In my first blog in this series, I talked about why insurers’ use of analytics needs to shift from predictive to prescriptive models. It’s one thing knowing the probability of an event occurring. It’s quite another knowing which actions (or sets of actions) will most likely optimise a result and deliver a desired outcome.

Companies in other industries (like banking and retail) recognise that data-driven, prescriptive decision-making is vital. And to make this happen, they’re re-engineering their end-to-end value chains around machine learning.

To illustrate the kind of breakthrough benefits insurers can realise by taking a similar approach, we’ve launched a new website that brings this to life for each of the five key stages of the customer journey: Win, Grow, Service, Claim, Retain.

As we demonstrate, the valuable intelligence generated by analytics enables insurers to provide relevant, tailored offers to customers – at the right time, through the right channel, and at the right price.

In this blog, I’ll highlight what this looks like at the first stage of the customer journey: ‘Win’.

The holy grail for insurers is to identify high lifetime value (LTV) prospects as early as possible and, having done so, increase the likelihood of converting these new customers. The key to doing so? Harnessing the full value of data and applying machine learning to it.

At a high level, it works like this: with the right data (and datasets that have been built in the right way) it’s possible to create limitless experimental scenarios (by prospect, location, incentive etc) and then use prescriptive analytics to identify the next best actions for conversions in every individual case. By crunching the data, the algorithm at the heart of the machine will tell you which combination of factors should be used at every stage to increase the likelihood of a successful outcome (that’s to say, converting a high-LTV prospect into a customer). 

Drilling down, but without getting too deep into the detail and the science, there are three priority areas for analytics at the Win stage. First, the web-to-quote phase.

By capturing all available data from browsing behaviour, insurance requirements, customer background, demographic information, and so on, and applying machine learning to that data, insurers can calculate the acquisition long-term value of each prospect. With that intelligence, they can adjust pricing and offers to win the highest-value customers.

Next, aggregator rank prediction. By using uplift modelling and the right blend of external data, insurers can predict an individual customer’s sensitivity to an aggregator’s rank and/or brand, and feed this information into the pricing model to improve outcomes.

The third priority area? Digital marketing. This is a critical battleground for insurers. Facing fast-growing competition from aggregators and direct players, they urgently need to identify new ways to acquire valuable customers through digital channels.

With machine learning applied to the right data, they can use programmatic media buying to cost-effectively target high-LTV customers through the right digital channel. Programmatic media buying uses real-time bidding technology to serve ads, as and when a targeted user is on a specific channel. By using a rich blend of data about each prospect, it’s now possible to target ads with extraordinary precision.

The results that can be achieved are hugely persuasive. We’ve found that with programmatic media buying, insurers can realise 5 percent in additional new business, along with 40 percent reduction in marketing spend over traditional digital techniques.

The fact is, from the moment a prospect first visits a webpage and starts to complete an online form, analytics can spring into action to derive hundreds, if not thousands, of variables. Combined with other data sources, these can be used to develop a rich view of the customer, their likely LTV and the next best actions for winning them.

The other beauty of machine learning? The more you use it, the smarter it gets. That means the predictions and recommendations get better and better with time. Next time, I’ll be looking at what machine learning can achieve at the next stage of the customer journey: ‘Grow’. In the meantime, if you have any questions, it’d be great to hear from you.

 

Submit a Comment

Your email address will not be published. Required fields are marked *