Indexing Matters: The Investment Sector Analysis
October 15, 2018
I read reports which explain why large monopolistic or oligopolistic companies alter the behavior of certain ecosystems. I don’t see that many because analysts are preoccupied with more practical matters; namely, their bonuses, appearances on Bloomberg TV or CNBC, and riding their hobby horses.
I read and then reread “Platform Giants and Venture Backed Startups.” The premise struck me as obvious. The whales of online are functioning like giant electromagnets. There companies pull traffic, attention, and money. At the same time, they emit beacons which are tuned to the inner ears of investors.
Looks tasty but only semi organized. And from what is this confection fabricated? Answer: Cow hooves. Intellectual Jello, lovingly crafted to delight the eye.
The squeaks of these ultra high frequency waves alert those looking for big paydays to put their money into startups which do not compete head on with the outfits operating like electromagnets.
The “Platform Giant” write up assembles observations from a report which asserts the opposite; that is, big electromagnets do not have an impact on start ups and most investors.
Put that aside.
The core of the write up makes clear that indexing and classification make a difference. The idea is that if one classifies and marshals data, the classification creates a way to look at the data, the world, and in this particular case the way investments flow or do not flow.
What goes in “Internet software” becomes the trigger for the conclusion. Invest to compete against the Google? Not a good idea.
The question becomes, “Who does the indexing, classification, ontology, and related bits of the taxonomy?”
Indexing is important. But more important is the creation of the knowledge structure and the categories which will be used to chop, slice, and organize data for analysis.
Get the knowledge structure wrong and the flawed categorization creates findings that are probably misleading at best and just off base.
Who takes the time to work out the knowledge structure before training humans and smart software to assign metadata?
The write up suggests that humans (either with agenda or without, with expertise or not, or with a wonky knowledge superstructure or not) do.
Net net: Counting is verifiable. Pegging what to count may be more like organizing cubes of a gelatin dessert.
Stephen E Arnold, October 15, 2018
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
