Google: Making Search Better. But What Does Better Mean?

July 17, 2017

I read a darned interesting (no, not remarkable, just interesting) article called “The Google Exec in Charge of Designing Search: ‘There’s Always This Internal Debate about How Much Functionality Should We Add‘”. At first, I thought this was an Onion write up, but I was wrong. The article is a serious expression of the “real” Google. Now the “old” and now “unreal” Google is not applicable. That’s why I thought the write up was like the content I present in HonkinNews.

Here are the points I noted:

First, the write up points out that Google’s core business is its search engine. This surprised me because I thought the firm’s core business was selling ads. I know the “search” system is the honey which attracts the bees (95 percent or so in Europe, for example), but the “search” system is not about finding relevant and objective information. Sure, that happens for some queries, but for most queries, the searches are easy to cache and deliver with matching ads. Examples range from the weather to the latest in the dust ups and make ups between pop stars and starlets.

Second, the source of the write up is an “expert” in “design for search.” I am not sure what “design” means. I am old fashioned and prefer the trusty calculations of precision and recall, the stale bread of Boolean queries, and unfiltered content.

image

I prefer to do my own censoring, thank you. I noted this statement:

The whole goal is to try to organize information and deliver it to you. That’s the problem we’re trying to solve. The design has to accommodate multiple people, multiple expectations, and multiple situations. When you’re looking for whatever answer you want, how do we give you the right answer in a way that you’re like ‘oh yeah, that thing?

No, the “whole goal” consists of sub goals designed to deliver the following, based on my research for the three books in my Google Trilogy (alas, no longer in print but I can provide pre publication copies for those who want to buy a set):

  1. Minimize computational demands on the query matching system via caching frequent queries, partitioning indexes to get around the federation of disparate content like Google Scholar, videos indexed in Google Video, and the gusher of stuff emanating from Google Blogs
  2. With clicks on traditional desktops falling and small screen video queries from smart software or humans (imagine!), Google has to find a way to make ads out of everything. Thus, the need to keep revenue ticking upwards while driving costs down becomes a fairly significant sub goal. Some, like myself, say, “Hey, that’s the actual goal.” Others who enjoy watching billions flood into solving death, keeping Glass alive, and building a new puffy office part would disagree. That’s okay. I think I am right.
  3. Maintain the PR and marketing offensive that makes Google the innovation leader in finding information. The methods involve generating mumbo jumbo that disconnects precision and recall from what Google generates: Results that are often off point or some type of content marketing. (I know content marketing works because the Wall Street Journal told me it does. I assume that’s why Google pays some people to write really rah rah articles about Google. As I said in this week’s HonkinNews, “One must be able to tell the difference between a saint who helps people and a billionaire who rides flying car things.)

The write up identifies the experience “things” which Google is incorporating into its search results. Some of these are content objects like tweets. Others are pages which look like mini reports which cobble together “facts” to make it easy for a person to “know” the answer to the question he, she, or a software module had not yet asked. (Predictive results are part of the pervasive search movement in which Google wants to be a player who gets the biggest payday and the most media love.)

I noted this statement which is worthy of one of the New Age types I bumped into when I lived in Berkeley:

When asked if there are any similarities between the design for Search and the design for Google’s new offices in Mountain View and London, Ouilhet pointed to the fact that both are becoming “more open and more flexible.” He said they were also both becoming more “inclusive between people that belong to Google and people that don’t belong to Google.”

Net net: Google has yet to find Act 2 to its Yahoo/Overture/GoTo inspired business model. Setting up more VC operations, incubators, and buying companies in easy to reach places like Bengaluru, Karnataka, and smart software offices in cheery Edmonton, Alberta are not yet delivering on Act 2. If the European Union has anything to say about Google’s search business, we will have to wait for more action from that Google watcher Margrethe Vestager.

Stephen E Arnold, July 17, 2017

PS. For information about the Google Trilogy, write benkent2020 at yahoo dot com and put Google Trilogy in the Subject field.

Comments

Got something to say?





  • Archives

  • Recent Posts

  • Meta