... Continued from Part I
RD> Tell me more how these game theory models actually get used.
BF> Through a series of surveys and workshops given to the executives and thought leaders of our client, we extract the three components required in a game model . First, is the “who” ─ the players in the game. These are current and future competitors, partners, regulators, and any stakeholder that can affect the outcome. Second, is the “what” ─ what strategic options are available to each player. Third, we capture the interests of each player. What is a player trying to accomplish? The discussion around this third part is quite revealing because we codify all the players interests using a game theory construct called “preference trees”. A preference tree is a stack-rank list of all the options of all the players from most important to least important. Each participant role plays being a player in the game to construct that preference tree. The preference tree expresses the importance of each option either positively or negatively. That is, a player desires an option to occur, or fears that an option will occur. Priiva takes these inputs, goes away, runs all our mathematical models, and returns to the client with analysis and predictions.
RD> Then what happens?
BF> The analysis and predictions drive the strategic actions that need to be taken. Metaphors often emerge that help communicate that strategic plan. For example, “a rising tide” might describe an immature market where all the players benefit from a technological breakthrough by any of the players. At the end of the workshop, the client is left with contingent scenarios, has a deeper understanding of their market, and is well prepared when big headline events actually occur ─ like a competitor announcing an acquisition, a government scandal that triggers economic turmoil, or new regulatory surcharges on product shipments.
RD> Are these preference trees a proxy for risk tolerance of a competitor’s management team, and not just a proxy for preferences? If yes, do you think it creates a data collection bias because the participant in the room will probably bring in their personal risk tolerance and preferences while role-playing Player X? If no, how do you elicit those “real” risk preferences of Player X?
BF> Yes, it is a proxy for risk tolerance as well. We attempt to eliminate personal bias by confirming each preference tree with group thinking. We also re-run our models tweaking key assumptions. In terms of the risk tolerance of competitors, that comes out in the role-play as the participants need to really understand what makes Player X tick, including their risk tolerance. The team needs to learn everything they can about Player X ─ their personality, their history, where their leadership team trained, the tenure of their CEO, their ownership structure, their swagger (or lack thereof) at investor meeting events, their words, their play, and every piece of information we and the client can get on that player.
RD> How do you drive consensus?
BF> Sticking with our Player X example, a small team would be assigned to role-play Player X, use all the available information, and construct the preference tree. Then that team would appear before the entire group and defend their tree by stating the behavior profile for Player X. This leads to a great discussion, and often the original preference tree is changed. Once the preference tree is agreed to for Player X, we will inspect where Player X’s own options are in the stack-ranking of all options. If a player’s own options are at the top of its own tree, that is an indicator the player is aggressive and will act without fear or consequence from others. Conversely, when Player X’s own options are well down in the ranking it means Player X is more sensitive to the actions of others before initiating their own actions. The models can be run few times on our system – in order to take away disagreements and give a more definitive answer. Insights are the value we provide.
RD> Because your insights are only as good as the information you and the participants bring to the table, do you see games and simulations fitting into your workshops? And do you think there is value in redirecting play online where you could create additional simulations, involve more participants, avoid “GroupThink” issues, or facilitate better decisions?
BF> Today, we use our proprietary software to facilitate the discussion we are leading. But we do not install, teach, nor provide software for clients to use. Starting in 2011, we will have the option to leave the facilitation software tool with the client so they can spark additional discussion on their own. The question about a crowd-sourced simulation model and service is an interesting one. I can envision a cloud service of hand-picked or anonymous participants simulating a game model, and creating a piece of publishable research that Priiva would monetize. Right now, we do not have that in our plans, but we are not ruling it out.
RD> Thank you Bill. It has been most lovely to hear that you and Priiva have monetized game theory and strategic decision-making, something I intensely dabbled with in my past life. Wish you all the best with increasing your client base.
BF> Its rewarding to help companies learn to make better strategic decisions. Thanks for listening.
Posts
No comments:
Post a Comment