Note from Denis: this blog post ended up being a lot longer than I expected. The first seven paragraphs focus more on the “challenges” these sites are facing, but I encourage readers to check out the bottom 3 or 4 about what Marketocracy has been up to this year- they could be hitting a very interesting tipping point in their model, in a good way.

In February of 2007 I wrote a post entitled The Perils and Promise of Marketocracy, a company that was taking a very wikinomics-type approach to finding the “best investors in the world by tracking, analyzing and evaluating their performance in managing virtual portfolios on the site.” While I was, and continue to be, quite intrigued by the possibilities of their model, the bulk of the post was fairly critical. The reason for this was a clear disconnect between the performance of the virtual m100 index, which was based on the stock picks of the best investors in the Marketocracy community, and the Marketocracy Masters 100, which was the real mutual fund supposedly based on the m100.

The main problem I had is pretty easy to demonstrate For example, if you looked at the performance of the virtual fund (as posted on the site), it seemed to perform remarkably well from (for example) January ’03 to January ’05- a net gain in the 70 % range. If you looked at the performance of the real fund over the same time period, it went from $9.93 on January 3rd 2003 to $10.86 on January 7th 2005. Suffice to say, that’s a lot less than 70%- which doesn’t make a lot of sense if one is based on the other. You can read the post for more details, which was an update on the original post (after I interviewed the CEO).

What inspired me to revisit the company was a TechCrunch story about Cakedex. The idea behind Cakedex is similar to the original idea behind Marketocracy – create a mutual fund based on the top performing investors on a their social investing platform. What differentiates it is that on Marketocracy, individual performance is based on their success running a fictional $1 Million portfolio (with transaction and management costs applied to mimic reality) – while the Cakedex offering is based on tracking the success of the actual brokerage accounts. This seems appealing – measuring people on how they actually invest their money. Notably, Morningstar research has indicated about 1/2 of mutual fund managers have no money in the funds they manage. Scary thought.

What I don’t like about Cakedex is a repeat of a disturbing trend that seems to be emerging among such sites. The graphic in the TechCrunch article represents the fictional performance of the Cakedex index since 2003- and it blows away the Nasdaq, Dow, S&P 500 (the same graphic can be found at their site, in addition to an explanation of how it works).  The actual Cakedex index went live in May of 2008. Since then, here is the performance versus the S&P:

See what I mean about a disturbing trend? Fictionally, these funds seem to do well – but make them real and something falls apart. I’m going to be watching this one closely to see what happens.

Which takes me back to Marketocracy. I’m still not sure what to make of their performance, and how they promote it – you’ll note that not only is the m100 virtual portfolio still presented (vs. the S&P) on their about page, rather than the actual portfolio performance, the chart hasn’t been updated since November of 2006. The “how have we done” part talks about performance in the first 11 quarters of inception (which was in 2001), and nothing more recent. That’s kind of weird. It’s also notable that their real fund seems to have been crushed over the last year or so – from a peak of $17 in July of 2007 to $11.72 yesterday, and is down to net assets of about $33 Million.

However, they do appear to show the actual performance of the fund here, and it appears they are handily thumping the S&P index. I must mention, however, that most of the finance sites classify the fund (based on holdings) as more of a mid-cap blend. Using this comparable, the performance gap is much lower – but they do still outperform this comparable last year, and over the last 3 years, but not the last 5. Could this be a sign that they’re really figuring it out?

We’ll see – but what I really want to (belatedly) focus on is some really cool elements evolving on their site. One example is this Barron’s article about Trader Mark, who’s garned $3M in pledges so far for the creation of his own mutual fund – and the reason he seems to be getting them is based on the performance of his Marketocracy fund. This article also provides more details on the interworkings of Marketocracy, and how their internal intelligence is being utilized.

What’s notable here is that Marketocracy how has 30,000 people with trackr ecords of over five years. 500 “research contracts” have been granted to participants. Moreover,

Since January, the virtual portfolios of 12 of the best-performing virtual managers have been made available to investors willing to plunk down $50,000 in a separately managed account format. So far, $5 million has been allocated to the top managers, and Mr. Kam said he expects that number to double by the end of the month.

This is a fascinating development from many perspectives. That 5 year track record, and the transparency into it, really highlights the fact that most investors have NO insight into the real track record of who’s investing their money – I agree with the article that a 5-year virtual track record may very well trump a newly minted MBA. This could be a situation where that m100 / Masters portfolio problem does not matter so much anymore – could the real value be in bubbling those star investors to the top and letting them do their own thing? If so, it’s not something Marketocracy just stumbled into – when we interviewed them way back when, Kevin Kam talked at length about their hopes in doing just that.

It’s also interesting because Ken Kam is showing a commitment to the long haul that is often lacking in new start-ups – particularly those tied to wikinomics principles. As he says in the article, there are no short-cuts: if you want a 5-year track record, it will take 5 years. In my mind, this not only applies to the participants, but the company he is trying to create – one where the intelligence being collected is evolving into the core asset. It also gives Kevin Kam something any investor would love – barriers to entry. New sites like this may pop up all the time, but you can’t suddenly create all that history.

Unless, of course, something like Cakedex could instantly tap into the multi-year histories tied to online brokerage accounts…