THE RKGBLOG

Engineering Triumphant: Google’s Building 43

Interesting article in today’s New York Times on Google’s organic ranking process.

The article offers quotes from Google Fellow Amit Singhal, who leads the team in Building 43; Udi Manber, who heads up search quality; Matt Cutts (of course); John Battelle (of course).

The take-away for marketers is obvious, and well summarized by Seth Godin:

Betting against Building 43 doesn’t seem nearly as smart as betting on them

In other words:

To win at SEO, spend your time creating excellent content. Content worthy of links, and content worthy of ranking. Black-hat trickery can yield short-term results, but the benefits will be short-lived.

Beyond the obvious SEO “white-hat will trump black-hat” take-away, I get another take-away from the piece in the Times: that Google’s success is the story of Engineering Triumphant.

That is, Google is not about Science or Math or Theory. Google is about Engineering. Engineering is about getting things done. Quickly. Cost-effectively. Engineering is about making things work, within design limits. Engineering is about pragmatism, rather than elegance as an end unto itself.

As the article describes, there is not a single “God” algorithm running across Google’s huge server farms. Rather, there are a large collection of specific little algorithms, kicking in in certain circumstances, bound together with “messy” weights and heuristics. Algorithms specific to local businesses. Algorithms to handle names of famous people. To handle names of non-famous people. For video. For images. And so on, and so on. Google is running a “society of algorithms,” working together to make sense of our queries and of the ever-changing web.

This “society” approach is the same design proposed by Marvin Minsky to explain how the human mind emerges from the human brain (see Minsky’s 1988 book, Society of Mind, highly recommended).

Google needs “messy” algorithms because we humans are messy. We think and speak in natural languages, chock full of special cases, inconsistencies, and nuance. Our desires are complicated, and often irrational. Building a machine to win at chess is easy, because the problem domain is “clean”, eg mathematically tight. Building a machine to have a conversation is very hard.

Pulling the thread back to online marketing: retailing, both online and off, is people trading money for stuff with other people, hopefully with both sides happier after the trade. Retail is a human endeavor, so retail is inherently “messy.” Retail is full of special cases, inconsistencies, rationality, irrationality, and desire.

Certainly, online marketers must employ excellent algorithms just to stay in the game (be that game search, analytics, feeds, etc). But winning at online marketing requires more than just strong technology: it requires smart people to guiding that technology. Just as Google’s success depends on smart engineers steering powerful machines, winning at online marketing today requires smart analysts. Marketers who can also think like engineers, and like psychologists. Who are hands-on. Who understand shopping and shoppers. Marketers who live in the data, tweaking the weights, and responding to the special cases.

Technorati Tags: , , , , , , , , , ,

  • Alan Rimm-Kaufman
    Alan Rimm-Kaufman founded the Rimm-Kaufman Group...
  • Comments are closed.