When a product leaves the shelf, its buyers go somewhere: to a sibling, to a competitor, or out of the category. Chosŏn learns where from your conjoint survey, so the range you keep holds on to the shoppers you can't afford to lose.
Cutting a product only pays if most of its buyers move to something you still sell. That transfer is the number the decision rests on, and it's the one standard conjoint gets by assumption.
Most conjoint software sends a delisted product's buyers to every remaining product in proportion to its share, or by one similarity rule chosen before the study.
They stay within a flavor but not a pack, or within a brand but not a flavor, and the pattern differs from category to category. One rule can't capture that.
A range that cuts the product with no substitute and keeps two that compete for the same buyers. The forecast for the new range comes out low too, so a good plan can look worse than it is.
The same choice-based conjoint survey. A model that learns how shoppers switch, and a simulator built on it.
Chosŏn estimates from the choices how closely shoppers substitute along each attribute: brand, flavor, pack, size, whatever you test. No consumer decision tree to draw first.
Remove any product in the simulator and see, product by product, where its buyers go and how many leave the category, each with a 95% range.
Take any set of products off the shelf, together or one at a time, and compare what the remaining range wins, with 95% ranges. The optimizer then tunes your products' features and prices against the competition.
The volume and profit level of the range you pick, so the business case is made on the right number.
In a controlled study of a 20-product category, with shoppers who switch along brand, flavor and pack, the standard approach, even in its strongest form, kept a range that earned less than the best one and made the same wrong cut every time. Chosŏn picked the best range in every case, and its profit forecast landed close to the truth while the standard forecasts came out low.
Range reviews with a transfer estimate for every product, not a rule of thumb.
Which variants carry their own buyers, and which ones share them with a sibling.
Ranging studies for clients, with a source-of-volume answer they can act on.
No. A choice-based conjoint where products share brand, flavor, pack or other levels across tasks. Shelf-style tasks with many products work well. Studies of up to 5 products per task are free; larger runs fit up to 60.
A decision tree is drawn before the study and fixes which products compete. Chosŏn estimates it from the choices, and it can find that shoppers switch along more than one attribute at once.
When shoppers don't see some products as closer substitutes than others, every method gives about the same answer. Chosŏn tells you which case you're in: the Switching tab shows how much of the substitution each attribute carries.
Small studies are free. Larger studies run on the free trial, academic and beta tester routes. See the four ways to start free.