In our previous blog post on this topic, we shared our initial thoughts on the FCA’s review of the potential car finance mis-selling. As the FCA announce that many firms are struggling with data challenges in mis-sold car finance claims, we take a look at some of the complications that loan providers may need to overcome while preparing to deal with this issue.
We won’t know the details of the FCA review until September, but there are still plenty of proactive steps that businesses can take over the next few months to ensure that they are prepared…
Do we have all the necessary data?
A remediation project on this scale will require a well-designed remediation database so that the data analysts do not waste time performing the same extraction and transformation tasks whenever they need to perform some analysis or check a few accounts.
Most firms will have a well-supported and documented database that contains all their active customers and accounts. But this mis-selling issue could include loans taken out as long ago as 2007. Where do we find the data for those?
Archive data
Most remediation analysts will have had to deal with archive data at some stage in their career. The initial challenge is finding the data in the first place. Most BAU analysts just do not have a need for data that is more than a couple of years old.
Furthermore, many lenders will have acquired other companies within the previous 15 years, or they may have purchased a segment of another lender’s loan book. This means that they will now be responsible for remediating loans that originated on a platform other than their own. This adds even more complexity to the data sourcing process.
Once the data has been found, it must be extracted. If the data is archived on a mainframe system, it will require a copy book to be understood. Data could even be held on tapes or other physical formats. This extraction process is a lot more complex than simply writing some SQL to pull some loans from a database.
Unstructured data
It may be the case that the only record of the original terms of the loan is the form that the customer physically signed, possibly stored as a pdf on an old archive system.
Previously, these documents would need to be read by a human then manually entered into a database. Fortunately, Dufrain have developed an AI based document processing system called Velocity that can convert scanned documents into data. Velocity is much faster and more economical than manual file reviews but companies will still need to act quickly to include this in their remediation process ahead of September’s judgement.
What if the data just does not exist?
After the above steps have been followed, there still may be cases where the data just is not held any more, especially when it comes to Personally Identifiable Information.
This particular data challenges in mis-sold car finance claims issue is unique in that each loan relates to a vehicle. The DVLA allow data requests for commercial reasons so this may be a useful source, for example, if a company still holds the vehicle registration with no other customer level data.
Any remediation project involving closed loans and ex-customers will require some level of track and trace process. Tracing agents like Experian, Equifax or Equiniti can provide current contact details for ex-customers. A track and trace mart, containing the latest address details as well as the dates that these are valid from and to, is a useful way of managing this process.
Data cleansing: ensuring confidence in usability
Once the data has been extracted, data quality checks will need to be made to ensure that the data is complete and in the required format.
Critical data components will need to be identified and data quality rules must be written for each of these. Once these rules are applied, each loan can be assigned a data quality score which can then be used to separate loans into data quality cohorts.
Cohort management is key to data quality remediation. This allows data quality issues to be investigated and resolved independently while also allowing the “complete” loans to move through the remediation process, avoiding delays in development.
The end goal
Following this process, businesses will ideally possess a meticulously organised and thoroughly cleansed database comprising the particulars of each loan within the impacted population. This will allow remediation analysts to carry out redress calculations and implementations seamlessly, free from any troublesome hitches or delays.
To find out more, go to our motor finance remediation hub for this topic.
Our team has extensive experience in this area, having recently helped just one of our banking partners successfully redress over 4 million customers.
Contact us and we can help you prepare.