What is a KVI (known value item) and why it is more politics than data

A KVI, known value item, is the item whose price the customer remembers and uses to judge whether the whole store is expensive or cheap. It is not the same as a high-velocity item. Not every best-seller is a KVI, because KVI is about price perception, not volume.

The rice, the 2-liter soda, the diapers. The customer knows the price and compares. If those run expensive, the whole store becomes “expensive”, even if everything else is competitive.

How the operation measures it

In theory, with customer price surveys, traffic items, and systematic competitor comparison. In practice, anyone who has worked the operation knows the KVI list usually comes from somewhere else.

It comes from the buyer’s gut. “These are the items the customer watches.” Sometimes it is right. Often it is last year’s list, inherited, never tested against sales.

What the number hides

The political list hides two things. First, that the real KVI set is smaller than the list, and shifts with the season and the location. Second, that items enter the list out of habit and stay, even when the customer stopped watching.

The result is the worst of both worlds: you protect zero margin on items nobody compares, and bleed margin on the ones the customer actually checks. The data would say which is which. Politics says which always were.

Why this changes an AI project

A pricing engine treats KVI and non-KVI in opposite ways. KVI tracks the competitor closely; non-KVI optimizes margin. If the input list is wrong, the engine applies the right strategy to the wrong item, with confidence.

Before any model, the question is which KVI is real, measured in behavior, not in memory. It is the same discipline as elasticity per SKU and markdown by source.

Think about your KVI list. Who built it? When was it last tested against real sales and comparison? If the answer is “it has always been this way”, it is politics wearing the face of data.

Tell me the pricing term your operation uses but defines by eye. In one hour I will send back a one-page card: the formula, the measurement points, what the number actually means, and where it misleads. If the card shows the list has to be rebuilt from behavior, the two-week Diagnóstico is the next step.