Continue in 2 seconds

Data Integration: Making Multichannel Real

Published
  • March 18 2002, 1:00am EST
More in

In my last article, "Channel Integration: How to Hear the Sweet Sound of CRM Success at Last,"(DMReview.com February 1, 2002) I talked about how close many businesses are to achieving real benefits from their CRM investments. Whether using a one-brand, platform strategy or a best-of-breed approach, companies that focus on integrating across customer touchpoints and achieving channel integration will be the ones that realize the promised gains in customer profitability and operational efficiency.

In this, the second article in the series, I’m going to talk about achieving integration from a data perspective. (The final two installments will focus on application integration and information access, respectively.)

 

You Know Who You Are

Hopefully, based on my last article, you recognized what kind of CRM platform strategy you’re following. That accomplished, it’s time to talk about your data.

Ask yourself a simple question: Is multichannel a big issue for me?

Your answer should be based on your customers’ needs. How often do they interact with more than one sales or service channel within a given twenty-four hour period? If you’re a bank, brokerage firm or a telco, chances are the answer is "often." Your customers visit the ATM to withdraw cash, pay bills online and then contact your automated call center to resolve issues they may have with complex transactions. For them, overnight processing simply won’t cut it. Without some form of real-time processing, customer service representatives will be ill prepared to help customers that have taken action via one channel and are now trying to utilize another. It’s possible that your best customers will sit on hold while your agent calls another division. In that scenario, multichannel integration is clearly a major issue.

When financial transactions are involved or when goods and services are changing hands in a time-sensitive environment – for example, when you order a baby shower gift from Amazon.com under next-day shipping, then frantically track your package in anticipation of the party the next afternoon – you have a multichannel integration issue with respect to data integration. It’s an easy problem to spot. Account balances rise and fall. Goods are shipped, routed and received. And customers log issues or request services through the wrong channels.

Another issue plaguing large organizations is the notion that everyone is in customer service now and should be prepared to solve customers’ problems. No other constituency has been more abused by this expectation than the field service representative.

Consider a recent client of mine who had what we like to call a "good" problem: New customers were subscribing to their service at an alarming rate. Business was booming. As a result, inbound calls to manage installation processes, troubleshoot service issues and handle billing exceptions increased rapidly, before they had a chance to prepare.

The result is not hard to imagine. Emotionally frayed service reps found themselves struggling to piece together disparate customer data to ensure a successful installation, while emotionally frayed customers fought to maintain their patience as they tried to recreate their order specifications for the service rep. This classic scenario also had customers repeating and reentering their phone numbers, addresses and credit card numbers enough times to make them hoarse or to give them blisters. In order to ensure a seamless customer experience – from order placement to installation to customer care – the client had to make severe process and technology changes.

When it has to be right in real time, you have a data integration issue. One caveat: Make sure that your real-time needs are the rule rather than the exception. Achieving real-time data integration is akin to open heart surgery. The survival rate is outstanding, and the health improvements are dramatic. But it’s nothing you want to embark upon if you’ve only got a simple case of intermittent heartburn.

But Do You Know What You Need?

Now that you’ve established your challenge, the logical question is: What do I do about it?

On the one hand, you may choose to implement a single database strategy, the operational data store (ODS). Or, depending on your current situation – the investments you’ve already made, the unique geographic requirements of your business or that recent spate of mergers and acquisitions that has complicated your data architecture – you may decide that data synchronization (the ability to seamlessly exchange, merge and synchronize databases across multiple database platforms) is the way to go. At the end of the day, the question can be reduced to this: Will you move to one ODS that each CRM application will read and write to or would you rather transfer data from one source to another, simply making it look like you have one database when you actually have many.

As always, there are several factors that will influence your decision. Consider the following questions as you chart your course:

How complicated is your current data architecture?

Mature industries are a blessing and a curse. On the one hand, customer relationships and the carrots and sticks that govern them are usually highly evolved. On the other hand, the spaghetti-like tangle of discrete source systems that have arisen over the years is an IT department’s worst nightmare. The more complicated the data architecture, the more I would recommend an ODS environment. When you’re synchronizing multiple systems, it becomes very difficult to map all permutations of systems and data, and to keep the architecture up to date. The ODS strategy solves this by creating one central place to read and write, bypassing the spaghetti, as it were.

How challenging will data transformation be?

One issue that differentiates data warehousing from a real-time approach is the ability to perform complex transformations, calculations and validations. In a real-time environment, there may or may not be enough time to achieve the complexity possible in a data warehousing batch environment. A centralized ODS can ensure that data is "close;" that way, if a calculation or validation requires information from several systems, performance may be better. If customer scores or customer valuation metrics are served to each channel, the ODS may be better positioned to accept those metrics from the data warehouse and disperse them to the channels. A data synchronization approach would need to either calculate the field on the fly, retrieve it from the data warehouse (which is not optimized for transaction processing) or have the data warehouse send the measures to all channels that need them.

How fast do you need to move?

Speed is another major factor to consider. Centralization will require achieving consensus across multiple departments. Centralization is political and, as such, will ask you to campaign on its behalf. Are you excited about trying to drive agreement on data definitions such as customer, shipment or product? Synchronization, on the other hand, is operational; it makes its own case without asking permission. Near-term efficiency is easier to sell than a long-term solution that may be perceived as bordering on utopian. The centralized ODS will have as many or more data conceptualization issues as a data warehouse.

How much information does each channel need? When do they need it?

Variable requirements may suggest a variable approach. Which of your users really require up-to-the- minute customer data? How much data do they really need? In what form? Synchronization approaches allow you to focus your efforts on locating, manipulating and delivering the most relevant data in its most useful form. You can harvest the trees and not worry about the forest, so to speak. Centralization and an ODS, on the other hand, asks you to account for the entire landscape, roots and all. So how much do you really need?

What part of "no" don’t you understand?

You’ve mastered opting in, but what about opting out? What happens when a customer wants the call center to unsubscribe him from an e-mail newsletter? Can your CSR get anywhere near the data needed to fulfill that request?

I’ve recently received six e-mail solicitations to join a popular online movie rental service. All were from rental service’s marketing partners. I received these solicitations despite the fact that I already joined the service three months ago! It seems as if their marketing partners are operating in a vacuum. Yet no matter how many times I complain to the parent company, they can’t seem to call off the dogs.

When data is everywhere, customer satisfaction is nowhere. Permission-based marketing requires the synchronization of all channels to make sure that each channel can handle opt in or opt out, then inform other channels – and partners – immediately. Remember, the customer views you as one company, not disparate departments. Consequently, the requirements of permission-based marketing and privacy come into play for your multichannel endeavors.

What level of CRM are you aspiring to?

CRM means different things to different companies. If you’re only interested in automating and monitoring sales force activities, issues such as real-time value calculation won’t be terribly important to you. But if you envision a day when your shiny new customer service professionals are making on- the-fly offers to your best customers – a time when customer value management becomes an integral part of each transaction – then real-time data access and integration had better be in your plans, no matter how challenging it might appear.

Real-time decisioning implies quick access to key customer information as well as understanding the context of the transaction. A data synchronization environment could more easily combine contextual information with customer value metrics by storing the data locally in the form the application requires.

Remember These Simple Things

As you consider the many factors just outlined, there are a few simple things to remember about achieving data integration:
  • It’s always hard.
  • The more consensus you must achieve, the longer the project will run.
  • No matter how beautiful your architecture is, true success will reside in the eye of the end user and, ultimately, the customer. Never neglect the interface.
  • In a phased approach, start with the basics: customer information (including product profile), contact and response history, and marketing/promotional history. Worry about constructing customer value metrics and handling unstructured data only after you’ve conquered the essentials.
  • The solution may be a hybrid. Many companies don’t completely ditch the application databases, since they fear the ODS could be a single point of failure. Instead, many companies may synchronize to the ODS or build a thin ODS.

Now that you’ve heard about channel integration and you’ve considered your database, be sure to check back in a couple weeks as I explore integration from yet another angle: your applications.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access