Practical dos and don'ts of enterprise prediction markets

Eﬀec%ve Enterprise Markets:  What makes them work and Why  Dr. Robin Hanson, Chief Scien%st  Rebecca W Munn, Senior Vice President  KM Chicago  September 14, 2010 

Agenda  •  Introduc;ons  •  What makes Enterprise Markets work  •  Why markets as a medium for predic;ng the future  •  Enterprise Do’s and Don’ts   •  Discussion 

What Makes Markets Work  Buy Low, Sell High  Will price rise or fall? “Pays $1 if Obama wins” sell E[ price change | ?? ] buy price sell Lots of ?? get tried, buy price includes all!

What Makes Markets Work  Example: Current Event Prices  70‐71% Republicans win US House in 11/2010  42‐47% California legalizes marijuana in 11/2010   6‐16%  California credit default by 2011   10‐20% US overt strike on N. Korea by 4/2011    24%     US or Israel overt strike on Iran by 2012  10‐19% US Sup. Court bans med. mandate by 2012  16‐18% Palin is Republican nominee in 2012  11‐30% Japan says it has nuke by 2013  45‐48% US Cap & Trade system by 2013  12‐14% China war act on Taiwan by 2013  15‐38% Higgs Boson seen by 2014  www.InTrade.com

What Makes Markets Work  Beats Alterna%ves  •  Vs. Public Opinion   –  I.E.M. beat presiden;al elec;on polls 709/964 (Berg et al ‘08)  –  Re NFL, beat ave., rank 7 vs. 39 of 1947 (Pennock et al ’04)  •  Vs. Public Experts  –  Racetrack odds beat weighed track experts (Figlewski ‘79)   •  If anything, track odds weigh experts too much!   –  OJ futures improve weather forecast (Roll ‘84)   –  Stocks beat Challenger panel  (Maloney & Mulherin ‘03)  –  Gas demand markets beat experts (Spencer ‘04)  –  Econ stat markets beat experts 2/3 (Wolfers & Zitzewitz ‘04)  •  Vs. Private Experts  –  HP market beat oﬃcial forecast 6/8 (Ploh ‘00)  –  Eli Lily markets beat oﬃcial 6/9 (Servan‐Schreiber ’05)  –  Microsoi project markets beat managers (Proebs;ng ’05) 

What Makes Markets Work  Example: Iowa Electronic Market  Item 1988 1992 1996 2000 2004 All # big 59 151 157 229 368 964 polls Poll 25 43 21 56 110 255 “wins” Market 34 108 136 173 258 709 “wins” % 58% 72% 87% 76% 70% 74% Market P-value 0.148 0.000 0.000 0.000 0.000 0.000 “Prediction Market Accuracy in the Long Run” Joyce Berg, Forrest Nelson and Thomas Rietz, Jan. 2008.

Why Markets as a Medium for Forecas%ng  Advantages  Incen(ves  •  Numerically precise  Self‐Selec(on  •  Consistent across many issues  Correct Biases  •  Frequently updated  •  Hard to manipulate  •  Need not say who how expert when  –  Issue is not  experts vs. amateurs  •  At least as accurate as alterna;ves 

Why Markets as a Medium for Forecas%ng  Collec%ve Forecas%ng  Forecasts  On Requested  User   Topics  Contribu;ons  User Scores  Engagement 

Why Markets as a Medium for Forecas%ng  Collec%ve Forecas%ng Ques%ons  What exactly is   Consensus  What exactly are my  my inﬂuence?  incen7ves?  My Forecast  My Score  How exactly do   I express my  Truth  opinion? 

Why Markets as Medium for Forecas%ng  Edi%ng Interface Is Transparent  If my edit increases   I directly   Consensus  the consensus chance   change the  of true state, I win.   consensus  Else, I lose.  My Edits  My Score  Truth 

Why Markets as Medium for Forecas%ng  Factors Might Inﬂuence Sales  E[Sales|Factor]  P[Factor]  •  Economy recovers fast?  •  Compe;tors introduce new version?  •  We do big promo;on?  •  We lower prices?  They lower prices?  •  We add distribu;on channel?  •  We add feature F?  They add feature F?  •  Our defect rate very low? 

Why Markets as Medium for Forecas%ng  Issues  Input: Contribu%ons  Output: Forecasts, Scores  •  What info can express?  •  What ques(ons can ask?  •  How account for costs?  •  How account for value?  •  Who let in where?  •  Use or validate system?  •  Enough Incen;ves  •  Should adjust outputs?  –  T‐shirts enough?  –  Zero‐sum scoring?  •  Who let see outputs?  •  Limit Costs  •  Sabotage & manipula;on  –  Awkward Interface  •  Legal, P.R. risks?  –  Wait for oﬀer accept  –  Retribu;on 

Why Markets as Medium for Forecas%ng  Requirements  Use: •  Ques%ons really want answered now  –  Will eventually know answer (or parts)  –  Suspect not geong frank info via usual channels  –  Don’t mind par;cipants knowing best es;mates  •  People who have or can get key info  –  Their ;me is the main cost  –  Lihle penalty for invite many don’t know  •  Incen%ves to en;ce careful par;cipa;on  –  Money, ahen;on, inﬂuence can legally oﬀer  –  Valued when ques;ons answered  Validate: •  Many fast ques;ons, status quo es;mates 

Enterprise Do’s and Don’ts of Enterprise  Markets  Customer  •  Do  –  Integrate into Enterprise processes  –  Nurture execu;ve sponsorship  –  Make accessible to all  –  Customize to your business  –  Make it Part of your Value Prop  14 

Integrate Into Enterprise Processes  Customer  Challenge  Ac;on  Result  • Facing too many  • Deployed internal  • Increased accuracy  unknowns, leaders  market, integrated  from market has  were planning for all  into exis;ng  streamlined short  possible futures  Enterprise  and mid‐term  processes and  planning  • Costly and ;me‐ dashboards  consuming  • An;cipated  • Appropriate ;me  improvements in  horizons matched  metrics, e.g. cash  Enterprise  management,  deadlines  balance sheet  15 

Nurture Execu%ve Sponsorship  Customer  Challenge  Ac;on  Result  • Market introduced  • Sought exec air  • Balanced top‐down  by self‐ini;a;ng  cover, not mandate  endorsement with  team, not C‐suite  • Targeted key  bohoms‐up culture  • Fit entrepreneurial  leaders as early  • Grassroots growth  culture  adopters  more sustainable  • S;ll needed to drive  • Champion is one of  • Consulta;ve  awareness,  top users  rela;onship with  par;cipa;on, and  • Opened market to  key leaders paves  execu;ve usage  everyone   way for followers  16 

Make It Accessible To All  Customer  Challenge  Ac;on  Result  • Accelerate the  • Incorporated  • Improved overall  process of  market into  ;me to market,  priori;zing ideas  broader innova;on  speed of idea  and product  plaqorm  ac;on‐ by 100 days  features  • Deployed market  • 5‐10X ROI  • Engage more  globally  • Tripled  employees in  • Feed ideas to every  collabora;on and  innova;on  business unit  cut duplicate ideas  in half  17 

Customize It to Your Business  Customer  Challenge  Ac;on  Result  • Looking for new  • Market well‐named  • Made it familiar  ways to engage  and branded  and instantly  their customer base  • Seamless  accessible to target  • Seeking  integra;on with  group  differen;a;on from  company website  • Drove par;cipa;on  the significant  • Develop custom  and accelerated  compe;;on in their  interface to match  engagement  market   user community  • Differen;ated from  compe;;on  18 

Make It Part of Your Value Prop  Client  Challenge  Ac;on  Result  • Customers  • Recognized market  • Provide clients with   represent broad  as amplifier of  beher data via  array of industry  exis;ng informa;on  public market  stakeholders  and service  • Enable client  • Universal client  offerings  specific experience   need: predic;ng  • Integrated market  • Provide superior   behaviors  into all aspects of  services and ;mely  • Tradi;onal research  how they create  insights, create  & forecas;ng  value and conduct  compe;;ve  insufficient given  business  advantage  industry changes  19 

Enterprise Do’s and Don’ts of Enterprise  Markets  Customer  •  Don’t  –  Run a 30 day pilot   –  Use a similar group of par;cipants  –  Run a market without promo;ng it  –  Assume leaders will act on insights  20 

Discussion 

Practical dos and don'ts of enterprise prediction markets

by KM Chicago on Sep 18, 2010

Accessibility

Tags

Upload Details

Usage Rights

2 Embeds 15

Statistics

Practical dos and don’ts of enterprise prediction markets — Presentation Transcript

http://kmchicago.blogspot.com	14
http://kmchicago.blogspot.pt	1