Data-Driven Customer Interaction: Opportunities and Roadblocks

Fastwater Rapids vol. 1.1, 21Aug98

In the age of targeted and one-to-one marketing, it is clear that customer data is one of a company's most important assets. Information about customer demographics, purchase history, and behavior can be used in sales, marketing and customer service to gain more revenue and increase customer loyalty.

The internet seems like an ideal medium to conduct smarter, data-informed customer interaction.

On the internet, you ought to be able to:

provide customer service about the products your customers have, based on your company's records
use the wealth of information on customer history to provide relevant personalized recommendations and offers
integrate information about customers' responses to ads and offers to customize ad campaigns and promotions in real time
use data from on-line information and customer history to respond intelligently to customer's e-mail, or to provide on-line, live customer service

Listening to the excitement about one-to-one marketing and hearing top-level marketing messages from vendors, you might think that you could achieve these goals easily today. You can get started on the path today - many companies are - but you can't get all the way there from here. In practice, there are substantial barriers to achieving the goal of data-driven customer interaction.

This issue of Rapids provides an overview of the technical and organizational barriers, and makes recommendations about steps to move from where you are now to where you'd like to go. In the coming months, we will further explore the issues in this article, drawing on the Fastwater case study research and providing more details about technology strategies to help with your business goals.

Technical Barriers

Today, there are many steps that companies are taking to offer personalized content, services, and promotions on their websites. Most personalization today is generated by the user's past and current activities on the web (e.g., ads displayed based on search keywords; personalized promotions and recommendations based on web purchase history; offers based on customer-generated profile). These are important first steps.

The next steps toward achieving the goal of data-driven customer interaction include building feedback loops, whereby broader sets of customer actions on the web can affect individual service and overall campaign decisions; and where web interactions are integrated with off-line customer interactions and history.

Today, getting to these next steps is difficult, and sometimes impossible, because important pieces of the technology tool set are immature or still under construction.

data analysis tools need to get faster and easier to use
web data needs to be integrated with other sources of customer data
web interaction tools need pipelines to customer data

Data Analysis Tools. For most businesses operating on the web today, the main challenge is simply figuring out how to deal with the vast amounts of data generated by web activity - the E*Trade web site, for example, generates 5 gigabytes of usage data per day. The volume of data will only grow as more customers use the web. Business-related analysis features are becoming a focus of competition among web site analysis products, so the state of the art will improve. Moreover, there are new classes of data mining and analysis tools - from suppliers including Epiphany and Personify - that help companies understand web-based data, integrate web data with other sources of customer information, and use the data to make decisions for sales, marketing, advertising, and customer support.

Integration with other sources of customer data

If your business started before the web, the web may represent only 5% to 10% of your overall customer interactions. If your business is more than a few years old, web interactions are probably only a tiny fraction of your repository of customer data. To inform web interactions with customer data, therefore, you will need to integrate the web with existing collections of customer data:

Data warehouses of legacy and transactional data
Database marketing systems
Sales/customer service systems

Data warehouses. Many companies already have efforts under way to build data warehouses for broad sets of strategic purposes - accounting, distribution, manufacturing, as well as marketing. It will become important to integrate web data into the overall company knowledge architecture. The main gap is technical - today, the leading web data analysis vendors are working to open their systems to conventional data mining tools.

Database marketing systems. Many companies don't have a centralized data warehouse strategy - but they do have ongoing, direct marketing efforts that it would make sense to integrate with the web. However, most marketing departments' campaign management tools are fairly low-tech today, and are not up to the task of managing complex campaigns in multiple media. Emerging "marketing automation software" is starting to help organizations plan campaigns, to multiple media, in a more efficient, coordinated fashion. This technology in a very early stage - we spoke to a couple of the hot start-ups in this area, and they have less than ten customers each.

Sales and support systems. Many companies have deployed sales and customer support automation systems that keep track of customer interactions with human representatives (these systems are sometimes called Front Office systems, Customer Relationship Management, or at the high end, Enterprise Relationship Management). The front office software market is maturing - as evidenced by the recent spate of mergers and consolidations. As this happens, these systems are becoming increasing common, and they're becoming the primary repositories for day-to-day customer interaction data.

It makes sense to integrate the sales and customer service data with web interaction systems, so that a customer can see tech-support information, based on the products she owns, updated for the most recent support activity or product purchase. Today, many web customization features depend on customers to re-enter their own profile information - to tell the system, for example, which products they own - even though the company has that same information elsewhere. One area that is particularly confusing at the moment is sorting out the claims of leading front office systems to provide web interaction capabilities - these web capabilities in most cases are still quite rudimentary.

Using data to drive web interaction systems.

The most important step is to use the customer data to drive interaction online. Today, most customized interaction comes from customer-entered profiles, or from information gathered on-line. Here, too, there are still technical barriers. The leading web customer interaction systems, from companies like Vignette and Broadvision, are just beginning to open their systems to accept customer data from external sources. Other components, such as e-mail communication tools and "chat tools" for online customer service, are also just beginning to open to off-line sources of data. Next week, we will look at some of the ways these systems can - and can't yet - use this data to drive and inform customer interactions.

Process Barriers

In order to make use of the technology as it evolves, companies will need to adapt their business processes.

data-driven marketing decision processes need to be faster
companies need greater control over outsourced marketing processes
web marketing and business management need to be tied to overall business goals

Companies need to hire web site analysts and web marketing managers to analyze web data and use it to craft marketing plans. For example, Amazon.com is currently hiring a "Manager of Planning & Analysis Marketing" who will be charged with "analysis of customer acquisition, customer retention, and revenue growth strategies," among other responsibilities. These types of jobs will be different from traditional direct marketing planning and analysis positions in that the decision cycles will be much faster.

Traditional direct marketing processes are too slow to be useful in web business; the typical direct marketing campaign takes three to six months to plan. Moreover, web-based marketing efforts are often disconnected from the rest of the company's marketing plans. Outsourcing can be a barrier to the integration of web-based marketing corporate marketing strategy. Today, many companies outsource their direct marketing fulfillment process and database management. Consequently, campaign planning is slowed by inter-company collaboration. Moreover, valuable understanding about the data and the marketing process might reside with an outside contractor. As customer data becomes more strategic, and speed becomes more critical, companies will need greater involvement with their own direct marketing process -- although print fulfillment and data cleansing may always remain outsourced.

The most important pre-requisite to integrating the web with the rest of the business is a thorough business commitment. Many web sites were started as "skunk works," at a remove from the company's existing business activity. While this might be an effective way to get started quickly, it stands in the way of using the web to best advantage. Companies need to commit to integrating the web into their overall business goals. This commitment should extend to company executives as well as departmental decision makers.

Laying the Groundwork

This business commitment is not the same as a top-down initiative to build a comprehensive, completely integrated data-driven system tomorrow. That sort of full-force effort is risky today. However, there are a number of practical steps you can take to start moving from where you are now to where you'd like to go.

develop faster process of making marketing decisions. If your web marketing efforts are not integrated with the rest of the company's marketing today, bring the right people together. If you're working with service bureaus, make sure you have access to the information you need in-house.
evaluate vendor claims closely. If you are looking to integrate your web site with existing sources of customer data, make sure you understand what integration capabilities your vendor can support today and the direction of their plans. If there is something you need, ask for it. At this stage of the market, your dollars vote.
make decisions based on your company's goals. Every web business has its own set of goals driven by the company's business goals. Integrating customer data to drive your web site involves a good amount of custom development today. Focus your resources on a few top goals, rather than attempting to do everything at once.

Highway under construction

The long-term goal is to build web systems that use customer data to inform customer interaction, providing better customer service and more effective sales and marketing. But the highway from here to there is under construction. Despite the roadblocks, it is important to plan a route, because data-driven customer interaction will become increasingly important to profitability and survival. Developing a strategy today will help you take advantage of technical capabilities as they emerge and develop organizational processes that will support success.

Bob Green, a Formula V race car driver and racing instructor, offers driving advice that pertains to this situation. "If you're looking at something close, you can't see things far away. But if you're focusing on something far away, you can still see the things that are close." Today, managing a web business requires attention to many details right in front of you. To make good decisions, you need to keep an eye on where you are going.