In a $500 billion subscription economy, companies lose 50% of their customers every five years. And while they have a wealth of data on these customers, the data is in disparate places, and even if aggregated, companies lack cutting edge analytical tools to optimize insights. Companies suffer from subscription cancellations that they could have prevented if they had sophisticated technology to identify both high-risk accounts and the optimal retention tool for each client.

Sparked’s Retention Radar provides this technology as we aggregate data from each of a company’s Salesforce data sources: case history, including facebook posts, tweets, emails, calls and tickets from Desk.com; purchase history from Salesforce; Web and product usage history from ExactTarget; and all of the profile information from data.com or any other data source that IDs a company’s customers.

Then we run the data through SparkedMatch, which is where the magic happens. Sparked match finds attributes and combinations of attributes that correlate with customer churn. We use machine learning, natural language processing, sentiment analysis, and modeling techniques to automatically discover the attributes that are most highly correlated with churn.

We then deliver the data via mobile device - using a beautiful data visualization that allows companies to quickly see problem areas - and to start a Chatter session with the team to collaborate on a solution. And when there's a red alert -a really significant problem, we'll send a mobile notification so companies can take immediate action.

The immediate impact on businesses is profound as reduction of the customer defection rate by 5% can increase profitability by 25 to 125% (depending on the industry). The cost- and time-efficiency of retaining customers is critical to any corporation, watched closely all the way to the C-level. Ultimately, Sparked’s Retention Radar becomes a vital tool for quick, tangible, and sizable bottom-line impact for any subscription business.

Share this project:

Updates