Specification of So-called Flight Patterns

Our clients often consult us and request placement of ads in a flight pattern which will bring about the greatest effect within a given budget. The optimal flight pattern depends on what category of commodity one sets or which key performance index is set. At this time we would like to introduce the case study of a flight pattern decision process which set, the key performance index as the maximization of the percentage of Unaided CM Awareness (Note that numerical data shown in this report are replaced with dummy numbers for the sake of confidentiality.).

Hypotheses Derived From the Figure

In order to conduct the analysis, first of all we must set the hypothesis for the optimal flight pattern in the category of the relevant commodity. We set the period to specify the flight pattern at 16 weeks and set out past campaign data to obtain the pattern. As a result, we came up with the following hypothesis:
(1) Awareness tends to rise fast and high when the GRP is injected early in the campaign period (between the 1st to 3rd week of the campaign period);
(2) Once the awareness arises, it does not easily decline (strong sustainability);
(3) Awareness recovers if the GRP is injected during a week with some degree of GRP amount.

Hypothesis (1) indicates that one should adopt the so-called Burst type of flight pattern. Hypothesis (2) indicates that it is better to raise awareness in the early stages and indicates that one can reduce the cost not by injecting GRP continuously but by injecting it intermittently. Hypothesis (3) indicates that there will not be any effect unless one injects GRP weekly to some degree.

Bench Mark of the Amount of Weekly GRP

In order to test the above hypotheses, we first determined the amount of weekly GRP mentioned in hypothesis (3). We plotted a scatter diagram of actual data after the burst period (from the 4th to 16th weeks) so as to gain the tendency of the data. We took the horizontal axis as the weekly GRP and took the vertical axis as a week-over-week elongation percentage of awareness. From this diagram below, we figured out when the GRP becomes 200 or higher, when the elongation percentage goes around 5% and is likely to result in a clear elevation of the awareness ratio. Therefore, we decided on 200 as a bench mark of the amount of weekly GRP.

Specification of Flight Pattern

Based on the information so far, we conducted analysis to specify the flight pattern which raises awareness the most. As the aim is to raise rate of awareness, we set the objective variable as the 16 week cumulative rate of awareness and set the explanatory variables as the following three:

(1) Burst Type between the 1st to 3rd week (Casting 40% of the 16week cumulative GRP during the 1st to 3rd weeks),
(2) Number of times over the 200 weekly GRP casting between the 4th to 16th week,
(3) Variance of the weekly GRP amount between the 4th to 16th week.

By analyzing the above variables with Hayashi's Quantification Theory 1, we found some characteristics of optimal flight patterns. Those characteristics are:

(1) During the 1st to 3rd week of the campaign period, awareness rises more when the flight pattern is the Burst type;
(2) During the 4th to 16th week of the campaign period, awareness rises more when the number of times over the 200 weekly GRP casting is greater;
(3) During the 4th to 16th week of the campaign period, awareness rises more when the flight pattern is the so-called Flighting type (This indicates that casting a temporally dispersive GRP performs better than a continuous casting.).

The flight pattern which, we have found from the analysis, would raise awareness is shown in the figure below:

As awareness maintains its level once it rises, the optimal flight pattern should cast a vast GRP from the beginning of the campaign to boost up awareness early. After that, it should cast more than 200 GRP in moderate intervals. Such a flight pattern leads to efficient usage of a given budget and to prevention of awareness decline.

Summary

This completes the introduction to our analysis for now. Actually, we conducted a further analysis in the actual project, and presented an ideal flight pattern according to budget level and creative type. Though we set the objective variable as aided CM awareness in the above analysis, it is possible that the ideal or optimal flight pattern may change if the index highly related to sales is not aided recall but unaided recall, or if the objective variable is not CM awareness but brand awareness. In addition, GRP reactivity may change according to commodity category or target.
We are able to present you with an optimal flight pattern according to brand, target, or advertisement goal. Please feel free to inquire.

Written by SPI

Please contact us with questions or for more detailed information.
spiindex@spi-consultants.net