Tuesday, November 02, 2010

Tapping Into the Wisdom of Clouds

Prediction markets have been around for quite a while now. The Iowa Electronic Marketplace is world renown for accurately forecasting the outcomes of US elections often many months prior to election day. Sports betting sites in the UK and elsewhere move billions of dollars of wagers on the basis of the collective expertise of those betting on outcomes.

More recently, companies like Consensus Point, Crowdcast, and InTrade (among others) have brought the tools and technologies to the boardroom that allow managers to tap into global pools of “experts” to attempt to predict the future.

  • Hollywood movie studios have made “prediction markets” a key component of their forecasting efforts in deciding how much money to spend on advertising campaigns. (Interestingly, movies rated very high or very low receive relatively little advertising as WOM is expected to play its role at both ends of that spectrum; only movies in the middle-range receive significant ad spend).
  • Retailers use prediction markets to make decisions on which products to carry in inventory, as well as how to promote and price them.
  • Technology firms use them to decide which new platforms to bet on.
  • Pharma industry leaders use them to determine pricing strategies years in advance based on competitive pipelines and regulatory approval processes.
  • And B2B industrial companies use them to model the impacts of changes in sales force size, structure, and compensation.

Perfect? No. Helpful and insightful? Definitely, in several ways.

First, by identifying possibilities not previously within the imagination of your own executive team, and by considering factors which any small group of managers might overlook, they provide a more thorough and comprehensive assessment of uncertain outcomes.

Second, even if prediction markets cannot provide an exact answer (which they rarely can – being better at offering directional probabilities than precise forecasts), they can significantly reduce the uncertainty surrounding what a given market segment might respond to, or how a group of competitors might react to a significant change by one. This makes them good tools for setting performance targets and expectations in the absence of historical perspective.

Third, with the help of cloud computing and social networks, prediction markets are declining in cost to the point that they are often much faster-to-feedback and far less expensive than traditional survey-based research, while offering far greater flexibility to have respondents explore “what-if” scenarios.

Like any tool, they can be dangerous in the hands of amateurs. Garbage-in, garbage-out is a primary risk. So is being too confident in the absolute numbers, when the directional insights are often the most valid level of granularity.

Nevertheless progressive marketing measurers are using prediction markets to help them better understand and act on the insights they’re deriving from their mix models, their web analytics platforms, and their customer satisfaction and referral studies.

There is a whole lot more to learn about prediction markets before you jump into using them. But in general, you can benefit from using them to answer questions you might be struggling to answer with your current data streams if/when:

  1. You have a suitable pool of “experts” to engage in your marketplace. These experts could be associates, customers, prospects, or industry monitors. The exact number required differs by purpose. Sometimes you can get pretty reliable data from as few as 20 participants; other uses would require hundreds (or thousands) of participants.
  2. You can define your questions in terms of “what would happen if…”
  3. You can engage expert participants by offering something of true value in exchange for their effort and energy. Offering monetary rewards, special recognition, unique access, or other benefits of great interest will help ensure a more vibrant and active prediction market that explores new ground.

Finally, two quick learnings about how to get the most from your prediction markets (based on experience)

  1. Include some “noise” traders who inject provocative suggestions or wagers to ensure you draw reactions (supportive or contrarian) from the smarter traders with better insight.
  2. Run your markets as shorter-term events, and not continuous commitments over extended periods. Request only short-bursts of participants’ time; provide feedback quickly; and progress continually towards a defined end-point.

There’s a great deal of un-tapped insight potential in the clouds. Creative approaches are generating terrific new insights into marketing effectiveness and efficiency at increasingly faster rates. And the subset of “difficult to answer” questions is getting smaller and smaller every day.
Pat LaPointe is Managing Partner at MarketingNPV – specialty advisors on measuring and improving the payback on marketing investments, and publishers of MarketingNPV Journal available online free at www.MarketingNPV.com.

No comments: