Stage 3 Pyramid Analysis Paper

We have historically segmented our clients by how many assets they hold here at Vanguard. A clients aggregated holdings determines their level of service. We have also identified patterns to determine clients who are likely to leave (increase in withdrawals over a small period of time) so we could call them and see if there were any issues. Another pattern is clients who are highly engage (amounts of logons) to market new products and services. Identifying these patterns is time intensive and an extremely manual process. We do not have enterprise-wide analytics.

We need to have the ability to analyze all of our data and use it to create internal and external analytics for a digital one-to-one experience, similar to Amazon. I feel over the past 40 years Vanguard would be in Stage 2 of the analytical completion pyramid. We were never ‘flying blind’ as described in in Stage 1 but the rigor of systematically being able to leverage data for fact based decision making was minimal. Marketing was determined more by what we felt the customer wanted or management intuition.

In current day, we have both descriptive (mine data to understand what has already happened) and some predictive (uncover relationships between data to understand why something happened or what will happen in the future) analytics. These are not enterprise wide analytics
An example of a growth model would be to identify high net worth clients who have propensity for advice. Instead of sending an advice campaign to everyone, we want to target it to our clients who we feel will benefit and sign up for advice. An example of a cost containment model would be to identify clients who have a low balance but continue to call us, using our expensive phone channel. We need to change their behavior and show them how they can self-provision on the web or be able to route their phone calls to less expensive phone