Watch this video where Liz McLaughlin (Data Scientist at Wilson Allen) describes how to leverage a data lake to combine previously siloed data from disparate systems, and build a holistic definition of (in this example) client health from a broad range of features.

Liz built this model to establish metrics to identify & segment clients by client health – bringing together data from financials systems (3E), sentiment scores (Clearlyrated), relationship strength (Introhive), and opportunities data (Salesforce). The algorithm groups the data points, each representing a client, into segments sharing similarities across the feature set. Through machine learning and data science, Liz has broken down a problem that wouldn’t be possible for a human to complete – taking 500, 1000, 2000 clients, and within seconds categorizing them into actionable client segments. Then you can drill down into the ‘wrong-fit’ segment – prompting decisions and actions that may then be appropriate – whether to funnel resources away from them into an ‘opportunity’ client segment or use this as an intervention opportunity for clients at risk – you’re now making these decisions backed by DATA.