Evan Schuman's StorefrontBacktalk

Guest Columnist David Taylor is president of the PCI Alliance and a former security analyst with Gartner. Today’s column is Part Two of his series on Tokenization. Last week’s column talked about suggestions for saving money and reducing risk by replacing cardholder data with proxies.

Nearly one-third of merchants dismiss the concept of tokenization because they consider their environment to be too complex, with card data in so many places.

But a related reason for the quick dismissal is they see their card data as very "dynamic" – being used by many different applications and service providers. These merchants said that, at best, tokenization is a long term approach for them.

When I point out that they are likely spending much more to implement and maintain multiple encryption and incompatible key management systems, most note that these systems are now considered "sunk costs" by management. These security managers would clearly like to remove the risk from their environment, but they see the decision as out of their hands.

Slightly more than one-third are seriously considering substituting token or proxy numbers for card data and other types of data as well. When I recently spoke with a security executive who manages the infrastructure for a large public university, the executive said that his employer is looking at tokenization more for Social Security numbers than for credit card numbers.

His point was that they had passed their assessment by using a combination of data encryption and compensating controls, and felt this was good for another 2-3 years. But, he said, they actually had many more Social Security numbers in their system from current and former students, and that these numbers were in systems not currently encrypted and with comparatively weaker access controls, vis-à-vis their credit card data.

This exec said that they see tokenization as the fastest way to substantially reduce their risk to this data. He also said that this data is "just sitting there" – again raising the issue of the "flow" of data and the interdependencies among systems as a gating factor tokenization of card data.

The final third of the folks I’ve talked with were somewhat familiar with tokenization and were aware of at least two of the three vendors I mentioned last time (Shift4, MerchantLink and EPX). For these merchants, which included companies in the hospitality industry, the specialty retail sector, and a couple of mid- level general merchandise retailers, they see their major vulnerabilities as being "from the POS inward."

But two of these merchants are looking to switch out their POS systems over the next two years and consider this the ideal time to implement tokenization. The primary driver is cost reduction, with risk reduction coming in a close second.