Best use of FastStats® 2014 Award
Using FastStats to create a Next Best Action platform
Ever since its foundation in 1897, the RAC has been consistently at the forefront in developing motoring services - from introducing uniformed patrols in 1901 and roadside emergency telephone boxes in 1912 right through to the present day when they boast the world's most advanced computer systems to deal with calls for roadside assistance.
RAC Members enjoy access to an enormous range of motoring products and services throughout their driving lives, ranging from the familiar and much appreciated roadside assistance in the event of a breakdown, to continually updated legal and technical advice and up-to-the-minute travel information.
The RAC is headquartered in Birmingham with approximately 4,000 dedicated staff operating from four bases throughout the UK and provides motor-related services to both consumers and corporate clients. The RAC offers vehicle and breakdown services through a nationwide, branded patrol force, focusing primarily on vehicle repair and recovery. In 2012, the RAC's patrol force of approximately 1,700 patrols attended to 2.5 million rescue breakdowns. In addition, the RAC provides motor insurance broking services and offers a suite of legal and motor claims services, which provide both consumers and corporate customers with legal cover and advice as well as a range of claims management services.
RAC and Occam first began working together in 2011. The provision of a Single Customer View (SCV – known to both companies as ORCHID) with a FastStats front end gives RAC a consistent and holistic view of its customers and products. This enables the Campaigns team to deliver multi-channel, multi-wave marketing campaigns to the customer/prospect base to support retention, cross sell, upsell and acquisition activities. In addition, it supports RAC’s Insight team in analysis, modelling and segmentation of customers, allowing better targeting of marketing activities based on this insight.
The next step following SCV completion was to use this solid foundation to create a Next Best Action (NBA) platform. If established correctly, NBA would allow RAC’s contact centre agents to understand the profile and requirements of an individual customer at the point a call was answered. Bringing up a person’s details from the database would enable agents to speak to people about relevant offers on RAC’s wide range of products and services, enabling cross and upsell opportunities.
It works for us, it’s fit for purpose and we have seen positive results.
The project gets off the ground
In 2013, when the project got off the ground, RAC had upwards of 20 million individuals in ORCHID: lapsed and live customers - many of whom hold multiple RAC products and policies - as well as prospects. The approach taken to build a successful NBA system was to manage costs by utilising software already in place at RAC: Apteco’s FastStats.
The rationale was to utilise the excellent modular capabilities and data management powers of FastStats, rather than buy in expensive off-the-shelf packages or build something from scratch, strategies which would both have exponentially increased the cost of processing and managing data. Occam set out with a team of 10 experts reporting to the Project Board, working closely in tandem with RAC and other third parties involved in the project, to a timescale of four months from scoping NBA to switching it on.
RAC needed NBA to present effective data to the call centre, identifying a customer or prospect and allowing agents to better manage the conversation, keeping cross and upsell opportunities front of mind. A crucial aspect of NBA is to make agents aware of what to say next, without them necessarily having to think about all the different offers they might make. They are prompted to discuss what might bring about the most successful outcomes. This makes for a smoother, more personalised experience for the customer and a more fruitful conversation for the agent.
Before NBA was built, the missing pieces of the jigsaw were the ability for agents to recognise who was calling them; identify which operational system the agent should be using, plus how to structure the conversation; and how to feed responses back into the system to drive incremental improvements to the engine.
FastStats was harnessed to replicate knowledge across hundreds of messaging treatments at the touch of a button, without the need for the expensive implementation of a brand new data development environment. RAC currently has logic, or treatments, within the NBA engine, which can be processed to serve any of the messages to customers and prospects through the contact centre.
The decision engine was built around a number of predefined functions written in t-sql\C#. The engine utilises these functions against both the NBA-specific metadata, and the base and derived data held within ORCHID.
Files are output directly from scheduled FastStats data grids. Additional attributes to support the NBA process are then added via the use of virtual variables. These are used to provide the call centre agents with customer-level insights from the Single Customer View, insights which are not otherwise easily available in the multiple operational systems in use by the call centre.
FastStats fulfils what RAC and Occam call the Eligibility process. This is the provision of selected messages from a library held within FastStats that pre-defines whether a customer is eligible or not for any of the messages based on the logic statements or “treatments”. This all supports allowing the agents to broach many relevant subjects on the call.
Occam created a batch engine system using FastStats. With the batch work taking place in the background, the agent has a pre-calculated priority to talk to a customer about. Insight is displayed simply and efficiently to agents.
It also allows RAC’s customer teams to change rules as required, rather than having to go through Occam to make alterations. The feedback response data from NBA-driven conversations is diverted straight back into ORCHID. The updated data is then presented back out to the marketing teams through FastStats and used to inform the next iteration of NBA messaging, meaning a complete loop is achieved. Data can be used for modelling, i.e. to predict the propensity of ‘lookalikes’ to respond to certain messages. This, in turn, can be used to inform future message delivery.
Ian Ruffle, CRM Capability Manager, RAC, states
“Because we had FastStats on board already, our NBA solution has been successful at a fraction of the budget other major organisations have paid to deliver this capability. It works for us, it’s fit for purpose and we have seen positive results.”
RAC and Occam are now collaborating on extending NBA to other customer touchpoints, including online enquiries, driven by FastStats.