In the beginning there was Archie – a simple search engine by today’s standards that hosted directory listings. In those very early days, users needed to know the exact title of a site when doing an online search. Over time, search engines have evolved and gotten a whole lot more sophisticated and more useful than old Archie. And they are getting smarter all the time.
Google was the first to use inbound links and a site’s authority on a subject when determining search rankings, and with its Hummingbird algorithm, the company is now focused on a new search development known as semantic search.
Semantic search means search engines no longer simply use exact keywords to get results, but rather aim to produce the most relevant results for users based on their intent, and the meaning of their query.
Semantics is the study of the meaning behind the use of language. As we know, words don’t exist in a vacuum – when someone speaks, we don’t just hear the individual words, but whole expressions of them, along with context, emotion, intent, and other factors. We then try to figure out what the person is meaning or trying to say. As we also know, this often leads to misunderstandings. And at times, it can lead to offence being taken! On some occasions, we might then hear the speaker complaining that their words were taken ‘out of context’.
Imagine that if we humans have difficulty understanding each other at times, and that we sometimes take the meaning of words wrongly, then machines are not likely to fare much better. However, search engine programmers have been developing algorithms to help overcome these types of obstacles.
So just what is semantic search?
As the name suggests, semantic search goes quite a way beyond keyword matching. It attempts to deduce the intent of the user and the contextual meaning of the words entered, before returning information. In this way, search results should hopefully be more relevant to the user.
Besides user intent, factors used in semantic search include current trends, user location, variations of words such as plurals and tenses, synonyms, whether queries are generalised or specialised, and concept matching.
As a simple example, if a user were to enter the query “what is the weather like today?” they should get back a weather result for their current location even without mentioning it. The same type of result is likely if they enter queries about news events – in this case search engines may consider not only their location, but also current events (trends) that are happening in their region.
Synonyms are another factor. Synonyms abound in language – English in particular. For instance if a user enters “largest trees in the world”, they are likely to get results pertaining to tallest trees, biggest trees, and even reference to widest trees.
Search engines may also consider the basic concept behind the sentence entered. For example, “health problems in Australia” is likely to bring back results about local health initiatives, Aboriginal health, and disease statistics – even if the searcher did not request these directly.
How well does it work?
Semantic search is not going to be perfect. Different people can enter the same words, yet actually have different meanings or intent behind them. However, search algorithms are improving and getting more intelligent all the time.
What does semantic search mean for businesses?
It’s no longer enough to simply stuff content with specific keywords and as many links as possible. Instead, content should be of the highest quality which addresses the intent of your target market. The same rules still apply to content – produce it regularly, ensure it is relevant to customers, and promote it across platforms.