|   |   
Customers: SUCCESS STORIES

Chordiant Decision Management—Media Planning

One of the Netherlands' largest players in the savings funds market had Chordiant Decision Management re- schedule its direct response television advertisements

The Challenge
Our customer regularly advertises on television, inviting the viewer to call a toll-free number for information or even to open an account immediately. Chordiant Decision Management was asked to predict which advertisements would yield the highest response and thus offered an opportunity for large savings; advertisements predicted to yield a low or even zero response were to be removed from the broadcast schedule.

The Solution
A database was constructed containing the response to advertisements broadcast in two previous years. It contained a number of attributes for each advertisement: the time of day and the day of the week of the broadcast, the network employed, etc. Subsequently, Chordiant Decision Management was employed to develop a model capable of differentiating, on the basis of these attributes, between low- and high-yield advertisements. The resulting model assigned each advertisement to one of four classes: good, above average, below average and bad. About 25% of the cases is assigned to each class.

The Outcome
Figure 1 shows the percentage of responders per class of advertisements. It shows that advertisements in the class labeled as 'good' (the top 25% according to the model) yield twice as many responders as the 'above average' class. In fact, these advertisements yielded almost half the total response. At the same time, the 'bad' advertisements turned out to contribute the lion's share of the costs (see Figure 2).

The Result
The Chordiant Decision Management model allowed our customer to achieve almost half the total response at about a quarter of the total cost, by broadcasting only the advertisements categorized as 'good.' Broadcasting twice as many 'good' advertisements would result in almost equal response at about half the cost. Finally, eliminating the 'bad' advertisements would save 38% of the costs while losing only 13% of the response.

[ Return to top | Return to Success Stories ]