These decisions have to be based on and backed by historical data. For example, Mashreq Bank uses its historical data to predict if its marketing campaign is going to work or not. In conversation with ETCIO, Dr. Allen Roy, Head of Analytics at Mashreq Bank, talked about how the company has built an attribution model to predict the return on investment of marketing campaigns with the help of historical data.
“Most of the marketing campaigns are below the line. For this we have chosen from the palette of offers we have, the customers we want to target and the channel we want to use. Before we go ahead with the campaign, we also need to have a good understanding if we are using the right channel or the right offering. While these decisions used to be a guess work earlier, we wanted to build a model that helped us figure out the best way through the historical data,” he said.
The company built The RoMI model, i.e. Return on Marketing Investment, which is able to predict to a certain amount of accuracy what could be the return on investment from the marketing model. Every campaign goes through this RoMI model and if it doesn’t meet the threshold that the business has set, the campaign doesn’t go through.
RoMI module is not based on somebody’s wise thinking and ability to articulate but on the actual data of all campaigns that the company has done in the past and what we got from the investments. It’s the historical data that helps the bank decipher and bring transparency to its decisions.
“All the campaigns that we have done with the bank historically for 4-5 years, are there with us. We have fed this data into the attribution model where we identify a few things such as profiles of people who are interested and those who are not interested. We get to know the people who have actually taken advantage of the campaign and have strengthened the relationship with us,” he explained.
“Today, if I have to do a new campaign I first need to take out the gamers, the ones who accepted the offering because there was a cashback or something else but did not strengthen their relationship with us. Further, I need to focus more on my offering and make it suited for the people who have previously taken my offerings and have stayed with me. Looking at the past data, I would pick the channel that the customer is most likely to pick. RoMI is going to help me take the best of decisions and this will give me a good sense of response, conversions, and ticket size that I should be expecting from the customer,” said Roy
The data from the past campaigns helps the bank identify the ones that have taken advantage of the campaign in the past, and build better relationships with them. This also gives them an estimate of ROI that they should expect from the campaign.
“Since I also know the cost my business has incurred in the campaign, I can easily calculate the ROI that I am going to get with some amount here and there. But broadly I will get a good sense of return. If the business has set the ROI to be 3 times, and the model expects it to be 2.95, we will have to look at it. But if it says 1.97, we don’t go ahead with the campaign or work it to reach 3,” he concluded.
(With inputs from Dhrumil Dhakan)