Chapter 5: Things Google is Bad at Finding

Share

Google can find many things on the Internet, and answer all sorts of queries. Google is specially good with queries where there is one obvious result for the query, so that one is always the most relevant, and it makes that one the first result. On the other hand, Google is terrible with queries that have multiple, valid results.

This type of query is very common, and in fact one of the things Google is very useful for, despite not being designed to do it at all. Many people who use Google will complain about how bad its results are when they encounter one of the queries we'll learn about. Personally, I don't really blame Google for being bad at them. It's just not a problem can be easily solved. In particular, other search engines will often have worse results for the same queries, because it's just not a problem a search algorithm can solve.

Essentially, just because the results are displayed in a list that doesn't mean Google is good with lists of results.

Searching by Category

Google can not show you a list of results that is a list of webpages that belong to a given category. For example, as mentioned in the previous chapter, Google has access to the Internet Movie Database, which catalogues movie information, so every movie has its own webpage. These webpages also contain in their texts the movie category, such as "comedy." With this in mind, if we search for:

comedy movies

We might expect Google to give us a list of webpages where each webpage is a separate movie. But that doesn't work.

The list of results we have is a rank. The first result is ranked the most relevant. How would we rank the comedy movies? By alphabetical order? Is the first one that made most money? The best rated movie according to user reviews or critics? Or perhaps the funniest movie according to a computer algorithm?

No matter how you rank the items, it would be confusing and incomplete. If they're ranked by alphabetical order, and you have ALL movie webpages on the Internet, the Z movies would be all the way to the bottom. Perhaps you can reverse the order, but then what about the letters in the middle of the alphabet? In any case, why can't you order them by other criteria, like by reviews?

And then Google would have to have something like this for every single thing it could categorize. Books, songs, local businesses, etc. And that's not even considering what would happen if we searched for something that's not a product, such as:

animals with white fur

The same thing happens. Google won't give you a list of Wikipedia articles, for example, for each single animal that has white fur. That's just not what its list of results is for.

Instead, what Google will do is search the web for webpages that contain these terms. Owners of websites know that people search for these things, so they also make webpages with titles that answer these queries.

For example, if you search for:

words with 5 letters that begin with t

You will find several webpages on the Internet that exist only to fulfill this query. Nobody would be in a website and click an article about "words with 5 letters that begin with T" just because they're curious. These webpages are made for Google, to attract visitors that come from Google.

This sort of website tends to have lots of ads, which you may find very annoying, but the owner of the website is not worried because they do not have any real users. Nobody even remembers the name of the website. They just come from Google, get the information they need on that webpage, and go away.

It's worth noting that Google, the company, is aware of these shortcomings, so they try to address it by showing an actual list of items above the search results with some queries. These list may even have photos, in the case of animals, or, in the case of movies, Google shows the cover of the movie. Clicking on these items generally gives you a link to a webpage that talks about them.

However, the lists are incomplete. Google will never show you a list of ALL the comedy movies, only a partial list. It can even find characters of a series sometimes, but the information will be incomplete, so most of the time it's a better idea to just go to a search result that contains an entire list.

Queries with Websites Competing to be The First Result

Google is very good with queries where your intent determines which result you will click on. For example, if you search for the name of something because you want its Wikipedia article, the order of the results don't matter because your mind is already set on which result you will click. The same thing happens if you want the official website of something and its Wikipedia article shows first. Google is bad when your intent can be satisfied by multiple results, so you will click the first one that shows up, and the websites are in a competition to be that first result that gets clicked on.

That's because in this competition the judge is Google. Google decides which one is the top result, and the criteria it uses to rank results may not be the criteria you wish it used, and website owners may try to game Google's algorithms to rank higher than other results you would consider better.

This can happen with a vast amount of queries. It happens when you search for products, for services, for local businesses, for tutorials, for recipes, for news headlines, for anything. You could say that every new website is in a competition with every established website for that top spot, even if it's just pure information.

For example, we know that IMDB is a movie database, so if we search for a movie that we want information about, we immediately click that one. But what if there was a better movie database on the Internet? And we just didn't know about it because IMDB always appears first?

A more practical example, say we search for:

how to make an apple pie

Imagine how many websites in the world contain the answer for this query. Everyone who knows how to make an apple pie can create a free recipe blog in some blogging platform and write an article about how to make an apple. As far as we know, any one of these articles is as relevant as any other one, so how do we rank them?

Daft Punk has an official website. Hollow Knight is officially sold on Steam. There is no official guide about making apple pies. It's a competition for the top spot where the winner isn't necessarily the tastiest apple pie, nor the easiest to make, nor the cheapest to make, nor the best-written recipe, nor the most complete article, nor anything like that. Because Google isn't capable of judging webpages in such way.

Google's algorithms will just try to figure out which one is the most relevant. Which one appears to have the most authority and expertise based on algorithms nobody really understands. And one of them will be the winner at the top and the rest will be the losers.

But don't be sad for them. The owner of these websites will make several other articles about several other things, like orange pies, sweet apples, and so on. And sometimes they will be at the top, and sometimes they won't be at the top. You win some, you lose some.

What's important is that it's hard for Google to judge the quality of these articles. If you search for:

how to draw a heart in Inkscape

And the first result seems to be a very bad tutorial, and you had to go to the 5th result to find a tutorial that's actually well-written, it's because Google has real no way to really judge the quality of the results.

This is particularly important when searching for news. Some of the news you can find on Google may be completely fake, because Google has no way to fact-check every single statement in every single news article on the planet. That's just not Google's job.

And this doesn't even have anything to do with something nefarious like writing fake articles to push an agenda. There are satirical websites on the web that look just like news websites—such as The Onion, The Beaverton, and The Hard Times—and Google may show them to you in the results as if they were just a normal news website if you are unfortunate enough to type something that matches one of their fake articles' headlines.

In conclusion, do not expect Google to be capable of judging the quality of the accuracy of the results it lists at the top.

That said, Google, the company, is aware of these shortcomings as well, and they have taken steps to avoid showing completely nonsensical answers to questions and prefer serious webpages from serious websites, but don't expect it to be perfect or to solve all the problems listed above.

Navigation

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *