Search For...

Thinking

Natural Language is Changing the Face of Search

Google’s algorithm changes around 500 times a year. This fact alone makes it increasingly more difficult for search marketers to keep up with the ever-shifting goalposts. With big overhauls like Penguin and Panda single-handedly changing the search landscape in recent years, we are constantly looking at new ways to stay on point and on top of search trends. Something we’ll be keeping a close eye on in the future is natural language in search.

What is Natural Language in Search?

Natural language is exactly what it says on the tin – it’s the way we naturally communicate with other humans. Logging onto Google and typing in ‘best hotels in Bali’ is just the way we’ve become accustomed to searching for information by using keywords with little attention paid to turning these strings of words into sentences. We’ve all got a Siri or a Cortana living in our pocket, but how many of us make use them and understand their capabilities? It’s said that by 2020, 50% of all search will come from voice, helping us to get the information we need and quicker.  Surely the more adept technology can become at recognising natural language in search, the less the goalposts will move? By this point, search engines will have mastered the way we produce language as humans, delivering the right search result first time, every time. Strictly speaking, this isn’t true or even possible.

Cortana Phone | Creative Thinking | Kingsland Linassi

What’s the Beef?

Words like beef are the exact problem. Technology has the ability to understand the syntax of a language (the way words are arranged in a sentence) as computers are pretty good when it comes to reading formulas, and this won’t differ from person to person. The problem therefore lies in that everyone uses language differently, whether it’s their pronunciation or choices in vocabulary. This makes it very difficult to create one catch-all system that is capable of understanding language universally, or even just one language like English. Every part of the linguist within me wants to say that the search field won’t ever be able to keep up. And that’s simply because language never remains the same. Of course, elements like syntax and grammar will remain consistent, but with language diffusion happening over time, our vocabulary and the way we pronounce words will continue to change on a day to day basis. This makes it difficult for search engines to anticipate change.

Accents, dialects, social status, profession… the list goes on. These are all things that affect the way we use language, as we look to accommodate those around us. Search engines will have their work cut out for them in this respect as they try to decode exactly what is being asked. Would a world where voice search dominates mean search marketers start anticipating these problems by turning to negative keywords in the case of accents causing misunderstandings? What happens when people searching for a ‘pool’ of the swimming variety are met with the location of the nearest ‘snooker/pool hall’, a synopsis of the movie ‘Paul’ or a Googlecard linking to ‘Pull & Bear’? Homophones are an issue. What if we begin searching for words that we’re familiar with, but are yet to diffuse outwards of our home town or even our personal social group (like beef)? Although algorithms look to learn from any previous questions asked as much as possible, natural language in search is clearly not without its problems.

Siri iPhone | Creative Thinking | Kingsland Linassi

It’s Not All Doom and Gloom…

The ultimate goal of mastering natural language in search is to reduce the number of steps between an initial search and the end result to as few as possible. In this sense, we are doing quite well. Take your iPhone for example – an ambient search device that’s always listening. This smart alec of a communication device is able to cut out several stages of search by retaining information and getting to know the user. Much of the future of search depends on the unique ability of apps to retain and predict the attributes deemed important by the user. Faceted search is moving towards dynamic queries being answered with tasks or a card in search engines. For example, if you search for ‘Manchester United’, the top result will likely be the latest live score or result shown in a card. The search engine thinks it knows what you want, and thanks to Google intelligently tapping into your info and patterns, it’s more than likely right. Commerce search has come on leaps and bounds in recent years. Just look at the Amazon Echo – it’s at the forefront of natural search and having an IPA integrated into apps is certainly a big part of the future.

Amazon Echo | Creative Thinking | Kingsland Linassi

The Echo can condense search down from an initial search > clicking through to a site > filtering a search, to the user simply making a request and it being completed through the retention of information. For example, saying to an IPA, ‘Order me another packet of guitar strings’ will likely be met with no more than ‘when do you need this by?’ as it knows your address, the type of guitar you play and your bank details’. Collapsing this sales funnel into just one simple step is a huge step in the right direction for IPAs and the future of natural language in search. That said, as long as language continues to change, (which it will forever until the end of time), as search marketers, we’ve got our work cut out for us….Maybe those algorithm changes don’t seem so bad after all?

Kerri - Kingsland Linassi

Written Bykerri

Published InThinking

View All Thinking