Last week I delved into the CV <çurrent value> tranch of data you need to collect. And ended at the scenario you might find yourself in whereby you cannot access that data directly from your internal finance system or any channel intermediaries that on-sell your products/service to the end customer base. If this is your lot in life, then you need to develop a proxy model to collect that all-important CV data.
The first key consideration is to define the possible characteristics of your relevant products/services that can proxy for inaccessible dollars revenue/margin data, at the individual customer level. At a primary fundamental level the possibilities are likely to, in some way, relate to quantifiable volume throughput/consumption of your products/services. Having set the parameters, then you need to define how you are going to gather that data across the customer base. This consideration, will inevitably involve the assessment on what role the sales team are likely to play in this pursuit. In very many cases , the team will be integral to good proxy CV data collection, if you have no other way to access it. Then the next consideration will be around how. Will you brief the sales team to ask each of their customers for volume/consumption throughput estimates of your company products/services from each customer as they make their rounds? Will they be able to act like an auditor and conduct indicative counts of product or service schedules that are visible to an “auditor”? Will you brief them to profile the customers on a set of characteristics that are auditable/investigable and then use these profile scores to feed a proxy “algorithm” to apply to higher level revenue data <higher than individual customer level, eg region, state> to assign a CV value at customer level? Will you use the latter technique across the board, without any subset statistical validation, or can you run a such a statistical validation subset and then extrapolate across the entire base. You can start to see that there is more than a few ways to skin this cat. Importantly, there is ALWAYS a way.
Come back next week and we will delve into the second critical data tranch – “potential value” <pv>.