Our Solution For Clean Water…..Data
Clean…. or murky? A common requirement for water companies is to understand how many true customers there are within their overall SPID level data (i.e. supply point); any one customer may have multiple sites, and multiple meters, hence multiple SPIDs. It is vital to link SPIDs within the same customer group in order to fully identify how many true customers there actually are. And clearly then to understand the value and margin distribution within these, such that High Value customers can be identified and treated appropriately; similarly low value customers against whom there is potentially a high cost-to-serve, and thereby running at a low or negative margin. Commercial data provides some very specific challenges. Business addresses invariably involve some sort of complexity in needing to detail a Unit number, or a specific location on a bigger industrial estate; and for meter-reading purposes the precise location is sometimes included within the address. Coupled with this, there is a proliferation of Garages, Troughs, and other unique locations that are each served with water and/or waste. Businesses are often named to ensure uniqueness from similarly named organisations; other times they are named as closely as possible to pretend to be the more famous brand. Of course they also include a variety of corporate suffixes denoting the legal entity upon which it was established (Ltd, plc, LLP, T/A etc.). And they make plentiful use of non-standard characters such as &,-.()# eg “J.Harris & Sons (UK) Ltd.”, which adds inevitable difficulty when merging many site records, if the company name has not been punctuated consistently on all records. Address details are also prone to complexity, either through the wrong element of the address being in the wrong part (eg the Postal Town in the County field); or through mis-spellings, or letters transposed. Further complexity still arises where sites with the same company name should be specifically not linked together eg franchises, pubs, churches. And conversely, where sites with different company names should specifically be linked together (eg Boots the Chemist, Boots plc). Our solution The reality is that even the most advanced de-duplication software cannot handle the complexity arising within SPID names and addresses. We have found that a combination of our own bespoke cleaning routines, coupled with manual eyeballing, is the only effective way to truly clean data like this. In the past we have put this into practice for Business Stream, and more recently for Thames Water B2B, now part of Castle Water. Powerful de-duplication software, bespoke cleaning routines, eyeballing and a methodical approach need to be applied in equal amounts to get that true understanding of the customer. Couple this with further enhancement services we can provide to add SIC, No. of Employees, Year of Establishment (and more); and data services to understand your customers’ out-of-area presence; and you have an incredibly powerful platform on which you can focus your retention activities, and increase your customer acquisition volumes.