What you get from conjoint platforms, from statistics software, and from Chosŏn. Rows marked coming are being built now.
| conjoint platforms | statistics software | Chosŏn | |
|---|---|---|---|
| Where Chosŏn is different | |||
| Learns how shoppers switch between products | a similarity rule | possible; hours to days at shelf size | yes, in minutes |
| No decision tree to draw | rule tuning or nests | nests by hand | none needed |
| Full shelves, with switching | rule only | not practical | 20 to 60 products |
| Forecast level | tends to run low | — | close to the truth in our tests |
| Menus in one model at every size | method changes with menu size | by code | yes |
| Which products to test next, and what to recommend, chosen as a set | no | general tools | coming |
| Choice-based conjoint | |||
| HB logit, simulator, willingness to pay, importances, elasticities | yes | yes, by code | yes |
| Segments: latent class and taste clusters | yes | yes | taste clusters; latent class coming |
| Constrained part-worths (for example, price always negative) | yes | yes | coming |
| Respondent covariates in the model | yes | yes | coming |
| Range, price and product optimizers | yes | by code | yes, on how shoppers switch |
| Volume, revenue and profit forecast with a hurdle | partly | by code | coming |
| Dual-response none, alternative-specific and partial-profile designs | yes | yes | coming |
| Adaptive CBC | yes | — | coming |
| Volumetric CBC | yes | by code | coming |
| Excel simulator | formulas or export | — | add-in running the full model (beta) |
| Other studies | |||
| MaxDiff, including anchored | yes | yes | yes; anchored coming |
| TURF | yes | yes | yes |
| Pricing studies (Van Westendorp, Gabor-Granger) | yes | yes | coming |
| Claims and concept tests | yes | — | coming |
| Surveys and data | |||
| Design generator | yes | yes | yes |
| Survey hosting | yes | — | coming |
| Import from other tools’ exports | own format | any, by code | long and wide files; other tools’ exports coming |
| Compute | |||
| Where it runs | desktop or their cloud | your machine | in the browser; real-size conjoint on the Chosŏn Engine on paid plans; large runs for full shelves |
"Conjoint platforms" and "statistics software" describe what is typical in each category, not any one product. "Close to the truth in our tests" means on surveys simulated from a known truth; the course shows how. Plans and prices.