More Advertising

Data analytics is really useful for understanding what is happening in a business, but often its true value is in its predictive ability.

We worked on a project with an American business that sold a range of products via inbound calls to a call center. These calls were driven by advertising through a range of channels in key target markets. Both the marketing manager and the sales call center manager were interested in understanding how advertising spend impacted on calls to the call center so we developed a data analytics project to understand the relationship between advertising dollars spent and calls received.

The first thing we did was had a specific phone number allocated to an advertising campaign, then we set the advertising spend at $1000 and tracked how many calls this generated. We stepped the advertising spend up in $500 increments while tracking calls generated. When we had reached $5000 we were able to demonstrate a clear linear relationship between money spent on advertising and calls to the call center.

We used this data to conduct linear regression which allowed us to develop a function that can be used by the business to predict how many calls they will get for any level of advertising spend (within reason). This function was tested by making predictions for what the call volumes would be for advertising spends between $5000 and $7000, and then actually running the advertising while monitoring call volumes.

While it’s hard to predict human behavior with perfect accuracy and a certain amount of randomness is unavoidable, we found that our predictions were inline with what actually happened and this function is still being used by the business to make predictions about how much to spend on advertising based on how many trained call center staff the business has available.