SEARCH ENGINES REVOLUTION | From Semantic Search to Visual Search

Any search starts in Google, but is Google always the best search engine?

Without any doubt Google’s search engine has become an inseparable part of our everyday lives’ to the point in which we ask ourselves, what would we do if Saint Google didn’t exist? There are so many of us that resort to this search engine every day that even ‘to google’ is, since 2006, a verb accepted by the Merriam-Webster Collegiate Dictionary.

We have such trust in the results that this semantic search engine offers us that we constantly resort to it. Even so, “if you do not appear in Google, you do not exist” or at least that is what we think. It does not matter the “what” or “where”. Google always has an answer for you, or better said, pages and pages of answers. From there to the existence of infinity of articles and manuals with tips on “How to use Google search more effectively?”. But from here we lay out another question, is Google always the best search engine? Even for this questions there are answers on Google. But do not Google it, here you have the answer: “No, Google is not always the best option”.

How would you describe this dress?

IRINA-SHAYK_SHORT-DRESS-COLORFUL_VISUAL-SEARCH_WIDE-EYES“Short dress geometric colorful”. This would probably be the description or keyword structure that we would use to find this dress in Google if we did not have basic information like brand, model or ecommerce in which you could buy it. Though there would be more differences if we had searched famous #TheDress (viral phenomenon).

I want it! Google answers.

It is worth it to mention that Google includes in its algorithm multiple variables like geolocation, search habits, activity in other Google services (Picasa, Google Plus, YouTube…), etc. that makes the order of the results personalized for every user. So with that, there could be some differences from our vision and yours, but in no case does there exist a manipulation of the results.


“Frustration” is maybe the best word to describe the emotion we feel when we see the results. It does not matter that the results are prioritized by “Web” or “Images”. In the first 5 pages there is no dresses that look alike.

Something similar happens when semantic searches are made in an ecommerce. Finding the exact word is almost mission impossible. Though in these cases, as tiresome as it looks, it is always more effective to navigate over categories and use the different filter options. Though the time inversion invested in the search is not always compensated.

And it is that the principal handicap of the semantic search engines is “finding” the exact words to find our objective: finding the dress from the image. Why? There are as many descriptions as perceptions.

Why don’t we use Google’s image search engine?

Since we start from an image, maybe this would be the next logical step. Google Image allows us to find information about an image in the web. So that we will know in a matter of seconds which sites that image has been used. With almost complete security, we will get to the ecommerce where it is sold thanks to the ‘Shopping’ classifier.

But what would happen if instead of being a photo from Irina Shayk, it is from an anonymous person? And if we wanted to begin a search starting from a capture from our mobile devices? Google Image’s answer will always be “This image has not been found in other sizes”.

The specialization and optimization of the search engines

Since Google’s apparition in 1998, the investigation and development of the search engine has been changing the way users understand the Web’s search on the Internet by adding new technologies, services and offering advanced search possibilities. Though Google is still the principal search engine, new propositions are gaining strength like the specialized comparative type search engines (insurances, flights, hotels, financial products, etc) or the search engines for mobile devices.

We capture what we like basically to share it in our social networks in search of interaction (Likes, retweets, comments, share…). The smartphones and social networks are changing our habits marking a before and after in the way of communicating, relating, shopping (social shopping, showrooming, visual shopping) and even searching.

The search engines are in constant evolution, adapting to the user’s necessities and the new tendencies of the information society. It is here that the visual search engines have a great part to develop. Although the effectiveness of the visual search engines is now a little bit doubtful, we recall that this was the case for Google’s semantic search engine at the beginning.

The development and perfecting of the images visual recognition technology is the key to make the efficiency for visual searches better, especially for very visually strict industries like fashion. A time will come in which it would be possible to buy a look from a capture or photo in a matter of seconds and without human interaction in an ecommerce or market place. Although this is possible now, the results do not always fulfill the consumer’s expectation. Visual search has a great potential, but there is still a long way to go.

Leave a Reply

Your email address will not be published. Required fields are marked *