Evolving a data enrichment strategy

Most brokers/insurers are doing some sort of data enrichment these days – even if that’s only a vehicle lookup to check validity/value/etc, or a postcode to validate an address.

But there’s a lot more you can do with data enrichment, and there is a lot of data out there.

The big question is: where to start, and having started, how should your data enrichment strategy evolve?

Why do data enrichment?
Perhaps the most compelling answer is that everyone else is doing it! But in truth data enrichment also can result in:

• Increased pricing accuracy
• Streamlined customer experience
• Reduced cancelations
• Reduced fraud

To achieve these lofty aims you have to do things with the data, not just have it. And that is where a phased approach to data enrichment is important.

Phase I – validation
Let’s assume you’ve chosen a data enrichment partner or partners. The likes of LexisNexis, Experian, and Percayso (all of which Ignite integrates directly to) all offer an array of data to enrich the picture of someone getting an insurance quote.

There are a number of data fields that may be relevant. Try to collect as many as possible to start with, even if you’re not using them immediately.

In Phase I, concentrate on validation: make sure the risks you’re taking on genuinely are what they say they are. For example, don’t ask a customer if they have CCJs – just get the data and quote/decline based on that. Don’t ask if they’ve got motor claims – use CUE to validate it instead. This has the added benefit of asking fewer questions to the client, which makes their whole experience better.

Phase II – expanding the appetite
After 3-6 months, have a meeting including your data enrichment provider, your software house, and your insurer to discuss additions to the strategy. The discussion might include expanding the risk appetite to cater for technically higher risks (like younger drivers or higher value homes) but with the protection of knowing the providence of the client.

Phase III – claims
After 12 months of quarterly reviews, it’s time to dig into the claims data.

There will be some high-level correlations with claims that you can pick out using just Excel and some common sense. This is where collecting lots of data from the start comes in handy because you can review it now and see what might be relevant.

Implement a small number of changes at a time in order to avoid confusion over what is and what isn’t working.

Phase IV – machine learning
Finally, there are some correlations within large volumes of data that can’t be gleaned with the human in Excel. These include multi-factor correlations such as Age vs Postcode vs Credit Score or suchlike. To understand these insights some machine learning tools are required. This is a relatively specialist and technical subject (which Ignite’s team can help with), but it makes sense to get the easy wins in the first 3 phases first.

Data enrichment offers a complete picture of customers and improved pricing accuracy that can help you achieve a real competitive advantage. Stay ahead of the curve by evolving your data enrichment strategy.

To learn more about data enrichment and how we can support you, get in touch!