How clean is your customer data?

Publishers are relying on data more than ever to drive their business. Data fuels advertising campaigns, audience development efforts, even editorial decisions. There’s a heaping, growing pile of information about your audience and your prospects – and collecting, storing and managing all of it is a growing challenge.

The two biggest issues: Keeping your customer data fresh, and finding a way to integrate multiple, disparate sources of information. These challenges are the major impediments to developing a clear, consistent view of your audience in order to increase the effectiveness of your sales and marketing programs.

The business case for clean data

But how do you go about organizing the mess? Every publisher has plenty of untamed data coursing throughout the organization, stored in multiple databases and maintained erratically by salespeople, web teams, editorial groups and marketing analysts. SiriusDecisions, a marketing research and advisory services provider, estimates that error rates involving customer or prospect data at a typical B2B company (not specific to publishing) are as high as 25 percent.

With limited resources, publishers may be tempted to say a 75% accuracy rate is good enough. But consider the escalating costs of ignoring bad data. Sirius offers a “1-10-100” rule: It costs $1 to verify a record as it is entered into a system, $10 to cleanse and de-dupe it after it’s been entered, and $100 if bad data is not updated – because mistakes based on that misinformation will continue to re-occur.

“When you put bad data in, it deteriorates the value of all the other sales and marketing assets you put in play,” says Marty Sunde, senior vice president of product management for Acxiom, a provider of data integration and database marketing products and services.

Publishers can (and should) build a real business case for improving the quality of customer data, says Sunde. Consider a customer file with 2 million names in it. Improving the quality of that data by 10% will expose your offer to 200,000 additional prospects. With a 5% conversion rate, that translates to 10,000 more purchases (or downloads, or visits to your site, or whatever other activity you’re promoting).

“You need a continuous focus on having the cleanest, clearest view possible of your customers and your prospects,” says Sunde.

Need more convincing? Sirius estimates that B2B companies that address data errors at the source of entry can show a 25 percent increase in converting inquiries to marketing-qualified leads.

Integrating disparate data

The second big challenge for data manages is integrating information from multiple sources. The problem, of course, is not new – for decades, companies have struggled to maintain consistent records of customer names, addresses and phone numbers in order to execute direct marketing or telemarketing campaigns. Add in Web activity, mobile numbers and social media, and the process of integrating that information becomes even more daunting.

“More data is better – as long as you can find a way to tie it all together,” says Sunde.

Service providers such as Acxiom are continually upgrading their consumer and business databases with new sources of information. Acxiom taps into 30,000 different data sources to maintain its customer database, which houses 300 million names in the U.S. alone (500 million including Europe). And that’s just for offline information such as name and address. This summer, the company plans to release a new version of its AbiliTec customer recognition technology that incorporates email, phone and cell phone information.

“In one pass, we’ll be able to integrate offline, household behavioral data with all the information a company has on transactions, click-throughs or other online activities and associate it and rationalize it,” says Sunde.

Acxiom worked with Hearst Digital last year to combine Hearst’s offline and online customer data into a unified marketing database. The mega-database “allows us to sync up our marketing efforts,” says Chris Wilkes, vice president of marketing and audience development for Hearst Magazines Digital Media. “We can use it to expand our audience – by analyzing what our customers really like, we can find people who are like them but are not consumers of that product and do more intensive marketing to those people.”

Trade publisher PennWell has adopted a similar approach. While each brand maintains its own database, files from brands covering the same industry, such as electronics, are rolled up into a “market validated” database used for cross-promotion, says Gloria Adams, senior vice president of audience development.

As with improving data quality, integrating data sources can lead to some clear business benefits. Sirius estimates that unifying customer data can lead to a 12.5 percent uplift in conversion rates to the next stage in the purchase funnel for B2B companies.

Practicing good data hygiene

How do you improve your efforts to manage customer data? There are plenty of data management “best practices,” and while most are not specific to media companies, the general concepts will apply to most publishers.

If you’re a data laggard – meaning you lack any formal processes for managing customer data – researcher Aberdeen suggests the following steps:

• Start by using data for activities that will impact revenue. This will demonstrate the value of these data-driven programs to help you justify further investment.
• Develop a periodic process to scrub the customer database. The importance of good data hygiene can’t be understated. If you don’t regularly clean up your existing assets, you won’t gain many useful insights.
• Invest in tools to help analyze customer data. Data analysis and marketing automation technologies will help you improve the efficiency and effectiveness of your marketing campaigns.

For companies that are more evolved in their data management activities, Aberdeen suggests the following steps to reach “best in class” status:

• Implement a formal data hygiene strategy. This is about managing the current and future health of your data, with repeatable processes for cleaning, appending and de-duping data.
• Engage multiple departments for data analysis. Encourage collaboration between marketing, IT, finance and sales (and, for publishers, editorial as well).
• Democratize all customer data. A centralized database “opens the door to performance measurement and optimization of multichannel campaigns,” the Aberdeen report states.

Another helpful tip is categorizing potential data quality problems, which can help you prioritize ways to address the issues. Progress Software’s John Wilmes offers four common types of data quality flaws, in order of increasing difficulty to fix:

• Invalid data – an impossible phone number, a nonexistent postal code, a birth date in the future – is easier to find and fix than other types of dirty data. It can and should be detected at the time of entry.
• Incomplete data is more difficult to find. Its detection requires a model of data relationships, and may involve multiple applications.
• Inconsistent data can be even harder to find, because its detection requires even more inside knowledge (substitute "rules" or "metadata" if you prefer). Even if the web of data for a customer is complete, it can still be inconsistent.
• Incorrect data can be the most intractable, because much incorrect data will not be detected by validation, completeness, or consistency checking. Even though it is valid, complete, and consistent, it's just wrong.

These types of problems cannot be addressed strictly through technology. Improving your data-management efforts often requires changes in the way employees view and process the information they are responsible for.

“It’s important to get everyone to accept a common view of the rules and guidelines that govern data management,” says Sunde. “There will always be many manifestations of prospect or customer names. The question is, which one do we declare the lead and how do we organize around that?”

The key, Sunde adds, is understanding that data quality is a continuous process, not a one-off or intermittent activity. “Those who do it best pay a lot of attention to it every day,” he says.

And while no one is likely to achieve the hallowed “single view of the customer” – despite what data management vendors may claim – that’s really not the point. You don’t have to do this perfectly – you just have to do it better than your competitors.