Near Team

by Madhu Therani

Chief Technology Officer

July 17, 2018

To meet growing consumer expectations in a digitally-driven world, companies are dealing with large amounts of data and aim to create personalized consumer experiences at different touchpoints to stay relevant. The various consumer touch-points of an enterprise generate a wide variety of different types of digital data and are also potentially great interaction points to improve an enterprises engagement with customers. Furthermore, as enterprises adopt digital technologies for their internal and external processes, a large amount of digital data is available about consumers, products, competitors and the marketplace within and external to an enterprise. Establishing the business-value that can be generated from this data and developing or deploying viable technologies in digitally-driven processes is the current focus of most enterprises. Artificial Intelligence (AI) driven technologies have a critical role to play in exploiting the ever-growing data deluge to improve products, services and customer experiences.

Numerous challenges exist as enterprises embark on this next phase - managing the flow of incoming data of all types and most of it being unstructured, data storage, aggregating and integrating data. Harnessing the integrated data and deploying it in various decision-making points by utilizing appropriate AI-based technologies across the organization is another key problem. A key aspect underpinning all these efforts is the need to maintain the data fresh and relevant - which varies across different touchpoints. Apart from the scale of the data - figuring out which bits of data are relevant and will have a business impact - is something every enterprise has to solve for at the earliest.

Harnessing the data with appropriate AI-technologies can support a variety of interactions involving an individual customer or a cohort of customers with similar needs and preferences. Data-driven decision making can enable better customized product or service design, optimize manufacturing or service delivery, enable better marketing and build customer awareness, support pre-purchase, purchase (online and offline) and post-purchase activities. For knowledge-intensive and/or highly priced items, post-purchase lifecycles can involve many customer interactions which need to be managed effectively. Data-driven technologies enable a new level of customization/personalization to be embedded in these days of mass manufacturing - where everything is standardized. This helps product and service offerings to be highly differentiated. Systems such as Alexa, Echo, Google Assistant, IOT systems such as different kinds of sensors and more - allow embedding different kinds of smarts in products and services. As the intelligence embedded product/service eco-system evolves, customer are going to expect much more from enterprises.

As a first step towards the comprehensive data-driven effort to improve customer experiences, it is important to have a good understanding of the issues involved.

Harnessing and Activating Customer Data - The Issues

Customer data currently sits across many silos in an organization - the marketing systems, the website, the Point-of-Sale systems, the returns systems, service and maintenance systems and many more. Integrating this data and providing a unified framework to access this data in a standard manner across the enterprise is essential. The data exists in many formats - structured, unstructured and a consumer has multiple identifiers across the different systems. Getting a unified 360degree view of a customer is a key first step.

In the context of modern marketing systems, this customer data is of two major kinds - existing customers who need to be retained and prospects. Data about each kind is available both internal to an enterprise (called first-party) and external to the enterprise - from a variety of vendors and data aggregators (called third party). Pulling this all together in a meaningful manner and analyzing it for detecting individual and group trends and patterns at scale is a major task in knowing more about your customers and prospects. Much of the data needs to cleansed and curated before it starts becoming valuable. Furthermore, data goes stale - as customers change their digital identifiers and personas. Keeping track of all these changes and propagating their effects is extremely important. Furthermore, the data needs to be made available at different touch points with the customer and also power a variety of decisions along the consumers journey within the enterprise’s purview. Once this data is made available, how to utilize it intelligently is where the AI systems play a key role.

Enabling Decision-Making - the role of “Intelligence”

Customer-centric decisions coming again in two primary kinds - generic decisions that apply to groups as a whole and individual decisions - customized interactions - with an individual. AI systems play a role in both of these. Much of the recent focus has been primarily on the second one which we will discuss briefly later in this section.

Decisions such as what product features to build, what product options to provide, what price points to refer to, what kind of merchandise to store at different retail locations, what combinations of products sell together, what kind of promotions etc are all powered by the availability of data. AI, machine learning and optimization systems help analysts explore the variety of choices available for each decision and then make the appropriate decision. For example, well known techniques such as market-basket analysis, pareto analysis and such power these decisions. Apart from understanding consumer behavior with your own product, it is important to understand how customers engage with competitors.

With the availability of new location intelligence systems, enterprises can gain granular competitor insights in real-time like - who are the audience walking into your own vs competitor stores? What are their profiles, how long do they commute? How often do they visit? Do they make purchases or just visit to check out prices and buy online? These insights can help you analyze and make long-term business strategy in capturing competitor audience, consumer retention, pricing, product design, product features, store format, identifying location for your store etc.

Decisions that engage a customer at each individual touch point are far more complex. Firstly, reaching out to prospects or engaging existing customers is a major task in any enterprise. Marketers engaging in omni-channel marketing monitor all channels and engage customers appropriately based on the channel. Tracking an individual across different identities and engaging appropriately is a complex task. Different types of messaging and media types are evaluated on-the-fly by analyzing consumer responses to different messages and taking appropriate action. Making product recommendations to a customer - both online and offline - is powered by AI-based recommender systems. Similarly, when a consumer is in the transaction phase, pricing and discounts are offered on the fly - based on a real-time assessment of the long-term value of the customer. Even loans can be offered real-time based on dynamic credit assessment. Post purchase engagements are mediated by a variety of text, video and voice-based channels. Service reminders and upsell/cross-sell opportunities are all triggered based on data and usage analysis. Systems that analyze emails, voice interactions, and web interactions are powered by AI/ML systems that look for key events and trends for each individual and trigger appropriate engagements.

A key aspect to realize here is that the “intelligence” embedded in these decision-making systems address the following key items: a) which data is relevant for each decision, b) how to get the data and summarize it, c) how to utilize the data to specify a customer-specific action that is relevant, d) how to deal with noisy/incomplete data, e) how to make a choice that satisfies the customer and is also economically beneficial for the enterprise. These steps need to be executed - some apriori and some at the point of engagement. Engineering these AI system requires a variety of issues to be addressed, which are quite complex. Furthermore, as consumer behaviors evolve, these systems need to adapt - leading to another layer of complexity. We are still in the early days of building such comprehensive adaptive systems.

Concluding Remarks

Exploiting customer data in enterprises requires a holistic and comprehensive approach. From gathering appropriate data, harnessing it and building AI-driven systems that deliver real value requires major investments in understanding the overall business processes and the underlying technology. Revisiting the core competencies of the enterprise and rethinking the overall consumer journey is essential to exploit advance AI and digital technologies effectively.

Contact Us to use data for superior decision-making.