Even though Google's (NASDAQ:GOOG) (NASDAQ:GOOGL) new mobile algorithm has been getting the bulk of the attention in that aspect of the company, it's another algorithm being considered that has me concerned about how it will impact the search giant going forward, as it has the potential to be a major disaster if it ends up deciding to make the changes it is flirting with doing.
At its essence it is considering changing from a link-based, page rank algorithm to one that relies primarily on facts it determines are more accurate than others.
Here's how Google describes what it's thinking of doing in a paper it recently released:
"The quality of web sources has been traditionally evaluated using exogenous signals such as the hyperlink structure of the graph. We propose a new approach that relies on endogenous signals, namely, the correctness of factual information provided by the source. A source that has few false facts is considered to be trustworthy.
"The facts are automatically extracted from each source by information extraction methods commonly used to construct knowledge bases. We propose a way to distinguish errors made in the extraction process from factual errors in the web source per se, by using joint inference in a novel multi-layer probabilistic model.
"We call the trustworthiness score we computed Knowledge-Based Trust (KBT). On synthetic data, we show that our method can reliably compute the true trustworthiness levels of the sources. We then apply it to a database of 2.8B facts extracted from the web, and thereby estimate the trustworthiness of 119M webpages. Manual evaluation of a subset of the results confirms the effectiveness of the method."
At a cursory glance this looks appealing, but when you consider the idea that Google will determine what facts are more accurate than others, it could result in mischief because of it aligning with what would be considered by most as "consensus" facts, rather than keeping the Internet a place where competing ideas could flourish, as it has until now.
In other words, if Google doesn't consider something a fact, it would punish a site by pushing it down the search memory hole to the point it wouldn't be found by those with an interest in a specific topic and ideas that may push against the body of knowledge and "facts" considered acceptable at the time.
It also could be used to suppress contrarian viewpoints in a variety of topics, including sites like Seeking Alpha and its peers, who may want to offer a variety of ways of looking at data and drawing different conclusions from them.
Assuming this goes forward, I see that opening up a huge opportunity for its search competitors if they have the courage to position themselves as an alternative to this new initiative.
(click to enlarge) source: StockCharts.com
Gossip site feint
In the paper Google referred to a number of gossip, celebrity type of websites as an example of inaccurate information. I look at this as a feint by the company to make it more palatable to the reader in order to keep them from thinking through the ramifications of Google deciding what a real fact is and what isn't.
For example, picture the company naming a variety of major industries as those it may decide to reevaluate, industries that may be spending a lot of money advertising with it. It's not going to point out those it may consider weak on facts, even though if it rolls out this new algorithm it could have a big effect on them.
I want to use the example of gold in regard to investing, because there are a lot of contrasting viewpoints as to the value of it and its importance in protecting against inflation and the fall in value of currencies.
What if Google decides many of those the like gold aren't factual because it considers proponents of the Federal Reserve and central banking as more legitimate? It could drive many of the gold websites or articles down the search results because Google wouldn't have a clue as to how to evaluate the variables associated with that one area. We could extrapolate that out to a number of different areas of investing, and that's just one topic we're talking about.
Since most people don't consider gossip sites as much more than entertainment, the idea of it being factual isn't the impetus behind why they go to them. They go there for the purpose of gathering some information that they can then share with their friends. It's a social exercise. Google knows that, which is why I'm a little concerned about them using it as an example of being fact-challenged as it makes me think it's seriously looking to implement this as part of its search algorithm.
Model at a glance
The premise behind the proposed change is software developed by Google which will determine the number of "false facts" any specific page on the Internet has.
According to the paper, "A source that has few false facts is considered to be trustworthy." That means in Google's eyes it's more trustworthy and would deserve a better position in the search results.
Google calls the information it culls from the Internet for this purpose the "Knowledge Vault." Apparently it looks for what the Internet considers a reasonable consensus considering a specific fact, and from there makes a decision on whether or not it's accurate or not.
There also will be some human workers laboring with the software results to make the final decision in some cases, although that has to be a very limited endeavor because of the amount of information and data that would have to be sifted through.
I see that as more of a nod toward making people or companies feel better about the results if it's detrimental to the company.
At the core, it will move away from links and reputation to more of a perceived accuracy of a Website, as it relates to where it is positioned in the search results.
Keep in mind that this won't be the only part of the search result process, but it is one that if implemented would have a strong sway on where a company or idea would end up in the search results of Google.
Major challenges
To give an idea of how this will without a doubt become a disaster, we only have to look at what some consider inconclusive areas of life that we have a lot to learn about and of which there is a lot of conflicting views that could be considered factual by those adhering to them.
One of those would be in the area of health. Many people consider traditional treatments as the way to go while others prefer more of a preventative way of dealing with health issues. So if there were alternative health sites, they could be pushed way down the search results because Google may extract facts from mainstream medicine rather than alternative treatments.
Another area would be big government versus limited government. Again, we have the traditional view measured against those wanting change. If Google draws from government data and those who support larger government, it could skew the search results because it deems that more factual than those opposing it.
We could go on and on. The point is that this could and would seriously hinder contrarian Websites that offer different points of view and which may provide an answer searchers are looking for, while pointing the way for future change.
The risk
Where I see the risk for Google is in regard to its competitors. What they, for the most part, haven't been able to do on their own could be handed to them by Google if it puts this new algorithm into play.
As I see it, the reason why is there would be a huge backlash against Google based upon it becoming a corporate "big brother" concerning search.
No matter how many people it may include in the algorithm, it will be primarily a software service. There is no doubt about that whatsoever. That conclusion comes from the vast amount of information on the web, which prohibits meaningful human correction on the majority of search terms or phrases out there.
Bing or Yahoo wouldn't even have to do much in response. All they would need to do is position themselves as search companies that don't base results on perceived consensus in regard to trustworthiness of facts, but on what a person is actually trying to find.
Imagine people, even in the gossip and celebrity area, not getting results on the front page, but having to dig deeper into them in order to find their favorite topics and celebrities. It wouldn't take long for them to quickly migrate to competitors. I've personally done that for less reasons than those given in this article.
How will competitors respond?
Where this could get interesting to me, and may not have much of an impact on Google at all, would be if its competitors foolishly were to follow in the footsteps of Google and do something similar. That of course would undermine my overall thesis, which is based upon competitors not doing what Google may do.
If Google does release this as part of its search algorithm, and it carries a lot of weight in regard to results - which it appears to be geared to do - it would almost certainly result in a lot of errors and confusion.
Things that it would struggle with would be how to determine what's true from what's error, and whether it's a legitimate exercise to do so. The examples I used above show that facts are not as simple to determine as it may appear, and in many cases expertise about specific subject matter would need to be included in order to discern that.
There's also the ongoing growth of data and facts about any given area of life, and to keep up with that would be extremely difficult, especially if it's being done with software used to access an existing database. Of course it would be built to include new data, but it would almost certainly lag behind in relationship to being taking as truthful because there would be little to compare it with and it wouldn't fit the existing narrative concerning a specific topic. That suggests it could take a lot longer for it to appear in the search results where it should, and it also points to search results being slow to reflect the fast-paced knowledge world we live in.
In other words, this could bring about less accurate results in the past, which is where I see the potential for its competitors to position themselves as more accurate and reliable sources. This time users could start to believe them. That could be specifically true if it makes people angry who are searching for frivolous things like gossip, which they don't really care is accurate, but a tidbit to be shared and talked about among friends.
How to identify Google's motivation
Since there's so much potential for mischief with this type of change, such as ideology and government pressure being applied to preferred search result rankings, there's one thing I see at this time that would counter my suspicions as to the reason behind this possible change: choice.
If Google really is attempting to improve the results in their search engine, in my opinion it would offer this particular algorithm as an option for users, not one that must be incorporated into their results.
In that case it wouldn't have much risk, and for those who want this type of result could simply use that as their way of doing search. Those not wanting the change would be offered the usual way of doing search.
On the other hand, if this is implemented and it's the only option, I really do fear for Google, as I believe it would suffer a backlash against it that it hasn't had in the past.
Many of the algorithm changes of the past have been railed against, but that came from many (not all) of those that had gamed the system. But with this, it would likely bring about much more resistance, resistance that would come from a much wider swath of users than ever before.
So if this is offered as a search option, I see it as not having much of a negative impact on the company. It will just be trying to improve the results with no other motive. If it's forced on people, I see it has having other reasons for being put into service, which would primarily be the control of information and data it approves of. I say that because it has been known for years that if a Website isn't listed on the first page of Google's search results, the number of people going to a site diminishes dramatically. That even happens after the first several results, although not to the extent of being listed on the second page. I know this to be true because I've owned a number of web businesses and measured that type of result.
What if it's something competitive we haven't seen yet?
Finally, I want to look at the other side of the issue, which is it's possible to consider Google may be attempting to punish its competitors by using this new algorithm to push them further down the search results. By that I mean its many mainstream media competitors.
Google knows the power of its advertising model and what types of ads make the most money, and if it attempts to boost that by generating rankings that cater to companies generating better monetary results, it could get a strong boost to its top and bottom lines.
Although I'm not strongly convinced of this as a significant probability, it is something to consider. Whether traditional companies or new companies, they can be made or broken by the decisions Google makes. That doesn't mean a giant company would be crushed, but it could dramatically impact its earnings results if people or companies can't find them.
Conclusion
This could become the greatest debacle in the history of Google if it's put into action for the many reasons already mentioned - the chief one being the imposition of its view concerning facts and data versus what people want to determine for themselves.
That would cause deep resentment and almost assuredly an exodus from Google search to a competitor not attempting to control what people are attempting to find. If Google does do this, I think the damage would be significant, even if it were to eventually stop the service and go back to how things were being done before.
If I had a position in Google, I would consider this the biggest threat to my capital in relationship to the company I ever faced. Sure, it's not going to happen real soon, but since this is without a doubt a trial balloon, and a paper has been published outlining the proposed changes, this is beyond just the exploratory phase. It's close to having a decision being made about it. And from the paper, it appears that the decision is one that is close to being in the affirmative.
I'm not saying it will go prime time soon, but that it's closing in on a decision being made, and once that is made - assuming it's affirmative - it will only be a matter of a fairly short period of time before it's released to the public.
My concern is that this is very different from past algorithm changes and if there is ideology as a major catalyst, the search results would be considered an attempt to control the dispersion of information, which won't sit well with millions of people in the U.S.
Could this be death by suicide for Google? We'll find out soon enough once a decision is made and whether or not it's optional.