Given the new developments in network theory and complexity analysis of large datasets, does the traditional analysis of financial data matter. When we look at data collected at companies it represents a action that has taken place in the past and now we are forcasting that something may or may not occur because of it. Now if we were to analyze trends as they occur and work off of that data it would be more relevant. For instance traditionally someone buys something and the transaction is recorded and after a somewhat extensive financial analysis it turns up in a financial statement. If we use network analysis and complexity theroy to analyze consumer data as the action occurs we can predict behaviour real time and be more accurate for consumer behavour for instance. It is important to note that in order to do this you need a active data network that people use on a regular basis. This is what a few companies have, most notably google (see last post) done. Then all you need (no easy feat) is the computing power and the application of some high level mathematics and you are there. You can analysis on real time data not after the transaction data. Your thoughts???
4 Comments
Comments RSS TrackBack Identifier URI
Leave a comment
[...] Does Traditional Finance really Matter?For instance traditionally someone buys something and the transaction is recorded and after a somewhat extensive financial analysis it turns up in a financial statement. If we use network analysis and complexity theroy to analyze … [...]
The question would be- what do we gain by doing this? If we’re looking at it as an investor, quick reaction to trends might help the day trader, but doesn’t do much to really determine a company’s future direction- for that, we need a blend of past financal performance data (to gauge reliability and stability), plus a small dose of what’s going on now, to gauge corporate agility. Even so, rapid response to an emerging trend is only useful in some industries- it’s all but impossible in the automotive industry, for example, with long engineering build cycles.
As a corporation, such rapid data could be useful in improving agility, depending on your particular market segment. Industries with long engineering/regularoty cycles, or companies that have given in to international outsourcing, could find a lot more money is wasted on the analysis than could be gained, putting them in a ‘one step behind’ mentality. Such companies are better off being trend setters and innovators, with the rapid data used only to gauge effectiveness- and that is possibly better analyized at a broader scale.
Just my 0.02 cents.
I agree if it was only a instant trend that one was analyzing, but given network and graph theory and the analysis of a scale free graph that is not a point in time as much as it is a analysis of a dynamic network of millions upon millions of active nodes that effect each other and the networks that link to it. The mathematics allow us to formulate very high probabilities of certain actions of happening. IN my first post I spoke of a book “the Structure and Dynamics of Networks” that works through the math that back up network science.
Consider me a skeptic. I have done some of the reading on this- and I liken it to psychohistory (reference Asimov). In my opinion, predictive trend analysis at this level cannot be effectively used by business- it gives a false sense of accuracy in attempting to predict human behavior, and it has not been proven in those businesses which cannot instantly react to the predictive data.
Instead, what’s being attempted here is long term business planning based on prior trends- at a high level, this has been done (albeit without the math) since business itself was invented. The question was, does traditional finance matter? I say yes, most definitely, because even the highest levels of analysis can still go horribly wrong- most likely due to missing data points. Using the math to back up the thought process makes more sense- but it’s an awfully expensive way of saying ‘I told you so’. Most of the theory papers here rely upon the assumption that a complete, or mostly complete, network can be formulated- which isn’t really possible in today’s world, and thus leads me to the conclusion that traditional finance still trumps network theory.