Preview: Emergent Data

[The is a pre-publication draft.  You can download a PDF of the entire essay here.  Please join the discussion by clicking on the "comments" button below.  I'm looking forward to your criticisms, witticisms, and queries.  As a small token of appreciation, I'll make sure to send a complimentary copy of the final essay to all respondents.]

Killer Platforms | Emergent Data
The Spark that Ignites Information Advantage

by Chunka Mui

"Too Much Information"

A few years ago, I asked Mike McGavick, the chief executive officer of Safeco Insurance Cos., to name the most disruptive development on his industry’s horizon.  McGavick, known in the industry for his thoughtfulness, answered: “too much information.” 

McGavick was not referring to the perennial problem of analyzing huge quantities of information in his data warehouses. Instead, he was worried about the strategic opportunities for and the potential challenges of handling much better and perhaps perfect information. McGavick was anticipating what I describe as emergent data, classes of information that were previously impossible, or at least impractical, to gather but now are made feasible by advances in information technology.

In a range of industries, including insurance, financial services, healthcare, and complex consumer and industrial products, emergent data is changing the strategic context in which companies manage their organizations, products, customers, and markets. In the right hands, new information about product and process conditions, customer preferences, and other environmental factors are sparking numerous innovations. But, because these data types are just appearing, their strategic significance is often not well understood, so it is usually ignored, and rarely leveraged.

This article explores different kinds of emergent data, the competitive opportunities they are yielding, and the challenges of taking advantage of such information. The most valuable uses of emergent data are often unforeseen secondary effects, that is, applications that lay outside the original purpose for generating or collecting the data. If companies design their business platforms to integrate streams of emergent data and make that data available across the enterprise, they can generate greater value. And the most effective platforms, the killer platforms, will be those that are designed to facilitate a series of innovations rather than isolated applications.

To read the entire essay ...

Download the PDF file.  EmergentData.pdf

Here is an outline of the paper:

1. “Too Much Information”
2. Collisions in Insurance and Beyond
3. Emergent Data and Business Innovation
4. Classes of Emergent Data
   1.     Identity
   2.     Location
   3.     Health and Diagnostics
   4.     Preference
   5.     Operations
5. Emergent Data and Information Advantage
6. Platform Plays
7. Guidelines for Getting Started
   1.     Finding Emergent Data
   2.     Levels of Learning
   3.     Organizational Context
8. Conclusion

To read the entire essay ...

Download the PDF file.  EmergentData.pdf

Comments

[brian saperstein]

Chunka,

I was made aware of your essay via Chris Curran and must say it is very well thought-out. Two areas that leave me curious:

- Privacy. Another post in the comments mentions privacy, but I am curious about the ramifications of "getting closer to your customer." As people get more and more concerned about monitoring, invasive data mining, and personal financial security, I would be curious to know if thought leaders see a tipping point in regulations on emergent data. Would a mobile customer really want the cellular provider to track their exact location, even if it meant precise GPS directions or faster 911 response? What will be governed by markets and what will be governed by legislature? This is especially poignant as identity theft increases and we celebrate a decade of HIPAA, which I have seen clients continue to wrestle with in their day-to-day operations as recently as last year. The issue of what is private, what will be regulated, and what data can be leveraged to continually get closer to the customer will loom large.

- Contracts / Diversification. Aside from issues of privacy, the operational implementation of emergent data will most likely go far beyond traditional firm boundaries. The example you used about OnStar entering alliances is valid, but the value in this example is driven by OnStar's information. The next generation of emergent data uses might very well be in combining customer attributes (identity, location, or any other type) with another firm's information to offer products and services which are increasingly customized and specific. The history of business has shown that many firms, rationally and often irrationally, think they need to diversify into new lines of business to leverage potential synergies instead of creating contracts to capture value. I would be curious about, in a manner similar to privacy, who owns these emergent data elements, where they reside, and, ultimately, how companies will work together to leverage these opportunities.

Again, excellent essay.

Brian Saperstein

[Amaresh Tripathy]

Chunka:

Very interesting thoughts. One point that I will like the paper to address :

1. Where in the organization should the responsibility to seek out emergent data and work with them to create experiments lie? For example, we worked with a client to design a statistically valid business process, very similar to the Hartford example you mention, where inbound customer care calls are mined and information is gleaned on various dimensions for both operational and strategic use. Our vision was to use the process as a platform for competitive advantage, However, because we dealt solely with the customer service organization with a very narrow focus on meeting the call metrics, our process is used today to do operational reporting and is not utilized to its full potential. How does one avoid such pitfalls?

Hope this helps

Amaresh

[Jim Taylor]

chunka...i read with interest your essay....i would offer you the following comment...if you go all the way back to shannon-weaver and the early information theory guys, i think they can offer a perspective that might make your distinction between "emerging data" and information a bit more understandable...at times you treat data as information and at others you split hairs on differences between the two....two key issues transform data into information...the first if a point of curiosity -- a question...data remains latent, inert until it is programmatically dedicated the constructing some kind of answer...without strategic curiosity (e.g....where is my package) data will lie on the floor and do nothing...data becomes information (shannon-weaver) when it alters the distribution of possible outcomes enumerated by a question (or more precisely, an uncertainty)...for example...if i am in new york, and you ask where's jim?...if the answer is at a ball game...the range of possible answers runs from a little league game in queens to the mets and yankees...if you know i have no children the probability matrix is accordingly reduced...if you now i hate softball, etc...the matrix  eventually reduces itself to yankees-mets...if you know (learn!) that there is no mets game that night, you have reduced your search to perfect certainty...1 location for 1 person...a major strategic choice businesses make is how certain do you have to be before you make a decisive move...sometimes it is better to search both ballparks rather than run the expense of additional data collection and management (the issue of defined levels of equifinality -- a major decision issue in reinsurance, terrorism and distribution) but the key to the whole game is not data...its sustained, strategic curiosity...i hope these remarks are of help...jim taylor

[Mike Shinnick]

Chunka-

Thanks for sharing with me through Chris Curran. Interesting topic and one that I think about daily as a portfolio manager. A potential thesis stated on page 5 resonates true --- ''emergent data has long been a driver of business innovation.'' As you say in this statement and in the sea clock example, this is not new idea, rather it is a centuries old concept and emergent data has been changing the strategic context of business throughout history.

What is new, it seems, is you are implicitly getting at the pace and intensity issue with advances in information technology making so much more emergent data possible(thru better, faster and cheaper collection, storage, modelling and analysis capabilities). The danger is that a company focuses too much on the possible (as much data as we can get), without sharp focus on the true sources of competive advantage; e.g. Progressive's great business. For these reasons, you are suggesting that business needs a nomenclature and a process around emergent data because of the increasingly overwhelming permuations of the possible. I would say this more directly up front if indeed, this is the thrust of the case you are laying out. Then, illustrate with stories and anecdotes or cases and then, move to guidelines and conclusion.

I share Vince's thought that OnStar receives far too much ink in this piece. Mention, yes. Half to 1/3 the case pieces, no. I would in particular encourage you to beef up the platform section with other examples. EBAY, Instinet (ECNs)and RSS(Real Simple Syndication) and publishing come immediately to mind. I would also call your attention to "Our Brave New World", a new book by Charles Gave et al which deals extensively with new business models of platform companies. I would suggest that the investment industry has many emergent data case examples; e.g. Bloomberg is a multi-billionaire because of a business sharply focused on emergent data and information about investment securities. The Bloomberg network and process distills the oceans of information re: investment securities into actionable and highly accurate screens and tools for which people will pay $1400/seat per month. Much of this data is publicly available for free or at a much lower cost. Do people pay the Bloomberg premium because they can hit the HELP key twice on a Bloomberg system and a competent person calls you at your desk within seconds? How much are they paying for the network, the organization of the data and the DISTILLATION of the most IMPORTANT and TIMELY information.

The Organizational context discussion re: valuation and ROI vs. OCI puts too much emphasis on the OCI and not enough on the ROI, in my opinion. This sections ends with a portfolio of options discussion which cannot be an end unto itself. Hypothecate and experiment with OCI portfolio, sure. Making major investments in new programs to exploit emergent data takes ROI analysis. Discuss how companies make this transition and move from experiment to business model change. Acknowledge explicitly that what matters is what will make $$ and be a source of differentiation and advantage. We know the Xerox PARC story, good experiments and experimenters are not necessarily good for shareholders. My guess is that GM shareholders feel similarly re: Onstar's contribution to their share price. OnStar is not a new offering and has been talked about for years. What long run advantage are you espousing for GM? All the while GM has OnStar, Toyota is taking share and making lots of money without developing OnStar which begs the question related to your My ProductAdvisor preferences feedback example -- so what? Why are Toyota's products selling better and does OnStar matter that much to customer?

The conclusion lacks clarity. Yes, there is no universal approach or dogma. Yes, it is important and urgent.(assume you make this case to the reader) What do you advocate to DO and why? It would seem that the implications for certain types of businesses (industry sector groups) could be asserted here as starting points for discussion and then as much specific language re: a proposed path forward would be helpful.

I hope this is helpful and look forward to reading future iterations. All the best.

/mike

[Jeff Wood]

Chunka,
I received an invitation through Steve Ritchie to review your essay. The feedback that I want to add that I thought may have been missing is perhaps an extension of Steve’s points. Although emergent data has the potential to create significant new value, not all of that potential will be realized because much of the value will be shared with the customer or subjectively you might even say, wasted by the customer.

Whereas Steve’s comments focused mostly on the regulatory side, I believe the free market will also play a role in how much value is realized through emergent data. As an analogue, I would point to permissions-based marketing where consumers are rewarded for accepting marketing pitches. One “permissions broker” that I am aware of in this field is MyPoints.com, which enables consumers willing to accept emails from marketers to receive points which can be redeemed for gift certificates and products. Frankly, it could be considered merely an evolution of the broadcast advertising model where there is also a value exchange.

As consumers become more savvy, as your citation of the no-call list participation suggests, they will likely become more guarded with their data and require value in return for forfeiting their privacy. Obviously, it will follow classic economics where many will sell their information for de minimus compensation while others will require substantial compensation to forfeit their privacy. Personally I would be happy to share the aggregation of my various accounts and transactions in my Microsoft Money file with marketers, for the right price!

An inverse example if you will that I am just slightly familiar with is the recent rise of the “no documentation (no doc)” mortgage where consumers are apparently willing to pay a higher rate to not provide the copious amounts of personal financial data traditionally required for a mortgage application. Clearly there must be an adverse selection issue here given that generally those with sterling credits are not likely to apply for such a mortgage but I suspect there are some who simply do not want to provide such intimate personal finance data and are willing to pay the price for that protection (or laziness).

So in conclusion, I would simply add that emergent data’s potential for value creation is huge, but only to the extent that the data can be obtained AND used cost-effectively. To end with a final example, I would cite Intel’s aborted attempt at embedding a unique identifier into every microchip it produced. In the end, it was soundly rejected by consumers because there was no value to them, only costs in that exchange.

As an aside, I wanted to make you aware of Online Insight, where I worked several years ago. I believe they provide a product that is at least functionally superior to the MyProductAdvisor.com approach. They actually use conjoint analysis (i.e., product feature trade-offs) to provide product recommendations to consumers. They clearly have a value exchange – better data for marketers and better product recommendations for consumers. In a nutshell, where MyProductAdvisor.com asks a consumer about what is important to them they may not actually know consciously. Furthermore, MyProductAdvisor.com assumes there are no trade-offs or concessions in product selection, which clearly does not mirror the reality of economics. Full disclosure: although the company still exists, it has not flourished as my worthless options indicate!

[Steve Ritchie]

Chunka, you provided me a welcome break from a pile of performance reviews, so I owe you a debt of gratitude. As someone in the middle of the information services industry (ChoicePoint), I may have a different perspective than others, but I thought I would add my thoughts to the mix. The order is a bit random, but here it goes:

A concern I have with your thesis is the implication that individuals would consent to have information about themselves collected and aggregated. To date, I think a lot of companies have collected a lot of information without the knowledge of consumers, and as consumers learn about it, they can get upset. My company has certainly seen the effects of this. Remember, as soon as the FTC/FCC empowered people to join the do not call list, over 50% signed up within a year.... The Federal Reserve noted that subprime lending and online credit scores have revolutionized the mortgage industry, provided capital to millions who were redlined in the past, and provided homeownership to many families, but the focus of many states has been on 'predatory lending' and not on the value of the data being opened up to the market.... etc etc

Following your Safeco case study, it would be interesting to see how the growth of centralized shared claims data (ChoicePoint's CLUE product) has lowered rates (at least rates controlled for higher car values, etc.) Instead of predicting future claims behavior, companies now share their claims data with each other, so they KNOW the claims behavior of an insured and can, in theory, set better rates. The theory makes sense, but has it really helped reduce rates?

Keeping with that theme, insurance statistics show that credit data is a good predictor of auto insurance risk, but its use has triggered an avalanche of complaints, new state regs (it cannot be used in MD) and other concerns about invasions of privacy. Society has not yet decided the tradeoff between the value of perfect information and the value of personal privacy, in this case at least.

Final thought -- more emergent data can create a TON of value, but it can also be abused. With the new data must come clear consensus on societal norms and ethics around its use. For example, today, you can tap a database to screen the background of any prospective employee for felony convictions for as many as twenty years back. Clearly, the ability to protect people from potentially-violent or theiving employees has value (think nannies, camp counselors, bank tellers, etc.), but I just read a story about a college graduate in the area who is destitute and unemployable because of a 14-year-old conviction for a crime he committed when 17 that shows up on background checks and has been clean ever since). Is that a good use of data? Hard to say.... Also, too much data made too accessible can lead to other abuses... I can easily find your home address and telephone number online (even if unlisted) and while I would do nothing with it -- what if I meant you harm? 99% of the time the address data would be used for good, but there are cases of ubiquitous data being used to harm the indvidual whose data is aggregated.

I am not trying to be critical, because I agree with your premise that emerging data will open up tremendous opportunities, will create wealth, and solve a lot of problems. But like any other valuable commodity, society must establish the rules of use before the commodity can benefit far more than it harms.

Good luck with the essays.. I really enjoyed the thought-provoking read!

[Rick Wilson]

Chunka-

Very interesting area of thought; my sense is you are hitting a contemporary issue in a potentially provocative way. Some feedback and reactions for you:

1. Your email referenced the following: "customers are usually as technologically sophisticated as sellers, I am also looking at the changing context of customer relationships and expectations". This is an area that is rich for new IC - others are thinking about it in the context of "customer co-creation" and how to deal with greater levels of customer knowledge, involvement, and influence. I was hoping you'd tie back to this more explicitly to your ideas about emergent data - especially given your Drucker references. But simply "having data on your customers (e.g., location, etc) is not having "insight". I believe Drucker was talking about this distinction.

2. The "strategic unrest" idea was intriguing but didn't seem to endure through the article which I would have liked.

3. The metered insurance "build up (data, graphs, etc) is unnecessary as the idea and concept is easy to grasp. You could reduce the entire two pages to about two or three sentences and not lose any of its power. I also thought the "adverse selection" and re-regulation track was distracting and not central to your main point of discussion.

4. The RFID example didn't feel as powerful as its level of reference in the article. Some might read that it's an "improvement on UPC" - and simply solves limitations of UPC, as opposed to being a powerful and discontinuous technology/data breakthrough. You may want to find a different illustration given the controversial views of RFID adoption, or lack thereof (it seems to be getting the eye roll from a lot of people as "convergence" ultimately did after 10 years of hype in telecom). I certainly appreciate the potential of RFID, but my fear is that it has the potential to cause your readers to discount your primary points just because it's a bit of a cliche/over-buzz area right now. The big "customer" issue (retailer as technology customer) is about out-of-stocks, which they still seem to be struggling with mightily.

5. Also, per RFID you use the language "...hoping to encourage the broad realization of benefits from RFID, both Wal-Mart.....are using their buying power to encourage adoption". This feels dangerously close to a "technology looking for a buyer" which most writers are quick to criticize. The evangelical nature of your RFID adoption prose may be at odds with the main theme of "emergent data".

6. I very much like the "myproductadvisor.com" ideas. I think that whole concept could be developed further - it seems (to me) to better hit the target of "emergent data" that I thought you were addressing. How can a company capture and leverage such data without a third party? Does a third party source of insights matter? How do you ensure competitive advantage with this type of data if it's third party? etc...

7. Be careful at bottom of page 11 when you write off traditional market research. People who are motivated to go to a .com site for info may be a subset of the overall population. 80% of the "myproductadvisor.com" buyers moving away from SUVs may or may not reflect the overall population. As a "lead indicator" I agree with your point, but it would be naive to suggest that the rigor of research science is no longer necessary with an online data source....it feels a little like the old "rigor of financial soundness and profit are no longer necessary with dot-com start ups"....

8. On page 13 you start a paragraph: "Better operational data is needed...". It feels like you're moving into some type of consulting or advisory mode as opposed to a reflection of what's going on with emergent data. Again, it may come off as evangelical about "the need for" as opposed to "look at what's happening already with emergent data". Not suggesting either is better, but they seem like two different articles and you may want to keep the theme consistent.

9. The CEMS case seems to be less about "emergent data" than it is about "smart analysis"....I can't confess to have read this five times, but my first reaction was that I was taking another side road from the main theme.

10. I liked the "on-star as technology integration play" idea. How does tech integration fit with emergent data?

11. I also liked the "information (ie: competitive?) advantage" vs. "Emergent data" distinction. Could you develop this further? Getting from the latter to the former is ultimately the quest is it not?

12. I found your page 22 attach on academics and dogma unnecessary and distracting. Personal preference.

Neat stuff Chunka. I'd be glad to look again if you want.

Cheers
Rick
--
Richard E. Wilson
Managing Director
Chicago Strategy Associates, Inc.

[Vince Barabba]

Chunka --

Your essay is very well written. The only thing relative to the Market Insight System portion is that in addition to fitting very well into your definition of emergent data it also offers the ability to open up the mind and to look at things differently because of the potential of seeing things not easily available in the traditional methods. For example since all the products in a category are eventually evaluated by customers, consumer preferences for all products are available to the analyst instead of just those he or his superiors felt needed evaluation. This allows the enterprise to profile the users of all products in their category and see which products had consistent or different user profiles...as well as those whose user profiles were unique to the profiles of all other products.

The element of always being on also opens up new opportunities for dealing with the unexpected. When an unexpected event occurs, instead of -- after the fact -- trying to determine how people felt both before and after the event, the firm can check market preferences of its daily contacts before and after the event to determine the events impact...no matter when an event occurs.

Relative to OnStar you might want to consider referencing Clay Christiansen's two Harvard Case Studies. They were very well done.

[Alan C.]

A few short notes:

You may want to include some thoughts on data privacy laws / restrictions (especially the tight ones in europe) and consumer backlash over some of this unanticipated tracking (see http://www.nocards.org for a good example) as often, consumers aren't explicitly aware that they are being "watched" and some restrictions do exist today on this (rendering some of these applications infeasible overseas).

Although the On-Star example clearly demonstrates some of the option value to this level of data evaluation, it seems to take up about 1/2 of the article. You may want to chop down some of the detail on this if you need to tighten things up.

I spent about 3 years working for a retailer doing all sorts of unstructured data mining on their loyalty card data. The CEO had a "build it and they will come" attitude with the data warehouse, that only marginally panned out despite significant investment in the technology, people, and lots of input from the business. After 3 years, we definitely had identified and implemented a few neat marketing applications (e.g. topographical maps showing competitive market share by zip code, targeted weekly circular by email programs, monthly targeted snail mail programs, advanced statistical attrition models that could identify products that customers leaving the store prefer, customer level elasticity measurements, market basket adjacencies for cross-selling opportunities) and explored some of the information asymmetry opportunities (health insurance based on spending habits, programs based on profitability), but in most cases, even with this super-detailed data, the signal to noise ratio was way too high to identify meaningful patterns, and without adequate statistician supervision, too many patterns that really didn't exist were "discovered".

About 2 years after I left, the analysis group was disbanded due to a lack of cleanly measurable patterns, and actionable ideas coming out of the data.

[Alan Matsumura]

Hi Chunka,

I think that this is a pretty important and very relevant topic to explore. It is something that we face a lot in our practice so I do have some positive and negative reactions. (No negative reactions to your thoughts - negative reactions because I have some bad experiences.)

1.  First just a general reaction. I think that in many cases our ability to gather data has exceeded our ability to use it. There are two paths that this can take:

        A. Process flexibility or execution time is wildly out-of-synch with the data that is being captured. For example, one could weigh oneself every 5 minutes when dieting, but it would at best be wasteful and at worst discouraging and destructive. There should be some structure/framework for insuring that the relationship between the processes that you are monitoring and the facts that you are gathering are aligned. In a world where you get deliveries once a day and ship once a day, knowing real-time information that a package that you are shipping that afternoon is laying on the shipping dock is not very powerful or interesting. These disparities in process cycle time and refresh rate might lead you to rethink your logistics or it might just consume a lot of disk space and CPU. We have reached the point where we can drown in data without any capacity to be informed by the new facts (I think that information capacity is interesting). I think that better, quicker and more precise/accurate are often considered synonymous when thinking of data, but these attributes are only useful in an APPROPRIATE amount.

        B. Heisenberg's uncertainty principle. I have been using this in several casual discussions lately because I had to help my high school son with his Physics. As often happens, I finally understood what the high school text was trying to explain, namely that with complex interactions, there are relationships between what is knowable. Knowing the speed of the atom with great precision means that I MUST know less about it's exact location. Knowing it's exact location means that I MUST compromise the precision with which I know its velocity. I can know exactly how many chips are on the casino floor but can I know how many are being bet at a moment in time? At that moment, some bets are being won and lost. How do I count those? Are they the casino's or the gambler's? Once you dig under the covers, and look at the macro and micro situation, you can find surprising dynamics. In the end, you need to have some sense for what you need to know and why you need to know it so that you can instrument the right way.

Perhaps this could influence your section on too much data or help provide some structure to your classes of emergent data. I think that the way to develop a path towards exploiting emergent data is to make sure that there is some alignment with the new processes that you might influence and the data that you intend to capture.

2. For your location and your insurance examples, I have recently had some interesting discussions that say that knowing where something is made much more powerful by knowing what it's near. At one farm equipment manufacturer they said that knowing that a device is in a location in a field is only interesting. Knowing that the device is adjoined to a reaper as opposed to a fertilizing device can tell you about the type of operation and the processes that you might want to optimize. Knowing the speed and direction of two cars as they collide is probably a LOT more interesting than knowing the location of the accident - assuming that it wasn't inside a building.

3. Is the $100 laptop an emergent data story? I think that there could be something interesting in the patterns of usage that emerge from the new class/type of users.

4.The ID stories are good. I think that there might be some interesting stories about motes here. A network ofRFIDs that self-organize and tell you about patterns of movement and what things move together might be compelling.

5. I think that the bio-examples are clear spaces where our ability to measure and our ability to respond are out-of-synch. Biological systems are complex and many times slow-acting, e.g. our endocrine systems. Instantaneous access to many facts may or may not be powerful. Again, it depends on the process that we are studying. Heartbeats, yes. Hormones, no. This may help structure some thoughts about what is useful.

6. Preference is the holy grail of marketing information but again, it must fit within a holistic structure that says that I can recognize and economically satisfy the preference before a competitor or a substitute. We should measure the things that we can act upon in a positive manner.

7. I think that the OnStar example is the most compelling because it marries a business architecture and technical architecture with the capability. I might be thinking too closely to the things that I understand as opposed to the things thatare possible, but this was the example that resonated the best with me.

8. I think that in your section about finding emergent data - your comment about knitting together facts in a coherent way is very interesting. One type of emergent data is the consolidation of facts. As credit grantors, Master Card had a more interesting view than Marshall Fields or Amoco individually, Even though each was complete within its domain. Experian and Trans Union had even more powerful views as they consolidated credit info. I think that the job of gathering facts and making them coherent across organizational or domain boundaries is very powerful. The ability to consolidate and federate databases across boundaries will probably be a bigger and bigger business.

9. I think that the messaging at the end is very appropriate. Would it be too much to try and think of a framework - perhaps I could talk through this with you - of how to coalesce the ideas? Where should they look for these opportunities and how do they structure an opportunity to exploit the new facts?

This topic gets more interesting as I think about it. I'm glad that you invited me to consider it with you. Perhaps we should try and discuss some of this over lunch or coffee some day. I have a lot of war stories that might provide some color to your good ideas.

Please don't take the thoughts that are critiques as criticism. As usual, you have developed an interesting set of ideas into a platform for productive discussion.

Alan

[Vladimir]

Very interesting material. A couple of thoughts. First, you may find somewhat useful an information-related article Carl Hugener and I just published in Banking Strategies http://www.bai.org/bankingstrategies/current/decoding/index.asp . It provides FS examples of "emergent" information and some other uses and steps. One premises of the article is that with "emergent" information (using your word) companies would become PRODUCT and INFO businesses - similar to GEs and IBMs of the world becoming PRODUCT and SERVICE businesses in the 90s. Please let me know if you have any questions about the article. Second, to strengthen your "How to" section, I would a) give more thought to the "water-source" of information: who owns it, who controls it, who can use it and how - because that would delineate what can and cannot be done with the information, including the privacy issues around it, and b) discuss the new business and revenue opportunities that would present themselves to those who owns or controls AND can leverage the information not just for their own internal purposes: existing customers, products, services, but also to sell to others (either commercial or retail users) - either doing it alone or with other "suppliers" in the information "manufacture" process.

[Scott Busse]

Great Article...in general I like the flow but I would have to agree with Jeff on the relevence of the sea clocks example.

To expand on the emerging data opportunities provided by On-Star and RFID, consider how the transportation industry could benefit just by using information that these technoligies already provide. On-Star, RFID, toll programs like I-Pass, and insurance tracking devices all provide detiailed information on traffic patterns and driving behavior that is relatively untapped. This information comes at almost no additional expense. However, the challenge will be the consumers that view this as infringing on their liberties.

[Henry Hwangbo]

Very good read - I like the information, especially about casino chips and RFIDs. Seems like "tracking" technology is more readily accepted in areas where people are accustomed to being watched. Why would these people agree to this type of scrutiny? Because they have the (perceived) opportunity of making some money and that is inherently worth the cost of giving up some liberties (at least the gambler has been trained to think so).

In the same sense, the OnStar & GMC situation is a great example of customers who are willing to give up liberties in order to have the opportunity to save money on their car insurance.

So, in some cases, perhaps the collection and application of emergent data would best be applied to "closed" systems where there is the opportunity, perceived or otherwise, for financial gains or cost savings that outweigh the sacrifice of certain liberties to certain individuals.

Let's apply this to the recent XCP rootkit debacle, courtesy of Sony that has back-fired and arguably set back the DRM debate for years. What if Sony provided two different price points for CDs - one for people willing to allow Sony to track usage and a higher price point for non-trackable CDs. Given the right incentive and cost savings - would an honest user go for the lower price point in exchange of certain liberties? This would put the power in the hands of the consumer, to CHOOSE over liberties and the cost of keeping those liberties.

It would be interesting if someone was able to QUANTIFY the cost of certain liberties. The price point in which someone is willing to give up these liberties would be an interesting talking point.

That said, I think another interesting fall-out from the creation of all this data is the inability for the average person to successfully find what they are looking for. Google no longer solves this issue as people are constantly "gaming" the search engines and even now, the ability to target and retreive exactly what the user is seeking, has been reduced.

For example - if I am looking for a review on a new product the top results from Google are usually online retailers with a review section that is unfilled. I can try certain tricks like, "-cart" to remove results that having a shopping cart on the page or the "site:" directive to search certain review-type sites, but shouldn't there be an easier way for the average person to find things they are looking for?

I find myself spending more time trying different search keywords to locate exactly what I'm looking for, then I spend reading the results of a successful search. As Google sets its sights on other ventures, I wonder if the simple commandline-like textbox as a search interface has reached its end, but we're stuck with it because Google is the only thing we have until TNBL (The Next Best Thing).

Here's hoping that someone figures out a better interface for leveraging all that information, emergent data, or otherwise.

[Jeff Hughes]

Chunka – I like the material. I have just a couple of thoughts on the overall theme that connects the examples.
 
To me, in reading through the examples that you offer – mileage tracking in insurance, genetic information, Baja Beach Club chip, RFID’s – are tied together by the fact that smaller devices, optimized for a single purpose, are providing much richer sets of information, particularly when correlated with other single-purpose data sets.  Concurrent with this trend is the fact that individuals seem to be dropping barriers to third parties collecting this information – again targeted for a specific, benign purpose – even at a time of heightened sensitivity regarding ID theft, etc.  I think the Information Advantage is still the right umbrella, but it’s this shift that is the new element in achieving that advantage.
 
If this is the unifying theme, the sea clocks example may not be particularly relevant.  Also, the $100 laptop, while certainly an example of a low cost information device, is more about providing capabilities to a population that never had them versus that population providing information back into a data set owned by a company, government, etc.  There is certainly a case, however, that once these laptops become part of a growing network, information will be provided by that population that may drive truly transformational change – not just lower insurance rates or more targeted customer segmentation.
 
Thoughts?