Identity Resolution Daily

[Daily Post from Infoglide Software] Resist the Urge to Merge Purge Data

“Mr. Jonas published a great post yesterday comparing identity resolution against match merge, merge purge and list de-duplication systems that is a must-read for the CIOs doing all the hefty lifting in financial services, government and insurance industries. (Note: identity resolution is called “entity resolution” in IBM parlance.)”

GovernmentExecutive.com: Privacy advocates wary of data ‘fusion centers’

Riegle called the initiative ‘a novel and different approach to information-sharing’ and said privacy was a top concern when each center was formed. Wobbleton said transparency is a core part of the mission. ‘We do not want to mess this up,’ he added.”

Washington Post: Patriot Act Provisions Voided

“In a case brought by a Portland man who was wrongly detained as a terrorism suspect in 2004, U.S. District Judge Ann Aiken ruled that the Patriot Act violates the Constitution because it “permits the executive branch of government to conduct surveillance and searches of American citizens without satisfying the probable cause requirements of the Fourth Amendment.”

Leadership Journal: Privacy And Security

Writes Michael Chertoff in the DHS blog: “But what about the tension between privacy and security? Is it true that whatever we do to strengthen our security must be at the expense of privacy? It is not. Our efforts to secure our homeland need not harm our privacy. Rather, in many cases they can actually strengthen it. […] Privacy and security are fundamental rights and we will continue to defend both in our post-9/11 world.”

EFF: “Secure Flight” Returns, Lacking Privacy Protections

“When it enacted the Privacy Act in 1974, Congress sought to restrict the amount of personal information that federal agencies could collect and, significantly, required agencies to be transparent in their information practices. The Privacy Act is intended ‘to promote accountability, responsibility, legislative oversight, and open government with respect to the use of computer technology in the personal information systems and data banks of the Federal Government[.]’ Adherence to these requirements is critical for a system like Secure Flight.”

Share This

Surely you’re aware of Jeff Jonas, identity resolution’s poster boy and first real celebrity. Mr. Jonas is the chief scientist behind IBM’s Entity Analytic Solutions (EAS) and the founder of Systems Research & Development (SRD). He’s a media magnet who’s been featured in CNN, Forbes, Newsweek, NPR, Time and Wired. While he hasn’t quite made it to the cover of People Magazine yet, he was recently sought out by an NBC affiliate in Philadelphia to comment on the plot of the new NBC show Chuck.

Mr. Jonas published a great post yesterday comparing identity resolution against match merge, merge purge and list de-duplication systems that is a must-read for the CIOs doing all the hefty lifting in financial services, government and insurance industries. (Note: identity resolution is called “entity resolution” in IBM parlance.)

For example, when two insurance companies get the urge to merge, they always run into data compatibility issues when they begin to look at the databases they have currently have in place. For claims and underwriting purposes, data synchronization is not a luxury — it’s essential to the daily flow of business.

Back in August when Dutch AEGON acquired the life insurance operations of U.S. investment bank Merrill Lynch in a $1.3 billion in cash deal, our own Glenn Hopkins commented

“Instead of implementing a master data management program, I know that AEGON would be better served — and save lots of capital — if they bolted an identity resolution solution onto their existing architecture. And instead of merging or purging all the identities both companies possess, an identity resolution solution can sift through all the information and keep it all in its native formats for future use.”

The always eloquent Mr. Jonas extends this argument further in his post, Entity Resolution Systems vs. Match Merge/Merge Purge/List De-duplication Systems.

If it’s not too late for AEGON, they should consider Mr. Jonas’ reasons below to avoid a “ground up reload”:

• Batch versus real-time

• Snapshot in time versus perpetually current

• Data survivorship versus full attribution

• Data drifting versus self-correcting

• Single version of truth versus every version of truth

• Outlier attribute suppression versus context accumulating

• Binary list processing versus “n” data source ingestion

• Limited scalability versus massive scalability

Please read the post in its entirety for a full explanation of these points and you’ll see why we agree with Mr. Jonas’ conclusion below.

“Entity resolution systems are best suited for real-time missions where processes require access to the most accurate and most current view of that which is knowable [to the enterprise].”

To conclude this post, we’d like to point out to insurance execs that there are other business applications to consider, as mentioned by Glenn in his post on the AEGON merger last month:

“With mergers and acquisitions, there’s always customer overlap issues in any industry. In the insurance industry, with some customers attempting to defraud insurers by using multiple identities, it’s critical to the bottom-line to cross-reference multiple identity records against not just watch lists but also the following business applications:

• Automated fraud detection

• Underwriting risk management

• Enterprise identity management

• Transactional CRM

• Compliance

• Background checks

• Producer risk assessment”

Share This

[Daily Post from Infoglide Software] Identity Resolution and Mergers and Acquisitions

“With mergers and acquisitions, there’s always customer overlap issues in any industry. In the insurance industry, with some customers attempting to defraud insurers by using multiple identities, it’s critical to the bottom-line to cross-reference multiple identity records against not just watch lists but also the following business applications…”

Christian Science Monitor: Airport screening raises privacy issue

“The TSA will also offer people an opportunity to provide their date of birth and gender, so that if their name is the same or very similar to one the watch list the TSA can easily, and permanently, distinguish between them. TSA officials contend that with just a name, 95 percent of the matches against the watch list will be accurate. With a name and date of birth, that percentage goes up to 98.5 percent. Add date of birth, and officials insist there will 99.5 percent accuracy.”

Government Computer News: DHS bares upgrades to immigration travel databases

“The Homeland Security Department has unveiled several important upgrades to databases that collectively contain tens of millions of personal records concerning immigration and travel. Some of the changes are intended to foster information sharing among organizations inside DHS as well as with outside government agencies. Others aim to reorganize the databases internally so as to make them easier to use. “

Retail Solutions Online: Grow Business - Reduce Shrinkage: The 2% Solution

“On a daily basis, retail loss prevention (LP) executives face the daunting task of diminishing corporate financial loss. Add protecting assets and increasing profitability to the mix, and LP executives have a lot on their plates. Still, one thing is on almost everyone’s mind: Shrink rates are on the rise, costing retailers upwards of 2 percent of their sales each year. It’s a constant struggle to detect and prevent inventory shrinkage, as well as internal losses like check fraud, cash theft, and Internet sales fraud.”

Washington Post: Former Guard Accused Of Hiding Muslim Ties

“Federal prosecutors now allege that Jackson intentionally withheld his Muslim name, Abdul-Jalil Mohammad, to conceal a connection to a controversial imam in Southeast Washington.”

Share This

It’s probably obvious to anyone interested in identity resolution that it has some important applications to both business and national security. From employee screening and loss prevention to IT security to Homeland Security, there are lots of reasons an organization would want to resolve identities over potentially complicated data sources. But what about all the recent talk about so-called “micro-targeting” or “behavioral marketing?”

At its best, it’s basically the collection of identity data for the purpose of highly-specialized messaging. At its worst, it’s - well - the collection of identity data for the purpose of highly-specialized messaging. Still, nobody really seems to be talking about the fact that without the right software in place, the data being collected is less likely to produce much of a return and more likely to create significant security risks.

Thus far this year, advertisers have dropped about $575 million into behavioral advertising, and that figure is expected to rise to $1 billion or so next year, according to information on Marketing Vox.

Some of that money is spent on contextual modeling like Yahoo’s SmartAds, which Yahoo’s senior vice president of display marketplaces, Todd Teresi, says will create “scaleable one-to-one marketing” (from the New York Times). In a post on his Web Analysis blog Anil Batra argues that “this will allow Yahoo to build [a] richer set of data” that will lead to higher conversion rates.

All of that seems great for increasing revenue, but it seems like identity resolution should be an integral part of this process as well. Without intelligent systems in place to create what Infoglide Chief Software Architect John Ripley recently called a “living context,” what you’ve got is a huge data store with tons of identity data but without any built-in accuracy or privacy protection, especially since the data’s purpose suggests a great deal of sharing.

Putting it another way, HP Labs researcher Marco Casassa Mont says that identity management and sharing, especially from “an Identity Provider to a Service Provider or between two Identity Providers,” has to have both data and protection policies in place. He says that those policies have to:

• Enable users to provide their (privacy) preferences in a more explicitly and fine grained way (e.g. in terms of consent, disclosure list, deletion, notification, etc.);
• Enable enterprise back-end Identity Management solutions to manage the association of preferences and (data handling) policies to data and keep them into account during data processing steps;
• Enable the exchange of data along with associated preferences/policies;
• Introduce accountability, tracing and auditing mechanisms.

Lots of people are saying that the identity data collection of behavioral marketing is going to produce incredible results for everyone from retail organizations to political campaigns (as an article on MSNBC.com points out). However, while there is a rush to get data collection and resolution solutions deployed, nobody seems to be asking if those solutions are working hard enough to protect privacy or even accurately navigate the vast stores of identity data being collected.

Share This

Missed some of the great posts here on Identity Resolution Daily? Well there’s no time like the middle of the week to catch up on some definitive reading on things like identity resolution, employee fraud, retail data and protecting your privacy. Here’s what you may have missed:

The many meanings of data in the retail industry

Well, OK this post isn’t about the meanings of all data in retail, but there are a few specific ways that data is being used that have special relevance to us. Specifically, an article by Cathy Langley, director of loss prevention analytics for Rite Aid, about data-mining and exception reporting got us thinking about how crucial data is in the retail industry and, tragically, about how not everyone with a database (or databases) makes the most of the data they have. more…

Employees are walking out the front door with retailers’ profits

Organized shoplifters and employee fraud - often go hand-in-hand since the former regularly depends on the latter. Combined, they cost the entire retail industry roughly $41.6 billion last year, according to a joint study by the National Retail Federation and the University of Florida. more…

Getting ducks into rows: the complicated world of data synchronization

Information, especially the kind of data gathered by governments and enterprise-sized corporations, is often stored in lots of different places and in a variety of different formats. While this information age is definitely creating more information, it’s not necessarily putting that information in close reach. more…

A message from Mike Shultz about identity resolution

I’m the CEO of a software company, but I was a citizen before I was a CEO and I’ll be a citizen after I’m a CEO. And the citizen that I am sleeps fine at night knowing what I know about how the government is balancing identity resolution and privacy. The people we work with there take that balance very seriously, and so do I. more…

Share This

You may have heard the recent news that an Albertsons group is canceling their customer-loyalty card system. As with similar CRM deployments, the cards were mainly a way to collect customer data. However, according to Bob Colgrove, Albertsons Southwest division president, their aim in canceling the cards “is to give all the consumers in the marketplace the same great deals regardless of having to have a loyalty card” (quoted on RetailNet).

Presenting a compelling case, eWeek’s Evan Schuman argues that “the program got pulled because grocery managers weren’t bothering to use the data.” Schuman says that while many retailers love the idea of CRM data collection, few do much with it after installation. Then, when the business doesn’t see much ROI, they conclude that the CRM solution doesn’t work, and they pull the plug without really trying to figure out more effective uses for that collected data.

Richard Boardman says on The CRM Consultant that organizations shouldn’t even deploy CRM if there is insufficient executive sponsorship, insufficient financial and manpower resources available to run a long term program or instability within the business. The third element seems pretty obvious, but the first two got us wondering about how and why a lack of resources and support spells doom for CRM and all the data attached to it.

In a typical CRM deployment like Albertsons’, the amount of data is huge, and that volume means that the amount of waste is potentially huge as well. The reason for that waste (especially in deployments that lack the proper infrastructure) is that a lot of businesses still don’t have the tools to pull useful information out of their data. Maybe the data is just sitting in silos. Maybe it’s in several databases and in varying formats. Many companies see the resolution of that data as an overwhelming task, but it doesn’t have to be.

One way to extract that data is by using identity resolution software. It can resolve data across remote, disparate databases and turn it into manageable, usable information. By giving a real-time view of that data, it provides a cost-effective way to gain identity knowledge from the data. With CRM in place, you’ve collected lots of useful data; you’ve just got to put it to use.

Share This

Here in Texas, insured drivers spend “nearly $900 million a year to protect themselves against those without coverage,” according to a Terrence Stutz of the Dallas Morning News. Texas is trying to sniff out those 3-million or so uninsured motorists by creating a massive database, but it’s been slow going.

The delay is primarily because the proposed database, with records on the 16 million motorists in Texas, has yet to be put into action. However, the reasons behind the slow deployment demonstrate some common database concerns: avoiding false-positives and protecting individuals’ privacy.

Some potential good news for the Texas insurance industry is that those are great concerns to have. Jerry Hagins, a spokesman for the Texas Department of Insurance, said recently that “We don’t want people flagged for not having insurance when in fact they do.”

Those “flags” leads to false tickets or even arrests, and that leads to big lawsuits once the error comes into the light. False positives become an especially large concern since counterfeit insurance cards often make in-the-field verification tricky. One major aim of the database would be removing those fraudulent cards from play.

As we wrote about yesterday, one of the biggest factors in a database like this is the sheer volume of individual entries. However, one of the things that makes the data in a database like this one even more slippery is what Jeff Jonas calls “semantic reconciliation.” He explains that semantic reconciliation, the ability to determine that two entities are the same even though they are “described” differently, has to have perfect deployment even “before other analytical processes (e.g., statistical analysis, market segmentation, link analysis, etc.)”

The Texas insurance database (and similar ones in other states) has to be able perform identity resolution to determine - well - who is who. If it can, counterfeit insurance cards are less effective, there’s a vastly decreased risk of false-positives and the database will meet the “goal of at least 95 percent accuracy” that Hagins calls for. Based on the problems the media reports about the database, it seems that kind of data reconciliation is a long way from being ready.

Share This

It’s no secret that the information age is shrinking borders and expanding corporate reach. The means of acquiring that information is becoming simpler, and businesses and laypeople can access vast stores of information in seconds (or faster). Well… sometimes.

The fact of the matter is that information, especially the kind of data gathered by governments and enterprise-sized corporations, is often stored in lots of different places and in a variety of different formats. So maybe a more accurate statement would be that while the information age is definitely creating more information, it’s not necessarily putting that information in close reach.

Jeff Jonas recently argued on his blog that people are becoming “trapped by their data trail,” referring to the ubiquity of “historical data” and people’s increased access to it. However, he also points out that navigating the vast body of that data means figuring out ways to resolve the contradictions we mentioned above: lots of data, variety of places/formats. Jonas says that “we are really calling for custom crafted lenses to intentionally narrow our perceptions.”

The precision demanded by those kinds of “lenses” is incredible, especially in light of the sheer volume of digitally stored information out there. Bill Bittner, President of BWH Consulting, wonders if the synchronization of data, even through specialized lenses, is even possible. In a recent post on RetailWire, he argues that

“Data synchronization is built on the assumption that there exists one version of the truth, but the fact is that the supply chain is never standing still. The parts are always moving and variations between units will always exist.”

He’s talking predominately about retail data, but the argument can really be applied to any diverse collection of information.

One solution is the creation of a single data source, as Bittner says HP did in reducing “750 internal data marts into one data warehouse.” However, those kinds of data warehouses address requirements and support uses that are sometimes different from the needs for fraud detection, for example. Collapsing data into warehouses through ETL or similar means can risk losing lots of information that’s useful in fraud detection. Centralizing data in that way certainly streamlines the storage of that data, but it doesn’t always guarantee its completeness for all applications.

Perhaps a middle ground can exist for identity resolution applications. Instead of an overly-specialized “lens” through which very specific data is retrieved or the expensive and complicated endeavor of creating a centralized data warehouse, disparate data should be resolved into a unified collection. As the sources of information grow in size and complexity, that kind of solution will likely end up being a very important option for identity resolution solutions.

Share This