Buy vs build in data analytics: four things you should consider

The unending quest of insurers to improve customer service and business performance has led them to embrace advanced analytics with open arms in the last 10 years. But moving from simple analytics and dashboards to machine learning and artificial intelligence has left many with an unexpected hangover.

Early adopters experienced various difficulties when deploying advanced analytics solutions in day-to-day business processes and retaining the staff to maintain them. Some life and health insurers have spent vast resources developing their own underwriting rules engines, only to start using a vendor's solution a few years later.

So, why is buying advanced analytics solutions so rarely considered in the first instance?

Data science teams are rightly excited about solving important technical challenges. More often than not, their answer is to develop sophisticated one-off models. If successful, a lot more work and resources are poured into the project to make it viable and valuable. To realise the benefits of advanced analytics, a repeatable end-to-end process with simple interface for business users must be developed and maintained. In this way, the process of embedding analytics is more akin to the discipline of implementing a new IT system.  And let's not forget, analytics needs to provide a good return on investment (ROI)!

In my view, the initial question when embarking on an advanced analytics project should always be "should we buy?" rather than "can we build?" This most important question will put you on the path to optimise your ROI in advanced analytics.

Here are four crucial evaluation factors that will help bring clarity on whether you should buy or build your own platform.

Clients and prospects often state that advanced analytics will be an important point of differentiation. However, we must disentangle a new / distinctive machine learning model that is unseen by customers from its differentiating effect on your customer value proposition.

Is it really differentiation in advanced analytics itself that drives your ROI? Or how you use the outputs from the analytics? And is it critical to get 100% performance from the analytics model?

It's perhaps better to procure an out-of-the-box analytics solution and free up your team to focus on differentiating your customer value proposition using the analytical insights. Isn't this where the real value lies?

Building, embedding, and maintaining an end-to-end advanced analytics solution is likely to cost many times more than the initial model built by your data scientists. While it is immensely satisfying to see results from the first model, massive commitment is required for it to be sustainable.

Invariably, embedding an end-to-end advanced analytics process is a big project on its own. It will need ongoing model development and the associated user interfaces. To manage business risks, you will also need to invest in testing, documentation, staff training, managing key person risks and ongoing maintenance and development.

In contrast, the value of working with an established end-to-end advanced analytics solution using an experienced provider will show through in the medium term. For instance, a platform like Swiss Re's Impact+ is ready to go and allows swift deployment. Also, we have designed friendly user interfaces and our solutions come with established product support and well-defined costs (including setup and license fees).

It can be costly to build and maintain a platform to support all your performance needs including the technical requirements of data scientists, ability to automate processes, data auditability and security, hardware speed and up-time, and scalability. When talking to prospective clients, these practical considerations often thwart the successful deployment of in-house analytics.

The buy vs build evaluation will likely come down to a comparison of the opportunity cost and ROI of the two options. The true opportunity cost of your team designing, building, and maintaining analytics solutions is likely substantial (and often underestimated). When considering your use of resources, if a problem can be solved with an established solution, then you should consider focusing efforts on other business challenges that can only be tackled internally. In doing so we maximise the ROI.

Find out more about Swiss Re Impact+ and how we can help you find growth opportunities & steer portfolios by harnessing data, advanced analytics and risk expertise. We want to help you unlock your next.