- "Propensity-to-buy" models
- Probability of default models
- churn Models
- Predictive models for "Strategic & Tactic decisions"
- Customer-Life-Time-Value forecasting
- Direct ROI simulation
- Optimal resource allocation & planning
- Next Best Activity/Offer/product
- Text Mining
- E-mail campaign optimization
Try the TIMi Suite
Test the TIMi Suite for free. Click here
to get a full 3-months trial
of the TIMi Suite.
"Propensity-to-buy" models: Business-Insight
TIMi is one of the very few datamining software that is able to give you an accurate and very succinct list of the only variables/columns that have an influence on the prediction. Using this (usually very short) list, you can easily have a clear "portrait" of your target: Here is an excerpt of a TIMi report:
The objective here was to predict if somebody will buy a GPS device: Here above, you can see a screenshot of a standard Analysis report, automagically generated with TIMi. This predictive analysis report accurately describes the "portrait" or "profile" of the "GPS device buyer".
Amongst the 200 columns/variables that are available in the analyzed dataset, TIMi decided to build a predictive model that is using only 10 variables. In particular: you can see in the above Analysis report that, in order to predict if somebody will by a GPS system, TIMi decided to use the following variables:
- "Number of Home Cinema Equipment"
- "Use Internet banking"
- "Buy MP3"
- "Number of purchase using internet"
- "Buy Digital TV subscription"
- "Time spent on internet"
TIMi decided to IGNORE the following variables:
- "sportive holidays"
- "Buy hiking equipment"
- "Buy a backpack"
- "Buy sport holidays"
- "Buy Safari/Nature holidays"
Out of this report, we can conclude that it appears that the typical "profile" of the "GPS device buyer" in 2007 had a very strong technical background (don't forget that this model was built in 2007: the "profile" might have changed now).
Interestingly, in 2007, one famous company (that is still selling GPS devices noawadays) made a marketing campaign where you could see, on the television, a spot with a "Robinson Crusoe" kind of man, with a long beard, coming out of the wood saying: "I found civilization back using my GPS!". This television spot was obviously a poor choice because of the nature of the typical "profile" of the "GPS device buyer". In 2007, a television spot involving a more technical, IT person would have been a better choice.
All good datamining software should give you not only good prediction accuracy but it should also give you good "profiling" information about the target that you are predicting (see the example above about the "GPS device buyer"). Our consultants (from the Business-Insight company) are calling these "profiling" informations" delivered by TIMi under one common term: the "Business-Insight":
accurate predictions AND accurate "Business-Insight".
Next: Predictive Accuracy = Money. How so?
Test the TIMi Suite for Free!