Your data on Google
Making search more relevant
Making our search algorithm better
You’re about to go out for dinner and have to choose between two identical-looking restaurants on the same street. One is full, the other empty. You’ll probably choose the bustling restaurant, and when it comes to Google Search, we often use the same principle when selecting which results to give you. To bring you the most relevant results for each search query, we look at the usage patterns of millions of people using Google every day.
It is by analyzing these search patterns via our logs data that our engineers are able to improve the search algorithms that determine the order in which our search results appear. If our engineers can see that people are consistently clicking on the top result for any given query, they know they are doing something right. If people are hitting “next page” or typing in another query, they know they’re not delivering the results that people are looking for, and can then take action to try and improve the search algorithms.
Personalization
Most people find what they are looking for on Google most of the time. But we know that our search results don’t always give you what you were looking for, and that is because there is inevitably an element of guesswork involved, especially when the search term is ambiguous. If you search for [Paris Hilton] do you want a hotel in the French capital or celebrity gossip? If you type in [Chelsea] are you looking for information about the football club or about different neighborhoods in London or New York?
An algorithm cannot
provide all these answers so it helps if we know a little bit about you. If we have some
extra clues to your preferences we will have a much better chance of delivering you the
results you’re looking for on the first try.
One of these clues is your location. If you live in San Francisco, you really don’t need a plumber from New York (imagine the call-out charge for a start!). That’s why we often give you results based on where you are. One of the ways we can do this is by using your IP address, a numerical code which tells us roughly where your computer is, to the city or regional level, so your search results won’t take you halfway across the world to find a plumber.
Another clue is in your
previous searches. If you search for [beetle] do you want to drive a car or find out
about an insect? It’s hard to tell, so we have to do a little bit more than look at the
general search patterns of other people to get you the results you want. If you’ve
searched for cars before, we can guess you probably meant the car and not the insect.
Normally we do this only within a single session, e.g. within the most recent couple of hours that you’ve been using Google, so if you searched for cars yesterday and then beetles today we may not be able to guess what you mean. However, if you sign in to a Google Account and use Web History we can offer you increased personalization and more control over how your searches are personalized.
It’s good to know that data can make Google search results more relevant. Read the next topic: How we use data to make our services more secure