Marketing Campaign Optimization

Scenario
This publicly listed large marketing agency helps some of the largest brands in B2B segment to generate leads and help prospects move forward in the sales funnel. The agency hosts several well-known websites, produces content to engage with prospects and improves their engagement level by offering white papers, admission to industry events, webinars etc. One of the key elements of their marketing was massive email campaign system across different segments. More recent observation suggested that more and more people were unsubscribing to the email campaigns thus closing the door for future engagement. This was an alarming situation and the agency was interested in learning how the attrition can be minimized.

Solution
Data & Analysis
This is a complex online marketing problem and required a lot of discovery. The team started to look at questions such as What is the marketing funnel being employed? What segments exists and how have the prospects been segmented? How relevant is the content to the individual and segment? What is the frequency of emails received by prospect? Is the attrition behavior predictable? What different hypothesis exists about the attrition?

Model
The data elements required to effectively analyse and build the model was available in multiple systems. A central database was developed to compile information from different sources such as customer relationship database, online user-action (clicks, download of whitepaper, registration for webinar, registration for conference etc.) database, as well as email campaign user actions to determine which content topics does the user respond or not. The final model helped in identifying those users at the risk of unsubscribing, it also helped understand what content each user responds to so that email campaign content can be customized accordingly and the user is not bombarded with topics that are of less interest, lastly it helped optimize the user-action whereas the website can serve content based on the individual history.

Decision
The model has enabled this organization to optimize their email campaigns, slow down the un-subscribe behavior, and improve online user-action. The power to respond to individuals based on their behavior and history along with incorporating cluster behavior (who else is like this individual in the target segment and what are their preferences.) is changing the way individuals interact and engage with brands.