In my last blog post
, I examined the increased importance of personalization strategies and what it takes to create an effective, multi-dimensional approach to personalization. In this post, we’ll discuss whether or not creating a complex multi-dimensional personalization (MDP) strategy is worth the investment, and if so, how we go about executing on it.
Is MDP Worth the Investment?
Some companies might think they can apply a machine learning capability and generate a single algorithm to try and develop personalized interactions. The limitations of machine learning techniques
notwithstanding, the variability required to drive an effective MDP strategy – newer and more robust channels, growing volumes and complexity of data, as well as evolving analytical techniques – is too dynamic to expect this approach to work. Yet many organizations still take this view. In fact, studies show that although customers expect more in terms of personalizing relationships, the experiences are still not meeting
expectations. Nevertheless, according to a recent PwC report
, the value of getting multi-dimensional personalization right is significant:
- The report states that 54% of U.S. consumers say customer experience at most companies needs improvement.
- It explains that bad experiences with brands are driving consumers away — fast. One in three consumers (32%) say they will walk away from a brand they love after just one bad experience.
- 82% of Americans want more human interaction with employees.
- More than 70% of consumers value speed, convenience, helpful employees and friendly service.
- Convenience — the seamless transition from tablet to smartphone to desktop to human — is a baseline expectation.
- Great individualized experiences allow brands the opportunity to charge up to 16% more on products and services – while still increasing customer loyalty.
The Five Dimensions of Personalization
Once we recognize the value in developing a personal relationship, we need to evaluate the available techniques to address the necessary components of each dimension. The following outlines the various elements contained in each:
- Personalized Characteristics are focused on the “who” of the interaction. This category of personalization provides a variety of very specific elements to assign a “face” to the person with whom we are interacting.
- Segments include any tactical or strategic designations to which a consumer belongs. These can be human initiated designations. For example, seniors are equal to anyone 65 years of age or older, as well as other analytically defined clusters.
- Targeted Scores include any number of assigned categories that can be used to prioritize, predict, or define values based on some predefined set of criteria.
- Preferences are characteristics that are defined explicitly (through input) or implicitly (via some analytical process) that impact how personalization elements are applied.
- Personal Traits are pieces of information associated with a customer. This includes data points like their name or address. Personal traits also include information that can be derived from other data, including their store preference or the location they most often contact.
- Optimized Collateral determine the appropriate action, offer, or message. This would best be informed by predictive analytics (e.g., this customer is most interested in this next offer) combined with business rules (e.g., if this is recommended, it must be in stock). Types of optimized recommendations include:
- Next Best Offers provide a financial benefit to a consumer that is relevant to their transaction.
- Next Best Messages include specific messaging around a subject matter. These messages have less emphasis on financial incentives.
- Next Best Events trigger an action either overtly or behind the scenes with respect to the individual. More information on real time decisioning can be found here.
- Customer Journey is an understanding of the steps a consumer is expected to take to reach a specific outcome. The following components are typically found in journey maps:
- Personas are the main characters that illustrate the needs, goals, thoughts, feelings, opinions, expectations, and pain points of the user.
- Timelines show a finite amount of time or variable phases of purchase. Examples of these phases include the traditional steps of the buyers – awareness, decision-making, purchase, renewal.
- Emotions track the peaks and valleys of the journey. Tracking emotions helps to illustrate sentiment – the positive and negative feelings the customer associates with the company.
- Touchpoints are customer interactions with a brand.
- Interactive Channels enable consistency in how brands interact with customers on a personal level, regardless of the channel. The channel you choose to reach your customer does have specific characteristics and capabilities with respect to personalization. Therefore, each interaction must consider the unique set of channel-specific requirements so the right content can be pulled together. The most common categories of channels include:
- Direct Channels are typically viewed as channels where the organization initiates an interaction to solicit a response. Direct mail, email, and SMS/text messages would be considered direct channels.
- Indirect Channels are locations where the primary method of communication is typically initiated by a consumer. As a result, they expect a personalized response from a representative of the company. Indirect channels include stores, service locations, and call centers.
- Digital Channels are typically autonomous, but they enable interactions that emulate human-like decision. These channel exchanges are initiated by either the customer or the business and can take place on the web, via kiosks, or through social media interactions.
- Chatbots are specialized visual or audio digital assistants that drive autonomous or semi-autonomous communications that leverage techniques like natural language processing.
- Business Strategy drives the kind of message the brand is trying to convey to the individual. These message types include but are not limited to:
- Sales Messaging is focused on conversions. These kinds of messages are meant to drive things like purchases, acquisitions, or renewals.
- Marketing Messaging is used to convey branding, create interest, nurture leads, and drive demand.
- Service/Care Messaging is meant to interact with customers in order to provide service or care.
Bringing these concepts together, we move from generating generic messages based on limited information, to providing specific and personalized messaging, which can only be achieved when considering a multitude of variables.
The “How” of MDP
With the five dimensions of MDP (Who, What, When, Where, Why) defined, there are various methods by which these personalization elements can be expressed by using analytical methods (The How). There are several analytical methods that can be brought together to address each dimension. Please see the diagram below for more information.
The diagram shows that MDP requires a multi-dimensional approach to applying analytics in support of this strategy.
There are two key guiding principles I want you to take away if you are looking to develop a robust MDP solution:
- The analytical methods you use are less important than the process you are trying to enable. While it’s important to have a variety of analytical tools at your disposal, do not think analytics alone will drive an automated, personalized, omnichannel communication strategy.
- Interactions should be adaptable in real time to reflect immediate feedback and the tone of the conversation. However, the majority of what needs to be determined in the moment should be derived from a centrally defined organizational approach and a single view of the customer.
MDP is not handing data over to single channel providers and letting each vendor manage a different strategy and set of functionalities for each channel. A sound MDP strategy requires a variety of technologies and approaches to enable any decision. It also needs to be part of an overall marketing ecosystem which includes things like customer data platforms
, real time interaction management tools
, and analytical platforms like Teradata Vantage.
Whichever way you approach your own MDP strategy you need to land on a solution that is scalable, extensible, semi-autonomous, and analytically driven. The worst thing you can do when trying to develop relationships with customers is to invest in technologies that your business outgrows too soon. Customers are providing you a lot of information about themselves, and therefore they require specialization and personalization. In today’s business environment, you need to know more than a name to develop that personal relationship.