Everything You Need to Know About Visual Search
The human mind processes visuals 60,000 times faster than it does text. Whilst Google understands the human language with 95% accuracy, it’s a different story when it comes to images.
But times are changing. With Pinterest, Google and Bing currently leading the way and developing visual search technology and applications, it’s time for brands to take notice – and action.
If you’re sat there scratching your head, questioning why visual search is important and how it affects you, then you’ve come to the right place. We’re about to tell you everything you need to know about visual search: what it is, how it works, and how you can optimise your images, to give you the best possible chance of appearing on the visual search engine results pages (SERPs).
Image credit: Sparktoro
What is Visual Search?
Ben Silbermann, Cofounder and CEO of Pinterest, stated that “the future of search will be about pictures, rather than keywords”.
This is essentially visual search: the act of using an image as a query, as opposed to traditionally typing text into a search box – or asking your voice search device a question out loud.
Don’t confuse visual search with image search. Whilst image search refers to a user inputting a text query and being returned a related image, visual search is much more advanced. Search engines are actually recognising images, much like the human mind.
Why Visual Search is Important?
There’s no denying that current visual search technology is limited, but it’s continuously improving.
Pinterest, Bing and Google are leading the way in creating and implementing this technology and over time, it’s set to change the landscape of search marketing as we know it.
So, if you’re wondering why visual search is important, it’s because if we, as SEO experts, aren’t willing to understand this method of search, and optimise our images for it; then in the long term, we’ll ultimately lose out on traffic, revenue and customers. It’s as simple as that.
How Visual Search Works
Stop reading this article for a second, and look around you. What do you see? Are you sat at a desk, with a cup of coffee beside you? Or are you reading this on your phone as you’re walking through the city?
Whatever it is you see around you, your mind is instantly identifying those objects from visual cues.
If you’re wondering how visual search works, it’s in exactly the same way, except instead of a human picking up on these cues, it’s a computer.
We’ve already mentioned that Pinterest, Bing and Google are leading the way in visual search. Now, we’re going to take a look at these three companies in more depth, as we discover exactly how their technology works.
Image credit: Matech Today
In February 2018, Pinterest Lens enjoyed an impressive 600 million searches – that was after just one year of its launch.
Pinterest Lens “lets you use the camera in your Pinterest app to discover ideas inspired by objects you see out in the real world”.
For those wondering how to use visual search on Pinterest, it’s pretty easy. Simply open the camera on your app, and point the Lens at an object – for example, the denim skirt your best friend is currently wearing. Pinterest Lens will then return similar skirts for you to look at, in addition to examples of what you could wear with it (how about a crop top or a pair of chunky Fila Disruptors – in white, of course). Whilst not all listings have the option to buy, many of them will – meaning essentially, you can purchase a very similar (if not the same) version of something you were looking at in real life.
This is great news for brands, with statistics showing that 93% of active pinners use Pinterest to plan their purchases, and 87% claiming to buy something specifically because of Pinterest.
It’s not just clothing that works well with visual search technology and Pinterest Lens. For example, point your Lens at some food – let’s say an avocado – and you’ll soon see a whole list of recipes appear right in front of your eyes.
Up until now, this is one aspect where Bing has been one step ahead of Google, with Bing Visual Search (however, as we’ll soon see, Google is fast catching up…).
On the Bing app for Android and iOS, users open their camera and take a picture, and wait to have similar results returned. Like Pinterest, this can work for fashion, but it can work for other things as well.
For example, let’s say you’re on a city break in Moscow and are blown away by the beauty of St Basil’s Cathedral in Red Square, and want to know more about it. Take a picture (you can upload it on Instagram later), and Bing will return images and links that you can then click on to find out more information.
Image credit: Gizmodo
Google has been working on Google Lens, aiming to simply a user’s image search. The difference between Pinterest and Bing’s current offerings compared to Google, is that Google’s search is native to Android smartphone cameras (note that it’s not currently available for other makes), as opposed to the user needing to open up an app. By removing this one extra step, Google could well see an increase in usage and popularity.
The “Style Match” function – very similar to Pinterest Lens – allows users to upload an image, who will then be returned with related images, which contain links that users can click on to purchase the product.
One recent update of Google Lens was the inclusion of badges in visual search. This means that when users are searching images within the app, they’ll see a badge in the bottom left hand corner. The crucial badge is “product”, which means the item can be purchased.
How to Optimise Your Website for Visual Search
So, now we’re aware of the three main types of visual search technology, it’s time to look at how brands can optimise their images and websites, to give them the best chance of appearing on the visual SERPs.
Upload High Quality Images
Before we even consider the SEO aspects, you first need to ensure the pictures you’re uploading to your website are of high quality, and clearly show what you want to convey.
This is something all brands should be doing anyway, but if you’re not, then it’s time to think of the best ways to showcase your products. Let’s continue with our theme of fashion, and say you’re an online fashion retailer, and have had a new faux fur jacket come into stock. You’ll want to show off this jacket in numerous ways:
- Worn on a model from various angles – front, behind and to the side
- Close up shots – are there any specific details such as pockets or buttons you want to show? And what about the quality of the (faux) fur?
You’ll also likely have an image of the jacket laid out. For all pictures, it’s important to remember that it should be of excellent quality, and the background contrasts with the product, to make it clear to see.
That means if you’re selling a baby blue faux fur coat, you’ll want to place it on a plain black background. Alternatively, if your jacket is a deep purple, you’ll instead want to photograph it against a plain white background, to make the colour pop.
This next tip is very much common sense – and if you’re an e-commerce site, you should be doing it anyway – is to make sure that your product images are placed closely to related content on your page. This means having your product description right next to your pictures, which really shouldn’t be something we need to tell you.
Optimise Your Metadata
This is SEO at its most basic, but if your image file name is “image1.jpg”, then you’re giving search engines precisely zero information as to what your image depicts. Instead, you should name your pictures in a way that details explicitly what it is, e.g. “celine-luggage-tote.jpg”.
Don’t neglect your image alt tags. Whilst predominantly there to replace images that fail to load on a website (so users know what should have been in that space), search engines use alt tags as another way of understanding the contents of images. Take this as an opportunity to optimise your alt tag with relevant keywords – for example: “Brand new Black Céline Mini Luggage Tote”.
Go one step further and include a caption for each of your images too. These will sit below your photos and add even more context to search engines, as they further describe your image.
Image credit: Moz
Don’t Forget Structured Data
Another key way to optimise your website for visual search, is to use structured data, as that will enable search engines to further understand your images.
For this, you should use schema markup – specifically product markup – to highlight relevant information that you want search engines to take note of. This includes:
The more information you can provide, the better.
Submit an Image XML Sitemap
You would never forgo a standard XML sitemap, relying on internal linking alone to let spiders crawl your website and discover all your pages.
This presents yet another opportunity to clarify what’s depicted in your images. Provide some extra context, such as:
- Subject matter
- Priority score – this tells search engines which images are most important.
The Industries That Should Start Optimising for Visual Search
Whilst there’s no denying visual search is still in its early stages, there are clear wins that can be had – especially if brands start to optimise now, to gain an advantage over competitors.
Visual search – as the name suggests – is by nature visual, which means that visual industries have the most to gain from it.
Judging by all of the examples I’ve provided so far, I don’t need to tell you that fashion and visual search is a match made in heaven. Having the ability to snap a picture of a friend wearing a jumper you’re secretly lusting over, or a couture gown on a runway, or a pair of shoes in a shop window; and having related pictures instantaneously appear in front of you, readily available to buy – sounds amazing to me… and I’m sure it does to lots of other people too.
But it’s not just fashion brands that can win the visual search game. Beauty, home décor – and even brands operating in the food brands – should be including visual search into their SEO strategy.
Imagine seeing a picture of your favourite celebrity wearing a dark red lipstick; and taking a picture of that image, and then being shown the exact make and shade, with the ability to then buy it. Or, what if you were in a really high-end home décor store, and saw a chandelier that was thousands of pounds – you could take a picture of it and be shown lookalikes for a fraction of the price.
That’s not to say visual search is limited to these industries alone – there are loads of opportunities out there for a variety of e-commerce brands. However, it’s these types of industries that have currently identified the potential. Just look at ASOS – its Style Match tool lets you upload a picture and then search ASOS for similar clothing or accessories.
Image credit: Medium
The Future of Visual Search
TinEye may have been the first visual search application, but this technology has come a long way since then. Nevertheless, there’s an even longer way to go.
However, this technology is extremely exciting. Already, we can take a picture of a model on a runway and be shown affordable options, but take this one step further. Let’s say we’re on holiday and take a picture of a restaurant we like the look of – wouldn’t it be amazing if we were presented with the menu, opening times, and the ability to make a reservation right there and then?
Yes, visual search is in its infancy – the reason being that it’s not enough for a computer to simply identify an image. In fact, it needs to register a range of colours, sizes, shapes and patterns in the same way a human mind would.
Whilst visual search is still in its early stages, now’s the time for brands to start optimising their products so that they can gain that edge over competitors to ultimately gain more traffic and sales on their website.