Interactions between businesses and their customers are typically defined from one perspective or the other. Businesses conduct marketing activities to acquire, cross-sell and retain customers. Many tools–including CRM, marketing automation and campaign management–are utilized for these purposes. However, with increasing commoditization of products and services, providing a better customer experience has become the key differentiator among businesses competing for customer attention. Despite such terminology as customer experience management and customer relationship management, however, businesses cannot manage the customer experience. Indeed, the power has shifted to customers, who want to be treated on their own terms.
Another term, Next Best Action (NBA) reflects the efforts of marketers to present the offer that a particular customer is most likely to accept, or any content, moreover, that is most appreciated by the customer. When the right offer is delivered to the right customer, via the right communication channel, the needs of marketers and customers are aligned. But when customers receive irrelevant offers and messages, it damages both their experience of the brand and the brand’s ability to sell products. When done well, NBA dramatically increases customer acquisition rates, cross-sell revenues, as well as long-term customer loyalty and value. It also provides a superlative experience for customers, underlying the point that what is best for the customer is also best for the business. So what do marketers need to do to implement an NBA strategy?
Enablers of Next Best Action
Figure 1: NBA Functional Architecture
Figure 1 depicts the high level functional architecture for NBA solutions. Let’s take a look at each of the key enablers of next best action, noted in the blue shapes above:.
Data Integration. First and foremost, NBA relies upon extreme personalization. The next best action for one customer is usually not the best action for another. Personalized messaging and offer management, however, requires a tremendous understanding of one’s customers. For most companies, customer data resides in disparate databases. Data may be siloed by business function (e.g., marketing, sales, customer service, and finance), business unit, or communications channel. The implications of fragmented data are numerous. When various functional departments have their own data sets, their interactions with customers cannot be coordinated in a consistent manner. This results in a fragmented customer experience for the customer. When each business unit or product group has separate databases, they cannot offer the relevant cross-sells to customers of other business units, inhibiting the growth of Enterprise revenues. When different channels are unaware of the customer’s activity in other channels, moreover, offers cannot be coordinated across these channels, also resulting in a fragmented customer experience and reduced revenues.
As customer journeys move from one channel to another, businesses need to provide continuity. For instance, if customers are trying to accomplish something on the company’s website, and encounter difficulty, they are likely to call the call center for help. When a customer service representative answers the phone, s/he should know what the customer was trying to accomplish on the website and where he or she got stuck, without having to ask the customer. As another example, if a customer already accepted an offer by signing up on the web, s/he will not be happy to receive the same offer via email. Offers that have been accepted in one channel must be suppressed from appearing in other channels. Each channel, moreover, needs to reinforce the messaging in other channels in order to efficiently convert prospects to customers and customers to more lucrative customers.
Unfortunately, the data not only resides within different databases, but is typically organized in disparate formats, where it is difficult, at best, to integrate the data to develop the needed 3600 view of customers. One solution is to develop an integrated customer data warehouse, leveraging ETL, whereby the data is extracted from the original database, translated into a common data format, and loaded into the integrated data warehouse. Of course, this requires a tremendous amount of storage and effort. When new data sources become relevant, these too need to be integrated in a similar manner. A more recent trend is to leverage “big data” or “in-memory computing” solutions. These solutions are able to take disparate data from each of the data sources and place them in computer memory where they can be analyzed and acted on in real time–to help derive the next best action.
For data to be useful toward presenting the next best action, it must help predict the likelihood that any given offer will be accepted by a given prospect or customer. Based upon the patterns of data that are mined, predictive analytics are used to score each potential offer or other form of content. First, predictive models need to be built that demonstrate their ability to predict outcomes of a marketing campaign. Second, the relevant data for each customer or prospect must be entered into the model. Models are typically modified based upon new data about campaign outcomes for continuous improvement. The results of predictive modeling comprise one of the key inputs into making the right decision about the next best offer for any given individual.
Business Rules. Every business has a set of rules that dictate how and when offers are presented to prospects and customers. For instance, to ensure customers are not inundated with more offers than they care to see, businesses establish rules as to how often customers can be contacted, how often an offer should be presented, etc.They often apply other types of rules as well. For instance, with all other things being equal, a business might decide to offer the most profitable, high-margin product or service to the customer. Industry regulations must also be translated into business rules to ensure businesses are compliant with the law. In the B2B world, contracts between suppliers and customers may dictate the types of products or services that suppliers can offer to their customers. Some contracts might stipulate, for instance, that a supplier may only offer a basic service or product to the customers’ employees. This requires a business rule to be established to suppress any up-sell or cross-sell offers for such customers.
Decision Engine. Once the competing offers or messages are loaded into the system, a platform is required to make the decision regarding the next best offer or message for a given customer based upon the available data, predictive analytics and business rules. These can range from generic decision engines to complete campaign management solutions that address the entire process. Some of the leading decision engines are Oracle RTD (for real-time decisions), SAS Real-Time Decision Manager and SAP Real-Time Offer Management.
A more complete multichannel campaign management solution includes some type of decision engine, but also provides list management, campaign management and channel delivery across multiple communications channels. These solutions are either sold stand-alone or as (somewhat) integrated product suites. Examples include IBM Campaign, which handles outbound communications channels (e.g., direct mail, email, SMS texts, outbound phone calls, and mobile notifications) and IBM Interact for inbound channels (e.g., web/mobile, social media, inbound call centers and point-of-sale or retail). Teradata also has complementary products to manage multichannel campaigns. SAS Marketing Automation handles both inbound and outbound campaigns and integrates its data science tools to improve predictive modeling. Adobes provides perhaps the most complete (and most expensive) solution—the Adobe Marketing Cloud—which includes modules for Campaign Management (outbound), Experience Management (inbound), Social Media, Analytics and others. Both Gartner and Forrester publish ongoing reports comparing the various platforms available from these vendors.
Whether a company needs to invest in a real-time decision platform depends largely upon delivery channels and the frequency of interaction with the target audience. . For instance if a bank is merely sending out occasional statements with marketing messages, only batch processes, easily handled by mainframe computers, are really required. However, for companies who have a broad number of offers that change rapidly in importance or value, real-time decisions are required. This is especially true when leveraging highly interactive channels such as the web, mobile and social. When an individual is on a company’s website, for instance, offers should be updated in real-time depending upon the latest actions of, or information about, that individual.
In summary, marketers and their customers have mutual interests. A customer’s experience is derived from the sum of their interactions with a brand. The most relevant and timely offers or messages will not only provide a better experience for the customer, but also increase take rates on marketing offers. That, of course, leads to increased customer acquisitions amd sales as well as greater customer retention and lifetime value, the key goals of marketers.