Danny Sullivan Discusses How Google Autocomplete Predictions Are Generated
Google’s Search Liaison Danny Sullivan recently wrote a blog post that gives small business SEO professionals some insights on how autocomplete predictions work. Google’s autocomplete predicts the next word to be typed by a user when using the search engine for queries. Almost all small business SEO consultants know how autocomplete predictions work as Google already explained it before, but Sullivan’s blog post focuses on how exactly it is generated.
Google designed autocomplete predictions to help users find the answers that they need immediately, allowing them to save more time. Generally, these predictions are generated based on real queries that users previously typed in Google’s search bar. The system focuses on four factors when generating predictions: the most popular queries, the user’s language and location, as well as freshness.
Trending queries are the very first factor that Google looks at, but this is not to say that it will only show the most common predictions to the user who typed in the question.
The system also considers the user’s language and location, so it can show predictions that lead to more relevant content for each searcher. For instance, if one is to type the words “driving test” on Google’s search bar, there may be predictions containing the user’s location, such as “driving test UK”.
On the other hand, longer queries are dealt with differently because they are much less common. If a user types in a long question, Google will not predict the whole query but instead shifts its focus on predicting certain parts of the query.
For instance, if a user types in “the name of the thing at the front” on the search bar, the system might have difficulties predicting the next words. Instead, they will focus on the end of the phrase and recommend words that were commonly typed in by other users.
In this case, Google receives a lot of questions for queries like “the front of a boat” or “the front of a car”. Using this information, the search engine can then predict the next words that the user is going to type.
Lastly, when generating predictions, Google tends to prioritise fresh or trending topics. If people are becoming more interested in a particular topic, the autocomplete prediction will show users words relevant to that trend.
However, Sullivan emphasises that the autocomplete feature must not be compared to Google Trends.
Most people have always thought that this feature is convenient when typing in queries; however, it does not always work out that way. Sullivan explains that there are some scenarios where Google does not activate the autocomplete feature for the user’s well-being.
For example, Google will prevent showing its users relevant suggested words if the user’s queries are shocking, unexpected, or lead to unreliable content. It works in two ways: automatically and manually.
Google will automatically prevent predictions if they are policy-violating or unhelpful, such as phrases that are sexually-explicit, dangerous, disparaging, hateful, or violent.
If the automated system is not able to catch these predictions, they will ask their enforcement team to remove such words to comply with the policies.
Nevertheless, Sullivan concludes that users can still type in any questions that they want because although the autocomplete does not show predictions, Google will not prevent search results from showing.
Google is the number one search engine today. Whether you own a well-established or a small business, SEO is a must if you want to build your brand reputation, sell more products and services, and boost sales.
Here at Position1SEO, we will help you achieve these goals. Our team consist of experts who have garnered years of experience in the SEO industry, and no project is too big or too small for us. Whatever industry you’re in, we guarantee you a place on Google’s first search results page!