
Rule five: the table is a living object
A benefits table rendered as static text — or worse, a scanned PDF — cannot be searched, compared, translated, or read aloud by a screen reader. Structuring benefits as data pays off in every direction at once: members can search their own cover, support agents answer from the same source the member sees, Arabic and English render from one canonical structure rather than drifting apart, and when the insurer revises the schedule at renewal, the change is a diff, not a redesign. In embedded distribution this matters doubly, because the platform, not the insurer, owns the screen — and the platform can only render well what it receives as structure.
The quiet metric
Teams measure benefits screens by conversion, and conversion matters. But the metric that tells you whether the table works is downstream: the rate of claims disputes and support contacts that begin with the words I thought this was covered. A falling curve there means the beliefs formed at purchase match the policy sold — which is the entire job of the screen. It is unglamorous work. It is also, per hour spent, one of the highest-return investments in the whole medical product.