- Dealing with incentives; HR types cringe when they hear things like “winners will get an ipod” or ”$500 goes to top trader in sales market”.
- Belief;The theory is easily understood but the gut level belief is hard to get through when running internal markets. Convincing managers who have been running their organization for years and say “no market is going to change my mind” (even when presented with experimental evidence to the contrary…) are hard to change.
- Attention and time; Getting peoples attention and time to play can take more energy then you expect.
- Incentives again; the incentives get players to the market sometimes but it does not keep them there. There has to be real value to the player other than corporate goals and prizes. It is the “What’s in it for me” factor.
- Software; the software has to hide as much of the terminology and market geeky stuff as possible. Most people know what a stock market is but jargon and terminology turn people off (maybe why saving for retirement and college is hard to do?).
- Even if the market is run and real data is available to make decisions, will a decision be made? Analysis paralysis kicks in and market data is second guessed (the market isn’t the problem, it’s the culture of decision making that is).

Thanks for posting here; clearly you’ve had some experience.

I have a couple of questions/ideas to run by you…

- When it comes to incentives, did you ever give any thought to less-valuable incentives? Bo Cowgill mentioned that at Google they gave away hats, t-shirts, etc… the standard big-company promo gear, and their markets were successful. Perhaps there’s a middle ground where it could still get people interested and not HR worried.
- Incentives to keep people trading is an interesting concept. In my current research project it hasn’t been a problem at all, mainly because it is sports-related. (So far I’ve offered zero financial or “swag” benefits, it should be noted!) My current thoughts are that people will participate in markets where they have some sort of interest or passion. Perhaps if through prediction markets they can feel part of the decision-making process that would be enough to keep them coming back.
- That of course leads into the corporate hesistancy towards these tools. If managers are hesitant, their employees will be as well. But is that all it takes? One or two enlightened managers to buy into these tools and make them work in their organisation?
- As for the software, I’ve realised that that could be a major stumbling block when working with novice traders. That’s why I wanted to use Inkling for my project; though it’s still beta and can and already is getting better, it’s much more intuitive than the standard “market” interfaces of other sites. I will guess that this element of prediction markets could be one of the biggest barriers to success; trading must be easy, intuitive, and (in some cases) fun.

I’d appreciate any comments/responses.

I just spend a couple of months designing a large-scale prediction market for a multinational corporation. Apart from the basic critics on adaptation and resistance from (especially middle) management, it seems to be that there is a more fundamental problem. This problem has been slightly touched in a previous comment but is in my opinion more severe: If no decisions can be made based on the market outcomes, the tool is just another management fad, not likely to become adapted in a business context.

An example: In my market, around 1.000 people (collective wisdom right?) will be trading in a market with underlyings defined as sales-forasting. Contracts would be index contracts such that the pay-off depends on stock price movements and a “bonus” in case of a correct forecast. Trading would be anonymous as to allow for unbiased information aggregation. Management would use the stock prices as indicators of a changing environments which leads to the following challenges:

1. They will have absolutely no idea what causes a stock to rise or fall as the stock price is just a derivative from an underlying event. They will also have no idea how to find that out for the “smart crowd” is anonymous and located in a large number of countries. While stock prices might thus say something, it would be extremely difficult to find out what. The HP experiment of Chen and Plott then is a good example of how a prediction markets should NOT be designed in a non-academic setting.

2. The design of the width of underlyings as well as contracts is extremely important and not covered by any academic or real-life experiments. As the distribution of shares and contracts will determine the accuracy and complexity of the market, the design should be carefully considered. For example, should you set all shares equal? Or should it be based on the sales of the last year? This however, is less of a problem than the first point.

3. People JUST DON’T get it. In this organization, employees are highly educated but have great difficulties understanding the value and principles of prediction markets.
4. Financial officers and controllers argued that, IF it worked, the prediction market would create an extremely nervous financial organization. For one, including uncertainty in the forecasts is not commonly done in practise and would cause great challgenges for SOX compliance matters. Secondly, an ever changing stock price might correctly reflect the dynamic environment but would be totally useless for financiall professionals when combining with normal rolling forecast procedures.

I’m rounding up my research and prediction market experiment at this moment, so I will be available for questions and suggestions on this forum!

Greetz,
Mark

Mark – I’m still trying to figure this out.

At a horse track, some people bet on #7 because the silks are blue and they like the color blue. In the stock market, some people buy a stock because they like the environmental posture of a company. The obvious point being, many people bet without any or enough objective analysis (or even just an objective opinion). Furthermore, there is a difference between “betting on” (akin to “voting for”) something you hope wins vs. betting on the thing that you can buy for the right price because you think it will go up in value.

I know I need to read more about these prediction tools, but if your goal in your organization is to get the wisdom of the crowd for the next product to build or a reasonable forecast to set, isn’t that at odds, at some level, with a trading market where the purpose is to make a return?

Mark Bauer

Mark,

I hope your project went well. I have a quick question for you.

When you say stock prices, do you mean that people at your company were trading on the stock price of the company itself? Or do you mean to say contract prices? I would agree that trading on stock prices is a very inefficient way of forecasting anything, but I would disagree with you a bit if you mean contract prices.

I think that if a market is tracking the next quarter’s sales of item X and there is a significant change, a manager will always be able to track that down to find out why. This is the whole point of the market; forecasting methods that tell you something you know are easy, but forecasting surprises is very difficult. In the marketplace, employees have an incentive to reveal their private information that may not otherwise be revealed. (Such as, there’s no chance in hell that Y will be done in time / meet sales goals / etc.)

Regarding your second point, I would refer to you Chris Hibbert here:
http://pancrit.blogspot.com/2006/09/continuous-...

Your third point is very true; some people won’t and potentially never will “get it.” However, you don’t need that. You just need a diverse set of people that do. Some will trade no matter what because it interests them, some will trade with enough incentives, and others will never trade. So long as you have enough traders to make your market efficient, this point doesn’t matter as much.

Finally, I think the forecasts that result from these markets are actually quite stable once they’ve been running for a while. Initially as people enter the market there will be some volatility, but this should dampen out with enough traders. (This will depend on the type of interaction you’re using: CDA, MSR, DPM, etc.)

Please feel free to discuss this with me at any time. If you search for Mercury Research and Consulting you’ll find me.

An initial point in this thread was that HR usually cringes at rewarding the players in prediction markets. However, the same HR folks don’t give a second thought to giving millions of dollars worth of stock options to the executives. While executives can probably have a direct impact on the stock price of the company, the average employee has almost no bearing at all. It seems only fair to reward people in the trenches for adding shareholder value when they help to guide a successful prediction market…