A guide on how the credolab solution can help insurers
Alternative data is the set of information from multiple data sources, such as mobile devices and applications, Web and touchscreen interaction, online and mobile purchases, and bill payments, among others, that are analysed by Machine Learning algorithms to predict users' behaviours. These algorithms constantly analyse and (re)learn behaviour patterns in real-time and reliably link them to user outcomes of interest – product purchases, balance (re)payment, or the filing of an insurance claim. In essence, they make it possible to understand and anticipate consumer needs, wants and behaviour.
In the case of credolab, the alternative data comes primarily from anonymised smartphone metadata, i.e. data on device usage patterns that reflect the typical activities of representative smartphone users. This allows companies like insurers and lenders to continuously learn about their target audience (current customers, potential new users etc.) while rigorously protecting their privacy.
The benefits of using alternative data can come in many forms. In the case of Insurers, alternative data offers new inputs that can be integrated into all the steps of insurers’ value chain. These include improving customer targeting and engagement, accelerating new premium origination, improving risk assessment in existing underwriting models, and providing insights into possibly fraudulent behaviour. The originality, relevance and wide coverage of this data, coupled with Machine Learning’s insights, provides insurers with a holistic view of the (potential) customers’ profiles, an understanding of their level of responsibility, and a deeper comprehension of their interests and most likely behaviour.
3 ways in which credolab’s solution can help Insurers
Manage Underwriting and Claims based on predicted behaviour
Depending on the policy under consideration (e.g., Life cover vs Critical Illness vs Device Protection), underwriting standards are the traditional tool employed to assess and manage the insured’s risk. Often, however, insurers have to deal with a critical trade-off: on one side, complex underwriting is effective but costly, complex and time-consuming, and it reflects poorly on customer experience. On the other side, no (or simplified) underwriting is simple, (deceptively) cheap, and customer-friendly, but it might have an outsized risk impact on policy book performance.
Credolab’s technology makes it possible to escape the tyranny of this trade-off by using its technology stack, dynamic data and deep Machine Learning insights. This enables automated, accurate underwriting that retains the ability to review and assess each application individually but at the same time minimises costs and the negative impact on customer experience.
In the case of simple products with small ticket sizes (say smartphone screen protection), it makes it possible to “underwrite” each individual applicant cheaply, in a matter of seconds, and request no applicant data. In the case of complex products like Health, it allows the current underwriting process to be simplified and shortened while improving its accuracy. The net effect is lower cost, higher speed, and a greatly enhanced customer experience.
Finally, credolab’s track record in predicting affordability, risk, and behaviours suggests that policy pricing and customer inclusion might also benefit reduction in fraud and delinquencies (i.e., costs) as well as being able to price for both baseline risk and behavioural risk. This will ensure that the largest possible pool of customers has access to proper cover.
Accelerate Portfolio Growth
Some insurance companies use credit bureau scores as a proxy for customer affordability and risk in an attempt to simplify and automate policy approval and issuance decisions. For instance, according to Fico, within those US states allowed to use credit scores to assess risk, 95% of personal insurers decide to use it for evaluation purposes.
Unfortunately, when Insurers use this strategy, many individuals who can afford to pay their premiums are declined or forced to pay a higher premium due to limited credit history.
In markets, or segments, where credit bureau coverage is lower (e.g. younger customers, recent immigrants, etc.), these patterns are further magnified.
In all cases, CredoLab's technology allows insurers to assess all new and existing customers for affordability and financial risk, enabling a more equitable and inclusive pricing strategy. Credolab's integrated scoring solution allows insurers to increase their approval rates without impacting their book’s risk profile: this is made possible by the use of alternative data via smartphone, available for nearly 100% of the target population, together with credolab’s proprietary technology and Machine Learning scores, able to understand and anticipate users’ likely behaviour.
Improve sales and customer experience
In addition to expanding the portfolio through a better understanding of customer behaviours and interests, marketing teams can use the data to convert prospects into qualified leads, pre-screened for purchase intent and risk/affordability, and deliver personalised offers for them. The credolab solution, engineered in partnership with Sapio Asia, implements predictive Machine Learning models that improve cross-sell, upsell and retention rates by assessing the probability that a person will buy a new product or accept a new service or promotion.
Importantly, credolab’s and Sapio Asia’s insights help also to increase the relevance of the product or service offered to the customer, improving their experience and overall degree of long-term retention and brand satisfaction.
Indeed, according to Appsflyer, a mobile attribution and marketing analytics platform, 1 in 2 apps are uninstalled within 30 days of installation, and applications from the insurance industry recorded up to 95% uninstall rates, mostly due to poor value and customer onboarding experience. credolab’s technology tracking web- and mobile-engagement micro-events, detects when the prospect hesitates, leaves incomplete information, does autocomplete or copies and pastes data that she should be familiar with. These insights provide invaluable help in redesigning interfaces and the customer onboarding process to create a pleasant engagement experience for customers while reducing abandonment/uninstall rates for digital products.
In summary, alternative data, and predictive Machine Learning insights, provide multiple and compelling use cases for insurers. Credolab allows Insurers to understand their audience better and make better decisions by using smartphone digital profiles and web behavioural metadata and its technology and simple integration model. Both insurers and their customer greatly benefit from that.