Research

How to use Geo- Targeting to improve lift from digital

A heathcare brand wanted to understand the impact of using different digital channels in a practical way – so that the quantum of investment on each could be decided. The idea was to create a baseline for a market and check the lift with the application of a campaign run on a specific digital channel.

RainMan first clustered the markets based on sales. The next step built a relationship wherein a specific market sales was predicted with a combination of the other markets. This algorithm is useful as it predicts the sale of “A” market – with the sales from other markets (B …n). Take was taken to validate this model and ensure that it actually predicted sales in “A” market with a low MAPE (mean absolute percentage error)

Now the stage is ready for the experimentation. Additional specific channel  inputs are provided to “A” markets over a number of weeks. The sales in these weeks are predicted for this market and then compared with the actual sales. If the actual sales are “significantly” higher – that is over and above the usual deviation, then that lift is attributed to the specific channel. Repeat for other channels with other markets.

The advantage of this methodology is that ROI of key digital channels can be compared and spends optimized in line with these insights, significantly improving overall efficiency.