Prediki PROMPT

Simulated test marketing (STM) – with market leaders Nielsen BASES or IPSOS DESIGNOR – has a number of well-known weaknesses, some of which make predictive test marketing (PTM) the method of choice.

A well-known problem for STM is its application across different countries. STM can do well in one market but fail in another (Korotkov, 2013). This is due to lack of comparability of the answers produced by respondents from different cultures. The extent of overclaim, for example, is quite different for a Chinese respondent and a British one. However, betting behavior is immune to these cultural differences. A Chinese prediction market participant will bet exactly like a British one.

Further, STM needs a highly detailed and firm marketing plan to come up with forecasts. This exercise adds significant time delay and preparation work, only to be invalidated should the need for agile changes to the plan arise. In contrast, Prediki PROMPT can get by with a broad brush marketing mix and plan in the early phases, and detail can be added as the project progresses through subsequent stages.

The same principle applies for certain product classes which are bought only rarely or only by a select few buyers. A low incidence of available respondents precludes the use of STM, making Prediki PROMPT the method of choice.

The ever present biases and cognitive errors in questionnaires used for STM have already been discussed above. Prediki PROMPT uses questionnaires only for the purpose of screening category users and target segment sociodemographics - questions which attract minimal error - as well as for priming participants with stimuli for the possible new products.

The absence of sufficient normative data or outdated data or gaps for the region in question is also a showstopper for STM. In contrast, two or three years’ worth of up-to-date market share data is all that Prediki PROMPT needs to create enough context for prediction market participants.

In certain industries, such as pharmaceuticals research, respondents are hard to find or expensive. A predictive community, however, can be built not just with doctors but also nurses, patients, or their carers.

Last but not least, Prediki is a Bayesian method: participants interact by betting on higher or lower outcomes, changes are instant. If things change, an update can be produced incrementally, as a matter of a few days rather than repeating the whole exercise.