Loyalty glossary · 7. Data and analytics (31)
First Party Data
First party data is information a loyalty programme collects directly from its own members through their interactions, with their consent, rather than buying or inferring it from outside sources.
Most loyalty programmes do not have first party data. They have transaction records, which are a different thing. A transaction record says what a member did. First party data says what a member said they want. The first is an accounting byproduct. The second is a preference signal.
Accrual accounting treats points as a liability and does not need preference data. Activity-based qualification treats a stay or a transaction as a trigger, not a signal. An actuarial model pools members into cohorts and discards individual differences. None of these mechanisms produces first party data, and none should be mistaken for it.
First party data is perishable. A preference stated 3 years ago is not first party data today. Consent expires when members become inactive or change their minds. Maintaining first party data costs real money, and most operators refuse to spend it.
Work the arithmetic. A programme issues 1 billion points in a year and spends 0.8 cents per point on collecting first party data, a total cost of 8 million dollars. If that data reduces breakage by 5 percent, it releases an additional 400,000 dollars from the liability line. The numbers only work if the data is fresh and the consent is current.
The trap is treating first party data as a free byproduct of the programme. It is not free, and it is not a byproduct. Operators then use it to justify programme changes that members never asked for, which pushes quiet disengagement. The only first party data worth collecting is explicit preference data with a defined shelf life. Everything else is an actuarial assumption wearing a data costume.