No time right now?
Google presents the advantages of its new Conversational Search: With several search queries in a row, the context determines the search results.
Google is constantly working to improve its search results and make them more relevant. Updates like Bert should help, for example, to better interpret the context of words in longer search queries and thus to understand the actual search intention. With the Conversational Search, Google is now expanding this understanding of context to include several consecutive search queries.
Conversational Search: Google brings more context to search results
Google explains the improvements with an example: Imagine you are talking to one person and asking several questions about a topic in the course of the conversation. Wouldn’t it be strange if your counterpart forgot what it was all about after every question? Just as we humans answer several consecutive questions depending on the context, the Google search will also be able to do it in the future.
In future, the results of search queries should also take into account previous searches and their results – just like in a conversation. This is particularly helpful for search queries that allow different interpretations and for which search results were found in the past that were less relevant to the actual search intention.
Google makes the improvements clear using the example of Thanksgiving. If someone has searched for turkey recipes and then enters the search term “carving”, Google has different interpretations. For example, it can be carving, chiselling or cutting – or carving a turkey. Previously, Google used the most common interpretation when creating search results. Thanks to the conversational search, Google understands from the previous search that “carving” is most likely about the preparation and serving of turkey.
Discover topics more easily
The new language comprehension functions should also mean that Google recognizes when certain topics are searched. This is intended to display relevant information on related topics. For example, if you search for the Christmas film “The Polar Express” and then for “A Christmas Story”, Google recognizes that searchers are likely to be interested in family-friendly films for the holidays, and displays a list of similar films to match.
Thanks to contextual learning, Google also wants to help ask more relevant follow-up questions. In the “Similar Questions” section, Google has been showing related searches for specific questions for some time. In future, the further questions presented here are to be coordinated even better with previous searches. To stay with the turkey example: If, following the turkey research, someone searches for how to fold a serviette into a fan, Google would suggest, for example, “How do you fold a serviette into a turkey.”