IBM Public Policy Prediction Markets: Collective Wisdom on Education, Transportation, Energy and Healthcare

IBM Smarter CitiesIBM recently launched its Smarter Cities initiative. Part of its overall SmarterPlanet project, Smarter Cities is an effort to find solutions to the problems that will occur due to our ever-increasing population growth in urban centers around the world:

In 1900, only 13% of the world’s population lived in cities. By 2050, that number will have risen to 70%. We are adding the equivalent of seven New Yorks to the planet every year.

This unprecedented urbanization is both an emblem of our economic and societal progress—especially for the world’s emerging nations—and a huge strain on the planet’s infrastructure. It’s a challenge felt urgently by mayors, heads of economic development, school administrators, police chiefs and other civic leaders.

IBM has the smarts and global heft to be a major voice in innovating solutions for the problems that urban population growth will bring on. And of course, it doesn’t hurt that there will be government expenditures to make sure we’ve got the infrastructure ready.

IBM CEO Sam Palmisano laid out three fundamental changes to global urban areas:

  1. Our world is becoming instrumented: Sensors and devices are coming down in cost, and increasing in functionality, giving us “for the first time ever, real-time instrumentation of a wide range of the world’s systems”
  2. Our world is becoming interconnected: With the rise of devices with these sensors, “systems and objects can now ‘speak’ to one another”
  3. All things are becoming intelligent: Better sensors, increased computing power and more information from interconnection mean that “intelligence can be translated into action, making our systems, processes and infrastructures more efficient, more productive and responsive-in a word, smarter.”

The sensors thing is interesting. I’ve heard both Tim O’Reilly and Paul Saffo talk about sensors as the big area of technology growth and opportunity.

As part of this initiative, IBM (in conjunction with Spigit) is running a series of prediction markets that you can participate in. The objective is to tap the collective wisdom of people around the world. Here  are the prediction markets for which they’re seeking your perspectives:


  • Which approach will be most effective in enabling better education outcome within a major city? (link)
  • In order to increase the proportion of the population completing high school by 10% over the next five years; major cities will begin transforming education in what way (link)


  • Which company offers the best portfolio regarding Smarter Transportation? (link)
  • In a major city, what will need to be improved in order to make transportation more efficient? (link)
  • What enhancement can a major city make over the next year to be a global technology leader in public transportation? (link)
  • What transportation enhancement will a major city, like New York, need to make to relieve its traffic congestion? (link)


  • Which of the following will be the most important to the rapid deployment and adoption of Smart Grids? (link)
  • Over the next five years, what changes should a major city first implement to reduce energy waste and use its resources efficiently? (link)
  • Which of the following will reduce household energy consumption the most within a major city like New York? (link)
  • Which of the following should be a primary objective for a major city over the next five years? (link)

Government Services

  • The current economic crisis will change plans for high priority projects in a major city in which way over the next few years? (link)
  • If you were a mayor of a major city, which method would you use to assess the needs of your city, the business community and your citizens? (link)
  • In 2011, what will be the primary method for citizens to communicate with their smarter city governments? (link)
  • What immediate step should a major city government take over the next year to emerge as a leader in e-governance? (link)

Public Safety

  • Over the next five years, what transformation will large cities make to their public safety systems to reduce the physical / personal crime rate against people, property, and infrastructure by half (50%)? (link)
  • If a large city wants to improve its overall public safety position (i.e. reducing traffic fatalities, decreasing gang violence, improving emergency response capabilities) in which public safety area (or related city sub-system) should it target investment over the next year? (link)


  • Which of the following sub-system improvement will be most effective in providing immediate benefit to healthcare delivery for citizens in a leading smarter city? (link)
  • Over the next five years, what will major city hospitals do to increase efficiency and deliver better quality healthcare to its citizens? (link)


  • What are the top challenges large cities (i.e. populations over 5M) within emerging markets will face within the next five years? (link)
  • What region(s) will recover most quickly from the current global economic crisis? (link)

If addressing these issues is something that interests you, check out IBM’s SmarterCities Predictive Idea Markets.


My Ten Favorite Tweets – Week Ending 061209

From the home office in Palo Alto, CA…

#1: RT @palafo Facebook URL rush should have been hashtagged #nerdolympics. “Just sayin’. ”

#2: Enjoyed the Building43 launch at TechCrunch’s offices tonight. Knock ’em dead @scobleizer Looking forward to following and participating.

#3: Reading: Why SaaS Has Better Functionality than Enterprise Software

#4: Left comment on New York Times post, The Stalled Promise of Innovation Really, it’s not bleak, we’re doing fine.

#5: New Spigit blog post: Medplus Built Its Innovation Program with 12 Moose-on-the-Table Questions #innovation

#6: RT @innovate Knowledge Management is more about “How do I?” while Innovation is more about “Why don’t we?” – #yam #innochat

#7: Participating in an ABC7 prediction mkt: Will Dianne Feinstein run for governor of California in 2010? I’m betting ‘no’

#8: RT @Hammarstrand Top 30 Failed Technology Predictions. #innovation #tech #future

#9: TV news story here in SF about the CA education budget cuts, shows a teacher out of a job as “layed off”. Guess the cuts are hurting already

#10: Kinda sad…took down the crib tonight. Our 2 1/2 y.o. is sleeping in her own big girl bed, our 5 y.o. long ago left the crib.

Google, Yahoo, Microsoft Want to Legalize For-Money Prediction Markets

$500 on the U.S. economy turning positive in the first quarter of 2010!

Wouldn’t it be great if you could put money down on your predictions of future events? If Google, Yahoo and Microsoft get their way, you just might be able to do that.

Money $20sBack in September 2008, Google and Yahoo, united under an organization called Coalition for Internal Markets (CIM), wrote a 28-page letter articulating their support for the legalization of small stakes prediction markets. On April 9, 2009, Microsoft added its support to Google and Yahoo’s letter. Here’s an excerpt from the CIM letter:

CIM believes that small-stakes event markets of the kind first developed by the Iowa Electronic Markets have the potential to provide significant public benefits and recommends that the Commodity Futures Trading Commission propose regulations under which such markets may operate, both as internal markets or as public markets.

I learned of all this through Oddhead, Midas Oracle and Bo Cowgill’s blogs. This has the potential to be quite powerful as a forecasting tool, and a way for people to profit from their prediction acumen.

Just how did this come about?

Commodity Futures Trading Commission Wants Input

The Commodity Futures Trading Commission (CFTC) is the government body that regulates the sale of commodity and financial futures and options.

In May last year, the CFTC put out a public notice that it was soliciting comments on the regulatory treatment of financial agreements offered by prediction markets. So apparently the idea of legalization is on the Commission’s mind. The CFTC distinguishes prediction markets as not including financial agreements on market prices (stocks, cotton, etc.) or broad-based measures of economic or commercial activity. Rather, they define them as:

Event contracts may be based on eventualities and measures as varied as the world’s population in the year2050, the results of political elections, or the outcome of particular entertainment events.

“Entertainment events.” Think American Idol, and putting your money down on who you predict will win. That Adam Lambert?

The CFTC notes that its staff has received “a substantial number of requests for guidance” on the propriety of prediction markets’ use. Sounds like a pretty healthy interest in this sort of thing.

Getting Ahead of the Regulatory Curve

In the CIM letter, Google’s Hal Varian and Yahoo’s Preston McAfee develop three themes:

  • CFTC has the right to regulate these markets
  • Prediction markets provide substantial benefits
  • Propose a set of sensible rules for regulation

Google states that it started operating internal prediction markets in April 2005, and that now it runs 25-30 prediction markets per quarter. The purposes of the markets include forecasts of product demand, internal performance (e.g. product release dates), company news and external business environment factors. Google also uses the prediction markets to assess the strength of relationships between different teams.

Yahoo operates internal prediction markets. It also operates public events, such as the Yahoo!-O’Reilly Tech Buzz Game, in which participants predict which technologies will be popular, and which ones lack merit.

The two primary benefits discussed in the letter for predictions markets are: (i) Generation of useful information by aggregating the opinions of individual participants; and (ii) Hedging exposure by making predictions related to some position an individual holds.

The two companies then smartly propose some rules that would govern the small stakes prediction markets:

  • Total exposure per market of $2,000
  • Maximum loss at $2,000 over the course of a year
  • Non-intermediated, electronic markets
  • Trading could be matching bids and offers, or there could be an automated market maker
  • Program to monitor trading
  • Maintain trading histories for five years

Generally, the letter asks for a fairly flexible approach to the markets, with adherence to core operating principles to ensure fair, open trading.

An Inevitable Question: Gambling?

Perhaps as you’ve read this, the thought occurred to you…isn’t the same thing I can do in Las Vegas? Bet on sports teams? What distinguishes this from gambling? Indeed, in its solicitation for comments, the CFTC asks this:

What objective and readily identifiable factors, statutorily based or otherwise, could be used to distinguish event contracts that could appropriately be traded under Commission oversight from transactions that may be viewed as the functional equivalent of gambling?

The CIM letter notes that gambling is generally associated with sports events and games of chance. It recommends the CFTC develop a definition of permitted markets based on a set of examples, and expand the list on a case-by-case basis.

This question will likely receive the most attention from the public. What will be interesting is how Obama’s administration views this versus Bush’s.

Count Me In

Add my YES vote to this. I think it’d be great to buy and sell positions based on predicted event outcomes. The example I led this post off with, the economic rebound, is a great way to tap public sentiment about the economy. We’ll have to watch how this unfolds.

How about you? Do you favor small stakes prediction markets?

How Much Scale Is Needed in Enterprise 2.0 Employee Adoption?

A couple recent items caught my eye with regard to the issue of employee adoption of social software.

In Reversing the Enterprise 2.0 Pricing Model, Julien le Nestour argues that pricing per user for social software should increase as more employees use it, because the network effects of higher participation make the software more valuable. It’s a great theoretical piece, tying pricing to value received. But in the harsh budgeting realities of the enterprise and in the comparison against other software pricing models, it’s not likely we’ll see anything like this.

Atlassian, maker of the Confluence wiki and developers tools, recently passed the cumulative revenue mark of $100 million. In the post announcing this milestone, Atlassian blogger notes that the company has no sales force. People just download the app. I know some of the Atlassian guys, and this kind of viral, bottom-up adoption is core to their philosophy. They don’t sell to upper management, adoption occurs at the departmental level. That being said, I am aware from my work at Connectbeam of some large-scale rollouts of the Confluence wiki by Fortune 500 companies.

What connects these two items? The first post describes the nature of Enterprise 2.0 apps and how their value increases as more employees use them. The second post points to the value that departments have received from Atlassian’s Confluence wiki, even without broad adoption. In other words, network effects are not a critical aspect of the Confluence value proposition.

From these posts, other readings and direct customer experience, the following occurred to me:

You don’t need a high level of adoption to get value from some Enterprise 2.0 apps. Others require broad participation.

In some ways, that may seem obvious. Yet I don’t tend to hear this distinction being made. Usually, all social software is lumped together under ‘Enterprise 2.0’ and there is a collective view that wide-scale adoption by employees is a necessity. It’s actually more nuanced than that.

Varying Adoption Levels Required

The graphic below depicts the relative levels of participation required for different apps to “deliver value”:


Here’s a quick summary of the graph:

  • Employee participation is defined as contributions and engagement (views, edits, comments, etc.)
  • Moving from left to right, the percentage of employees involved gets higher

This graph has a couple of implications for Enterprise 2.0 vendors. Before that, here’s an explanation for why I put the different applications where I did.

Consider the Purposes of the E2.0 Applications

Before discussing these applications, I want to note this. All social software applications get better with higher adoption. There is no disputing that. The distinction I want to make is that some apps require increased participation before they deliver value.

Blogs: The nature of a blog is a single person’s thoughts, observations and ideas. Inside companies, these applications can be tools for the ongoing recording of things that fall outside the deadlines and process-oriented activities that make up the day. Making them public is a great way to share these contributions with other employees and establish your record of what’s happening. If only a few key people blogged inside a company, there will be value in that.

Wikis: Wikis actually have two purposes: (1) knowledge repositories, and (2) projects and collaboration. It’s that second purpose that makes wikis particularly valuable even with small participation. I’ll use Confluence as an example. We use it as our low home for putting up documents accessible to anyone else, and for free-form contributions on all manner of things. It is very much a utilitarian use case for us. If we weren’t using Confluence for this purpose, we’d share documents via email. In larger organizations, Confluence may replace usage of SharePoint or the company portal.

Using wikis as knowledge repositories, such as [Company Name]-ipedia type of implementations, requires a larger percentage involvement. Sparsely populated company versions of Wikipedia are of little use. As are wikis that are not updated regularly with new information. I’d put wikis-as-knowledge-repositories up there around prediction markets in terms of required participation.

Forums: The old man of Enterprise 2.0…forums. These are the place where topics can be posted, and a scrum of conversation occurs. To really get value out of these, it helps to have larger participation. Blogs are solo voices with interesting content. Wikis can have a very specific collaboration purpose among a few employees. Conversations around a topic require a wider variety of voices. Otherwise they fail to give people a sense of what others are thinking. Nothing sadder than forum post with no comments.

Social bookmarking: Bookmarking sites you find useful has value by itself. So in that sense, “social” bookmarking can work for very few employees. But it’s not really “social”, it’s simply a replacement for your browser bookmarks. You get value by finding those gems your colleagues deem interesting. The odds that any single bookmark will be useful to you are small, so you need a healthy amount of bookmarks to increase the chances of finding links that will help you. And to get a healthy amount of bookmarks, you need broader participation.

Microblogging: In some ways, microblogging could be compared to forums. Both are public places to serve up topics. But they’re fundamentally different. And that’s why broader participation is more important here. Forums have a distinct purpose – the discussion of a particular topic. You need participation by those who know something around the topic.  Microblogging is a more free-form, personal activity. You don’t need a distinct purpose to post something. You post all the things that occur to you during the day. Some of which will have value, although it can be hard to predict for whom. It also helps to know that people are seeing these posts, because there is a conversational aspect to microblogging. The free-form, who-knows-what-might-be-interesting, conversational aspect of microblogging require larger participation than forums do.

Prediction markets: Prediction markets thrive on having a variety of ideas, events and initiatives. They also require the different perspectives of employees, leveraging different perspectives, knowledge and experiences. This is true wisdom of crowds work. Limited participation limits the value of prediction markets. These benefit from broad employee involvement.

Social networks: I put these at the top of the chart in terms of employee involvement. Perhaps one of the best use cases for social networks is finding colleagues with the knowledge or interest in projects you’re working on. This requires large-scale participation. If a social network only is used at the departmental level, it doesn’t provide value. In terms of expertise location, you’re probably already aware of what others in your deparmtent know. It’s breaking out of that traditional sphere of contacts where social networks shine. I know I’ve heard many instances of large corporations suffering from “reinventing the wheel” syndrome because employees lack visibility about what others know. Broad participation addresses this issue.


Three implications of this view about required involvement come to mind.

Greater required participation correlates to greater impact on a company’s value: Generally, you could change the metric in the chart above from percentage of employee involvement to impact on company value. The increased participation means the associated application will also have a larger effect on the company’s strategies and operations. It’s not an tight correlation, but a general trendline. Exceptions will abound.

Top-down vs. bottom-up: General observation is that broader participation requires a greater amount of senior management support. That’s the way things work inside companies. Employees will listen when the executives of the company push something. For applications that need lower participation, the name of the game is to provide a compelling application with a low entry cost. Departmental budgets and the green-light from employees at lower levels of the organization are all that are needed.

Time for application to gain traction: With applications that require low levels of participation, there is plenty of time for the application to grow virally. It serves its purpose for a select few, and over time others will see the value and elect to participate. These apps can be resident inside companies for long periods of time. Those that require higher participation to see value will need to show results sooner. They are on senior management’s radar, generally cost more and have a greater number of employees who will be watching to see the results.

So it matters what type of application we’re talking about when it comes to Enterprise 2.0. It matters for companies and vendors. It impacts the skills required for everyone’s success.

A nice post that complements this one is Adina Levin’s Scale effects in enterprise social software.


See this post on FriendFeed: