By Mike Poyser
CLM represents the journey that a customer goes through in their relationship with your brand. There have been many different ways to describe these customer interactions, and the specific stages of a customer lifecycle will often vary by industry, sector and each individual company. However, at a high level, they fall into three main phases:
- Attract – Identifying, acquiring and onboarding customers
- Grow – Encouraging product trial, repeat purchase and adoption
- Retain – Customer engagement, loyalty, advocacy and churn prevention
In reality, the lines between these phases will always be blurred, and there is no such thing as a specific “desired customer journey” as each individual customer’s journey with a brand is unique. But by using data and analytics, it is possible to look for trends in the ‘typical’ lifecycle and, more importantly, influence a customer as they move along the different CLM phases.
Managing a successful CLM program requires the organization to be completely aligned on the benefit that this approach provides. It requires the different functions within a company to work collaboratively to design and operate the ways that the company can influence a customer’s interaction with the brand.
This can be structured in a multitude of ways, but typically we see best practice as the Marketing team taking the lead in defining the different CLM programs and customer touch points that are to be influenced. Analytics are responsible for providing the customer insight to facilitate CLM program design and to then develop the predictive models to be used within the programs. Data Management owns the provision of the required data to enable the analytics. Operations ensure that each program is set up in an automated, scalable fashion, to allow for rapid deployment of communication across multiple channels. Digital owns the majority of the channels, whether directly or through third parties. And underpinning all of this, IT plays a critical role in ensuring the right technology platforms are put in place, so that this process can be as seamless as possible.
Building a CLM program therefore requires considerable cross-functional working, and to develop this takes a significant investment in time and resources for a company. However, once up and running, results will quickly demonstrate the value of this approach.
Within the ‘Attract’ phase of CLM, data can be leveraged to dramatically improve the success of customer prospecting. Combining data on retail transactions, loyalty program engagement, campaign responsiveness and channel preference, as well as using predictive models, we have seen improvements in the success of acquisition reach by over 300% within the energy sector and increased response rates of 4.5x in the car rental space. In every case we have worked on, there is always an improvement in acquisition when companies take an analytical approach. Combining data sources also leads to a significant improvement in the model performance.
Once customers have been successfully acquired and on-boarded, the ‘Grow’ phase is about developing customer habits and engagement. Typically this features a strong emphasis on cross-sell or up-sell, and encouraging customers to trial new areas.
Understanding a customer’s “next best action” is hugely important. Determining what the best offer or product to communicate to a customer takes a similar approach to acquisition, and sees similar results. Next best action can be used to determine what to feature when a particular customer logs in to the website or to include in a communication. This can be based on that customer’s past behaviour, or through analyzing ‘look-a-like’ customers to infer the action with the highest chance of success for an individual. In online retailing, we have seen behaviour from regular shoppers used to determine the next best action for less frequent customers where the data is sparser.
While the ultimate goal is to be so successful in the ‘grow’ phase that customer engagement, loyalty and advocacy are natural outcomes, it is impossible to stop some customers from leaving a brand. This is the natural lifecycle, and while it cannot be prevented entirely, it can be dramatically decreased through an analytical approach. The maxim again here is to use as many variables and data sources as possible, and combine these into a regression model to improve the success of predicting likely churners. When done well, such an approach can not only significantly improve churn prediction, but it can also enable prediction to be done early enough with the use of fewer data points. The value of such “early life churn” prediction is that there is much more runway to put in place the right engagement strategy for that customer so that potential to churn becomes a distant memory.
Mike Poyser is the vice president of analytics at Aimia Inc.