Monday, May 11, 2009

The end of the PPX.

As some of you may know, I've been an active trader on the PopSci Predictions Exchange (PPX) for the past two years. During that time, I've marched my way to the top 15 (#14 as of this writing) out of over 33,000 traders. As of last Friday (May 8, 2009) it was announced that the PPX will be closing effective June 1st. So, I wrote the following commentary to express my thoughts on why and how this came to be.

{Disclaimer: The following is pure speculation and is not a real reflection of PopSci or Bonnier, nor am I employed in any way monetarily compensated or tied in to those respective organizations.}


There's pretty much only one way to sum up all of this: economics. To dig down deeper we look at the history of PopSci, and to do that we turn back to January 25, 2007, the day that Bonnier Corporation bought PopSci, along with 17 other magazines from Time Warner. This was during a time when magazine subscriptions and sales were dropping, while online readership increased (and if you’ve been keeping up with the news lately, you know that this has only accelerated.)
Unfortunately, the ad and subscription revenue model changes when things start moving towards the online world. There are a few reasons for this. For one, a good bit of the content that can be found in the magazine can also be found online, for free (to the reader.) But, keep in mind that for every PopSci magazine sold, there are about 5 readers to that magazine. So, in other words, of all of the readers, only 1 in 5 is actually paying for it. And those that do are only paying around $1 per month. You can take that $1 and divide it up into readers and that means that each reader contributes about 20 cents a month on average. So the with the online world there really isn't much lost in the way of subscription revenue, it’s the ad revenue that counts.
You'll notice that in the back of each PopSci mag, there are lots of what I call "secondary" or "classified ads." These ads sell everything from you-know-what enhancement and remedies to an assortment of gadgets and the likes. If you notice, those ads aren't found here online. Then there are primary ads for cars and trucks, shavers, flashlights, laptops, you know, the stuff that you're more likely to buy, or would at least be interested in. Chances are that if it is advertised in the magazine, then at one time it may have had an entry into my favorite section, "What’s New." Otherwise, there's a good chance that those companies sell products that compete against those things that are new. There’s no doubt, that a lot of marketing on behalf of the PopSci /Bonnier staff that goes on to seek potential advertisers who may be interested in placing their ads near an article that may be related to their product.

Now, the price paid for that advertisement space is all driven by supply and demand. Supply is the amount of ad space within a magazine while demand is the amount of people who want to advertise in that space. Those advertisers look at a set of more complicated factors, like the number of readers who are likely to someday purchase a product that that company offers within a given time period. Essentially, it’s the advertiser's goal to drive brand recognition. They know that you may not buy that truck that you see in the magazine today, but, eventually you just might. Generally, the higher the price the product is, the more you have to sustain your sales pitch to your audience.

So, that brings us back to our problem: economics. With the economy the way it is, the demand for print ads is down. This has been driven by the fact that we, as consumers, just aren’t in much of a mood to buy things, and besides we’re too busy hanging out on MySpace of Facebook to really care. Well, that changes the advertiser’s focus. Suddenly, they’re not as interested in placing ads in print magazines; they’re more interested in capturing your attention online. So they tell the print publisher, “Well, I’m just not that interested in placing an ad this year.” The publisher replies with a lower price, and eventually the advertiser agrees to run the ad since the price is now more affordable. So with that, you can see the problem: you’re still printing the same amount of ads, but now you’re just getting less money for them.

That brings us back into the selling of PopSci to the Bonnie Corporation. One company, Time Warner, sees that it’s simply better to sell off its underperforming brands while Bonnier sees the opportunity to gain new market share and turn the revenue stream around for its newly purchased publisher. Suddenly, new creative ideas sprang up to solve this problem of decreased print ad revenue, and increase its online ad revenue. The solution seems simple, if you can create a way to get readers to return to your website and do so, repetitively, over the long run, you can increase the demand for ad space, especially from those companies that need to keep brand recognition high. Those advertisements work best when you have a returning audience. You can charge advertisers for a pay-per-view (1000s of views) in addition to a pay-per-click fee. So out of this realization and creative process, PopSci comes up with idea to do an online exchange, yes, the PPX.

The PPX seems like it is the magical bullet needed to solve the problem. Not only will the readers be engaged in an online activity that will keep them returning, but each visit may require several clicks and thus increases the chances of pay-per-click revenue. At the same time, our ad space becomes more attractive to those long term brand recognition driven advertisers, and that could lead to pay-per-view revenue. And, to top it all off, the PPX becomes a scientific tool itself, as it becomes a science experiment to determine the viability of predictions markets.
Out of the gate, the PPX was a bit flawed. For example, despite the claim, the price of a proposition doesn’t necessarily reflect the percentage of people who think a proposition will come true. The claim is that if the price is at POP$50, then 50% of the audience thinks it could happen, and 50% does not. But in reality we know that if one person buys 1000 shares, then the price goes up by POP$ 0.25. (Now, I know a parameter has been added so that a stock’s could only move after it achieved a high enough volume - apparently to one direction.) So, if only one person buys 1000 shares, then according to the price, 50.25% percent of the traders of that particular proposition think that that event will occur. In reality, 100% of the traders of that stock, so far, agree. (On the other hand if there were 10,000 traders, that would only represent 1/10,000th of the market.) Once the stock price starts to move, Most traders bought or sold in the direction of the movement, whether they agreed or not. Now, it was argued that over time, the price would better reflect the opinions of the traders, but we all know that there was a disconnect between trader’s opinions and prop price.

Another example of a flaw was the limitation to buy only 1000 shares. Though this was important given that the price went up or down based on 1000 shares of movement, this limited the ability for a proposition to reach its true price. If the pricing was based on a scale that was determined by the total number of outstanding shares of that stock rather than 1000 shares of movement, this concept would have worked better. (It’s important to keep in mind that for every new trader who owns a prop, either short or long, the amount of outstanding shares increases by the amount purchased. This is unlike the real stock market where there is a limited quantity of shares available, for every share bought, one must be sold. For arguments sake, ignore stock splits and issuance of new stock, etc.) In other words, as number of outstanding shares per a given prop increases, the smaller the price change interval would be. The price change interval would be determined by a function of the number of newly purchased shares as a percent of current outstanding shares. There would of course be a minimum and maximum price change interval, for example: at least $.01 no greater than $0.25. Another reason that the maximum shares were limited, was to encourage people to buy more shares of different props (portfolio diversification), but the principle to not put all of your eggs into one basket would have kept this in check. Finally, the 1000 share max rule could make sense to give more traders an equal chance to participate. This is a fair and valid reason, but I think that a limitation of buying more than a percentage (based on outstanding shares) of a given prop per a predetermined amount of time would have helped. Also the pricing model, as stated above, would have reduced the propensity for market domination by a select few elite traders.

Furthermore, the game was complex and perhaps too complicated to understand for too many of those who wanted to participate. This can easily be seen by simply counting the total number of players who never traded and those who never earned as much, or lost a significant portion of their original portfolio values. Finally, those flaws were matched by visual and web design mistakes. It seemed that every time a new feature was rolled out, it was met with more problems. For example: a menu wouldn’t work, a broken links, etc.

So those flaws that I’ve described reduced the number of would-be traders. (Satisfied traders encourage new traders to join.) This all drives me back to making my point: economics. As discussed, the ad revenue is the primary driver, and the goal should have been to sale ads based on both concepts: per view, and per click. Unfortunately, the way PPX is set up, it simply doesn’t cut it. With each user, you can only get so many ad clicks per ad per a given time frame. When you think about the mechanics how the PPX is set up, most of the ads are ignored as traders are focused on trading and not the ads. Once you’ve seen the ad, each new time you see it within a short time frame, isn’t going to drive you to click on it any more than the previous times. It’s the rule of diminishing returns. After awhile, similar ads just become a blur during the trade process.
To overcome this lack of clicking, the only way to create more revenue is to have more registered users, and those people need to be return users who visit frequently. Unfortunately, out of the 1.3 million subscribers or 6.7 million readers, there were only 33,000 registered PPX users. We can only guess that a small fraction of those actually returned to play frequently. That’s the problem that we have today. PPX was set up on a 2 year contract with Virtual Specialist, the operators of PPX. This of course likely had the option to renew. As far as we know, PopSci had full intentions to renew the contract over and over again, but only provided that it brought in enough of a revenue stream to compensate for the monthly licensing and site maintenance fees and then some. PopSci knew of its problem long ago. It knew that growth was limited and that there was a diminishing rate each passing month. It tried to solve the problem by correcting some of the problems cited by its users, and they were able to do this by hiring better staff members who understood these dynamics. They even spent the money to upgrade the look and feel of the site. Unfortunately, this didn’t go as smoothly, and it really wasn’t making much of an impact. So, an important business decision had to be made. Do we spend the money, fix the flaws, and find new ways to encourage new people to play, or do we cut our losses and retool our site with a different function for our readers while maintaining the function of creating an ad-based revenue source? I’m sure that in the end, it was a tough decision to make by those who were directly involved, and probably a much easier one to make for those who count the change and keep the budget.

Now, as a final note, you do have to question the decision to bring on a quiz model. If the picture of the school aged children that was used to promote the leagues, which never worked, is any indication to the direction of this website, it shows the desire to hit a larger different spectrum of the readers. You see, it appears as if a majority of those 6.7 million readers were school aged (9-15 years), and not 16-60 year olds who play PPX. The only question is, do they have the money to buy things, and if not do they have the ability to convince those who do have the money to buy things?

Read More!