Local authority focus: the power of data and predictive analytics
Simon White, DWF Business Intelligence Analyst, provides an insight into predictive analytics and how local authorities might begin to use their power to reduce the cost of claims.
What are predictive analytics?
‘Predictive analytics’, the mining of information from current and past data which is then used to predict trends and behaviour patterns, is currently a real ‘buzz phrase’ within businesses and organisations..
It is a statistical discipline increasingly utilised across all walks of life. Consider the 2015 UK General Election and specifically, the main exit poll results released soon after polling booths closed. The exit poll predicted a Conservative majority, with Labour predicted to perform much worse than expected. Political commentators and experts largely dismissed the exit poll predictions and did not consider a Conservative majority likely at all, yet when all constituencies had declared, the final result – a Conservative majority – had come to pass with the final majority achieved remarkably close to that predicted in the exit poll. The reason for this closeness? The use of predictive analytics.
In very simple terms, predictive analytics utilises past experience to predict the future. In the General Election, results of previous elections in each constituency would have been analysed statistically to detect voting trends, those trends themselves being informed by a wide range of influences – effectively, determining which factors cause voters to vote in the way that they do and what degree of influence is exerted by each factor.
Insight from claims data
Consider the challenge for local authorities of high claims indemnity spend. What are the likely key influencers of high indemnity spend?
A high volume of claims presented?
A low proportion of successfully defended claims?
A high proportion of outright fraudulent claims?
Long claim lifecycles leading to increased claimant solicitor costs payments?
High litigation rate leading to high defence solicitor payments?
The answer is likely to be a combination of all of these plus many other influencing factors, all of them influencing the indemnity spend to different degrees. It is very difficult, if not impossible, to attempt to capture structured data to uncover every single factor that might influence the high indemnity spend.
However, when predictive analytics are applied to your data, a whole wealth of insight into your operational challenge can be gleaned. For example, your data will tell you which types of claims ‘perform’ better for you, namely, the commonalities within the data for such claims – claims that see low indemnity spend may involve claimants with a low average age for example, or claims involving certain specific individual claimant solicitors. Claims that see high indemnity spend might naturally feature claims involving more serious injury but may also have a range of commonalities that may be unknown to the local authority.
Formulating strategies to effect change
At this stage of the process, it is worth pausing. Having gleaned detailed knowledge about your operational performance challenge from your data, you find you are at the stage of ‘….all very interesting but so what?’ This is a critically important part of the process. If you cannot operationalise that knowledge – making it accessible and usable - then the use of predictive analytics has not really achieved anything worthwhile. The ‘So what?’ test is passed when the data insight is converted into operational procedural change or strategies that can be deployed in the live environment to effect positive change.
An example of this may be that the data indicates claims presented via a certain claimant solicitor may be performing much better than average for the local authority in terms of indemnity spend. Might this suggest that that claimant solicitor is an opponent to ‘take on’ even where the evidence to defend an individual claim might not be particularly strong?
At DWF, our predictive analytics consultant works closely with our legal experts in our local authority team to effect the necessary procedural or strategic change and work with our clients to deploy these changes in their live environment. We also work with our clients to measure the effect of that deployment over time – a quantitative measurement that can be easily validated.
Predictive analytics is not a panacea to the challenges faced by local authorities or indeed any other business or organisation. Rather, it should be seen as a significant value-add tool to help the local authority make quicker and better decisions to improve the performance of their organisation against specific key performance indicators. There is a risk that the overuse of predictive analytics can lead to self-fulfilling prophecies – for example, ‘our historical data shows that we always settle claims involving x so we should continue to do so’. Analytics is most sensibly utilised to support and enhance decision-making but never to become the sole decision-maker. The data informing predictions must be continually refreshed with new data at regular intervals to ensure the predictions being made are as accurate as possible using the most recent historical experience.
DWF are capable of analysing any dataset that a local authority may wish to provide to try to identify general trends and challenges that might be hidden within that data.
For further information on the potential benefits of using data analytics and how it may benefit local authorities, please contact:
This information is intended as a general discussion surrounding the topics covered and is for guidance purposes only. It does not constitute legal advice and should not be regarded as a substitute for taking legal advice. DWF is not responsible for any activity undertaken based on this information.