David Amerland
The Visual Web and Visual Search

Visual Search and How it Affects your Marketing in a Semantic Web

The web without search crumbles into a mass of data that makes little sense and where everything we need is hard to find. That’s why search is such a big deal. Done right it provides a service, instead of a product, that can only grow and grow.

The web itself is changing. Information is now no longer locked up in text. It also exists in videos and images. In the mobile web, images acquire tremendous power as the small-screen format is perfect for arresting our attention and prepping our minds with the right visual image.

This is why Pinterest’s Guided Search is such a big deal. On the back of the acquisition of Visual Graph, a semantic technologies company, Pinterest feels it now has the data necessary to begin delivering some very targeted results in search.

Pinterest Acquires Visual Graph 

Is Visual Search a New Idea?

Visual search of course is not new. Google has been refining image search for some time and as my test search using “breakfast” as a query proves it can deliver an impressive array of results.

Google Image Search

In addition Google has invested a considerable amount of time and money in semantic image search that allows its search engine to look at a picture and understand it the way a person would.

When it comes to semantic analysis and data Google is still ahead of the rest by a wide margin. That does not mean however that there is no merit in what Pinterest is doing. Currently with over 750 million boards and 30 billion pins, carefully curated by people who have a passion for collecting material that groups into specific subjects, Pinterest represents a treasure trove of very specialized information.

Can Pinterest’s Guided Search work?

Pinterest’s Guided Search has a couple of additional layers that Google’s Image search does not. As the image below shows not only do you find out a visual guide to your subject (in this case ‘Breakfast’) which brings up individual pins (your own or those of others) but you also have the ability to discover human curated boards (that represent semantic ontologies of a sort) or discover Pinners to follow, increasing the social connectivity.

Pinterest Guided Search

Within the closed but very rich data set that is the Pinterest visual web this works very well indeed. Does it replace Google search in any way? Well, I don’t know if we would really use Google Image Search to look for breakfast ideas but if we did the comparison is not the same.

Within Google Image Search the web is indexed and the results returned correspond to individual web pages. Within Pinterest the results are already curated collections that are conveniently grouped by subject matter, so even individual pins are thematically linked to other pins within a board. This is powerful stuff, albeit currently limited.

To understand the power of Google and the limitations of Pinterest consider how with Google Goggles Google’s Image Recognition capabilities allow you to point your mobile at a picture, barcode or QR code and Google will then surface everything it knows about it (and it even reads the text on the picture itself). This, almost seamless integration of the offline and online worlds is what Google does so well and where semantic search and semantic connections really come into their own. Pinterest has quite some way to go before it can deliver anything that’s not mapped and user-tagged in its own strictly online data set.

What Marketers Need to Know

So how does all this affect your marketing? If you are curating your own boards you need to make sure that the descriptions you use are accurate and the pictures sufficiently evocative so they can resonate with the audience. When you create your marketing materials the words to remember are “Visual Web”. If your pictures do not tell a story, stop the eye and capture attention, you’re missing out on a vital opportunity of discovery.

Visual is the breakthrough element of the semantic web. While Google is busy building semantic search and joining the online and offline worlds, semantic technologies, used in smaller data sets such as Pinterest are becoming part of the greater semantic web. All of these are porous, transparent and interlinked and they do lead to serendipitous discovery in a variety of ways. As a marketer, or a business owner you should be building your identity on the visual web with the same care and attention that you are building it in the more conventional text-driven one.

Will Pinterest rival Google search? That’s rather hard to say. The two are actually different services. Both lead to discovery but in very different ways. There is a strong element of complementarity here rather than outright competition. For now, certainly, smart marketers should be keeping an eye on both.


What You Missed:

Semantic Search and Disambiguation (How Search Gets Smarter)
Ten New Questions on Semantic Search Answered
Google+ Hangouts for Business

 

© 2017 David Amerland. All rights reserved