There are number of interesting Use cases in Marketing function that can be powered by AI solutions for making better decision and exponentially increase the desired outputs and ROI.

Xencia provided an interesting AI powered predictive solution for a Marketing platform by combining the power of predictive analytics and Azure ML services, for a Marketing platform that caters to hiring social media influencers.

The customer is an influencer marketing medium that enables businesses to explore and connect with influencers and run multiple marketing campaigns. They help top retail service providers hire influencers instantly.. These influencers can belong to different social media platforms like facebook, twitter, instagram etc. Tens of thousands of influencers are registered on their online platform where they network and collaborate with each other. The end customers can hire the services of influencers to run their campaign based on the campaign requirements and budget.

One of the key insights the platform demanded was – an accurate prediction of the cost of influencers both new and existing which is based on their past performance or historical data on similar campaigns.

As a company that is backed by Machine Learning & Artificial Intelligence tools to provide their service, they intended to utilize a proactive approach by investing in high-powered predictive analytics to filter out accurate results, based on certain specified criteria.

At Xencia, we did a detailed analysis of solution and built a solution that was both custom made as well as leveraging Azure platform services for fast tracing the solution Implementation.

We went over the data thoroughly which was being logged in from various social media sources and with different templates and form. The data was examined, cleansed and analyzed and stored using Azure Data lake as the data landing Zone.

The ML Models were built using Python Libraries that recommended a Pricing model for the influencers based on Follower Count, Following Count, Average Engagement, Average Likes, Average Reach, Number of Subscribers, etc. to estimate their price. These features were trained using the pruned data from Data lake that was split us training & test data. The trained model was tested using Azure ML Studio, and the model was finally deployed with the Azure Web App.
The entire training, testing & deployment was automated using Azure Devops, for an Automated ML ops to cater to future development and Integration.

The deployed solution has proved to unlock valuable insights into the current market for hiring influencers, pricing and also has made the budgeting process competitive for marketing campaigns. There has been a significant increase in campaign effectiveness measured on budget and campaign reach and has proved to be effective over 30-50% as compared to before. Also there has been increase in influencer hiring by 20% over the past quarter since the AI powered solution has gone live.