Start Scheming

Just a few weeks ago, Google, Microsoft and Yahoo announced that their search engines would jointly support a new semantic markup language convention called microdata, the details of which can be found at a site called schema.org.

Semantic coding? Microdata? Okay, I promise not to go too deep into the details, but this is important stuff. In a nutshell, this new schema -- which will be recognized by all major search engines -- allows websites to present their information to the search engines in a more structured fashion. The schema supports standardized HTML tags. By tagging data at a very granular level, search engines not only get smarter about the data they are indexing, they can manipulate and process it just as if it was information in a conventional database.

Yes, by encouraging website owners to add these standardized tags to their content, the search engines want to marry the power and precision of structured data with the richness and depth of textual and graphical information.

As a data publisher, should you be concerned? Well, let's look at just a few of the specific business types already defined by the schema: travel agency, child care, financial services, real estate agent, shopping center. If enough of these types of companies ultimately add these tags to their websites, the search engines will be able to pull them out with complete precision.

Not precise enough? Doesn't impact your industry? Well consider too that the schema allows for tagging of postal address and geo-coordinates. That means that over time the search engines will with precision be able to list (and even sort) all businesses at any location or zip code. Precise mapping will also be a breeze.

And let's not forget products. The schema provides tags to identify specific products, and even if a specific item is being offered for sale or not. There is a way to tag associated product reviews, and even to identify a product ID code -- either a proprietary code, or an industry standard code system.

Most significant of all: this schema will allow search engines to do easily what they've never been able to do before: parametric searches. Want all self-storage centers in Fairfield, Connecticut? Done. All self-storage centers within 50 miles of Wichita, Kansas that have units for rent? Done. Yes, it starts to hit a little close to home, doesn't it?

Will this schema take off? It depends on wide-scale adoption, of course, but it's free to use and there's the huge carrot of potentially improved search engine results rankings. The schema is not sufficiently robust as of now to scare too many data publishers, and it does feel heavily weighted to local, retail businesses, but by improving the precision of search engine results, it reinforces the thinking that the big search engines are "good enough."

Can you use these tags to improve the results of your own web scraping efforts? Absolutely. Like almost everything with the search engines, this one is a two-edged sword.

Inferred Opportunity

The FICO score -- that numeric summation of our credit histories -- continues to find new and novel applications. Most recently, the healthcare industry is using it as a predictor of whether or not patients will take their prescriptions. Whether or not you will take your pills twice a day as instructed, called both adherence and compliance in the world of healthcare, is actually a big issue. (Check out how 2010 Model of Excellence Nominee HealthPrize is making a business out of this). If you don't take your medicine as instructed, you ultimately end up back at the doctor or in the hospital, pushing up healthcare costs for everyone. There are lots of people who for one reason or another don't take their medicine, and with doctors spending ever-less time with patients, it's harder and harder to identify exactly who would most benefit from some follow-up calls and encouragement. Enter the FICO Medical Adherence Score.

This new product is interesting in its own right, but it comes on top of numerous other applications for FICO scores and credit data. Many employers now inspect the credit of potential hires, as sort of a proxy for reliability and character. FICO scores are used by auto insurance companies to set rates (apparently, the higher your credit score, the lower your propensity to sue -- who knew?).

There are many, many examples of this type of data application -- we at Infocommerce Group call it inferential data -- where data point X can be used to infer something totally unrelated. Almost all of this is occurring with consumer data, not business data.

Yes, there is some inferential data in the B2B world. Sometimes called "trigger events," some data publishers understand, for example, that a company that just moved to a new office is an excellent candidate for office supplies. But B2B inferential data is still in its early stages. And that got me musing.

Inferential data is all about trying to discern something you don't know from something you do know. Companies with multi-state operations need more sophisticated accounting, HR and payroll support services. Companies that lease office space disproportionate to their number of employees are likely planning for near-term growth. Companies operating from residential addresses tend to be smaller and have a distinct set of business needs. Companies that get taken to court a lot in contract disputes are probably not great prospects.

It's likely even possible to assign companies to behavioral buckets. Are companies in high-rent districts free-spenders? Should companies with press releases touting many joint venture and partnership deals be sold differently from other companies? Are companies that constantly recruit for the same job titles revolving doors? Do family-owned companies (often spotted by the same surname in multiple executive positions) do business differently, and do they need a distinctive sales pitch?

If a lot of this sounds like business intelligence, it is. But there are types of business intelligence that can be divined on an automated basis by data publishers if they assemble the right data, and understand its inferential value. And by identifying and packaging some of these inferential indicators for our customers, we simultaneously build revenue, value and customer annuity. We are at the earliest stages in the development of powerful and useful B2B inferential data. Now's the time to start staking your claim.

Judgment Day

In a 6-3 decision yesterday, the U.S. Supreme Court held in a case called Sorrell v. IMS Health that states may not prohibit the use of physician prescribing data for marketing purposes. The decision in this closely-watched case is seem as a strong win for all data publishers.

What's the background? Several states, including Vermont, passed laws prohibiting the use of physician prescribing data -- information on how many prescriptions doctors write for specific drugs -- for marketing purposes. That was a blow to the pharmaceutical industry, where this data is relied on heavily for deploying their field salespeople, called "detailers." It was an even bigger blow to data publishers such as IMS Health, who make a lot of money aggregating this data and selling it to the pharmaceutical industry, among others. No surprise: IMS Health sued, and with different courts deciding different ways, this case quickly made it to the U.S. Supreme Court

What was decided? Well, as noted, the Court came down in favor of IMS Health. What makes the decision interesting is that the Court saw this as a First Amendment, free speech issue. If I am boiling it down correctly, the Court said that IMS Health got the data legally, and the data are factual and accurate, so therefore, IMS Health can sell the data any way it wants, and government can't restrict how the data are used.

By the way -- both the Court and the lawyers involved refer to IMS Health and other data publishers as "data miners," but rest assured this case has nothing to do with web scraping. Further, this case has nothing to do with information relating to individuals; all the prescription data was "de-identified" as they like to say in healthcare, so it was all anonymous in terms of who the prescriptions were for. This is a case about purely commercial data.

So let's play lawyer -- What are the implications of this case? I'm still not clear how much the circumstances of the case influenced the Court's decision. That's because the Vermont legislation in particular was passed for the express purpose of making it harder for pharmaceutical companies to sell brand-name drugs to doctors, which pushes up healthcare costs. That's where a lot of the free speech issues come into play.

But the sense I get in the commentary I have seen so far from real lawyers is that this case is significant, and that it will be much harder for states to place restrictions on the sale and use of data for advertising and marketing purposes. Databases are now considered "speech" giving them strong First Amendment protections, and the rights of commercial speech generally have been further strengthened. And all this is generally good for data publishers in terms of both accessing and distributing data.

Rating the Better Business Bureau

It's no secret that online reviews have changed the way we shop and interact with businesses large and small. Despite this, the granddaddy of business ratings, the Better Business Bureau, even with its incredibly powerful brand, continues to drift around, seemingly unable to capitalize on this trend.

I've read article after article recently in the consumer advocacy column in the New York Times taking the organization to task for policies that appear to help errant members look better than they really are. The newsmagazine 20/20 raked the BBB over the coalsin a blistering expose late last year. Many consumer complaint websites carry numerous complaints about BBB. What went wrong?

A lot of it seems to stem from its business model. First, BBB is much closer to a chamber of commerce than a consumer advocacy organization, but by accident or design, BBB has created a consumer advocacy aura around itself. That leaves lots of consumers underwhelmed when they reach out to BBB for help.

The second issue is that BBB is supported by its business members, creating an inherent conflict much like we have seen with the big bond rating agencies. BBB thrives when its members are happy, so it has strong incentive to give its business members the benefit of the doubt when complaints arise.

Finally, the BBB rating system is poorly understood. New members of BBB start out with a high rating, which may be reduced over time if many complaints are left unresolved. The problem with this approach is than an unscrupulous business can pay the modest membership fee, secure a great rating, and have an extended window to plunder and pillage before its rating gets reduced.

The BBB response? Well, quite recently, it announced that it was going to offer detailed profiles of all its members -- much more than the Spartan contact information it had provided before. It is also going to allow consumers to directly post reviews against these merchant listings, a radical departure from past practice and potentially unsettling to its membership.

Will this new approach work? My sense is that this move may be too little, too late, and I still think BBB policies may be too opaque for the consumer-driven world we now live in. At the same time, I've learned never to under-estimate the power of a strong brand, and BBB is one of the strongest. Here's hoping this new positioning will move BBB back to the head of the pack. After all, consumer reviews, third-party ratings and an integrated dispute resolution capability is a powerful set of features.

Car Talk

It started out with an intriguing headline: Toyota, Salesforce.com Build Social Network. I admit, I was intrigued and I tore into the article only to come away more than a little confused.

Apparently, Toyota wants cars to communicate with their owners. To make this happen, Toyota has partnered with Salesforce to use its private messaging network, which is described as “like Twitter and Facebook,” except that it’s not, which is a really strange choice for a pure consumer application. Leaving that oddity aside, the basic idea is interesting: Toyota wants your car to tell you, with messages to your mobile devices, for example that the battery is low. You would also get scheduled maintenance notifications and the like. It’s a clever application, arguably a personal workflow application, and it makes good sense.

That is, it made sense until the Toyota social marketers got involved. While this new feature doesn’t actually utilize Facebook or Twitter, Toyota marketers saw the value of you being able to forward these messages to Facebook and Twitter. Yes, now you can easily tell all your friends and associates that your car battery is low. Huh? And better yet, your new best friend your car will now be able to send you helpful “product and service information” (sounds like advertising to me). Yes, a good and useful idea gets bizarrely contorted just so marketers can slap a “social media” label on it. And your new car will be sending you a stream of sales messages along with critical safety information.

Object lesson for all of us: there’s a role for social media and it’s undeniably powerful. But to contort good products to force in social media components is a great way for Toyota to get de-friended … fast.

This Toyota announcement comes at the same time Nissan is promoting its new "NV" line of work vans that are being positioned as mobile office environments. High-tech features include such things a mounting rack so you can install a filing cabinet, and a magnetic bulletin board. Perhaps it is Nissan and Salesforce who should be talking!