Visual Search: Search Beyond Words

January 23, 2020

Bailey Chancellor

Share this:

Visual search beyond words

In the English language, there’s around 180,000 words and we only use about 3,000 – 5,000 of them. Think about that a bit. We only use a few thousand words but there are infinite objects and billions of descriptors out there in the universe. Applying this to search, it becomes clear that people, via voice and text, only search for a tiny set of topics. But what about the topics we cannot verbalize?

Have you ever wanted to turn to Google to search for something but couldn’t put your search into words? Maybe you see an eye-catching pair of sneakers and want to know what brand they are or where the person purchased them. Text-based search isn’t the best solution in this case. For this situation, visual search is the answer.

Brian Rakowski from Google explained it best: “Being able to search the world around you is the next logical step.” He couldn’t have been more right.

What Is Visual Search?

With visual search, the image itself is the query. Thanks to artificial intelligence (AI) and machine learning, search engines (like Google) are able to understand the content within an image and return relevant results back to the searcher.

It is a great tool for answering questions that are difficult to verbalize or search for items that you have little information about. To quote Purna Virji from Bing: “(Visual search is) the ideal and only way of searching for something when we don’t even know how to name it.” Searching this way returns answers to questions that are not necessarily simple to verbalize:

  • Who is this?
  • What is this?
  • What goes with this?
  • What does this mean?
  • Where can I buy this?

Visual search is perfect for shoppers who face two common dilemmas:

  • “I don’t know what I want, but I’ll know when I see it”
  • “I know what I want, but I don’t know what it’s called.”

What’s The Difference Between Visual Search and Image Search?

Although they may sound the same, image search is actually quite different than visual search. Image search begins with a typed query that leads a user to a search engine results page showing a collection of images that match the query.

Shopping via image search
Image Search

Visual search is when the image itself is the query, rather than a keyword. A user takes a photo, uploads it to a visual search supported search engine (more about that below), and gets a list of results that closely match the content in that image.

Shopping via visual search
Visual Search

Visual Search Engines

Most of the major search engines (and many big brands) now offer visual search features.

Visual search engines
Visual search providers

You may be most familiar with Google Lens, Amazon, and Pinterest visual search features as these are the most widely used today. But they all have one thing in common: visual search engines transform smart phone cameras into a visual browser for learning about the world around you.

Examples of Visual Search

Gather Information About the World Around You

Interested in learning about that historic building? You can now take a photo of it and Google will return information about the building – the name, address, and history, perhaps. Or, maybe you’re looking at a menu at a restaurant and wonder “what is ratatouille?”. Your visual search engine can pull text from the image and surface more information about that particular word you are questioning. Both incredibly valuable features for travelers and explorers alike.

What is this?
What does this word mean?

Discover How To Locate and Purchase Items You See in Your Everyday Life

Like your friend’s outfit? Tap the camera icon in the Pinterest search bar and take a picture. Pinterest will return results that closely match the image, suggesting the brand, name, and price of the item searched for. Similarly, you can take a picture of a room and Pinterest will tell you where to purchase similar pieces and can even offer direct links to purchase.

Visual search is a particularly great tool for shoppers who are looking for a specific item. This is because it reduces the number of steps the user would otherwise have to go through – they simply upload their image and get matches to the desired product immediately, sometimes eliminating the browsing step entirely. In fact, studies show that visual search leads to check out twice as quickly as text-based search.

Who makes this shirt?
Where can I buy this?

Learn About Local Businesses and Use Technology to Determine Where You’ll Go Next

Are you ever out and about and come across a new restaurant? Take a photo of the storefront and your visual search engine can return a link to the restaurant’s website, menu, and reviews information.

Let Visual Search Be Your Personal Assistant

Visual search can also act as a personal assistant. Take a photo of a business card to gather phone numbers for your contacts or take a photo of your meeting notes to copy and paste text into a digital file.

Copy text with visual search
Add this phone number to my contacts

Google even uses visual search technology to offer accessibility to those with visual impairments. Google Lookout was designed to help the blind and visually impaired navigate their surroundings. The app uses a mobile device’s camera to gives users audio cues about their environment; for example, telling them if there’s a chair or a cute dog blocking the way.

Google Lookout visual search

What Marketers Should Know

Your peers and competitors aren’t thinking about it. A study cited by Search Engine Journal, which includes responses from more than 1,000 creative and digital marketing decision-makers on their top marketing priorities for 2019-2020 found that just 35% say they plan to optimize for visual search. So, if you’re looking for an opportunity to stay ahead of competitors, optimizing for visual search is a great way to stand out.

It’s expected to make a significant impact on sales. Gartner research predicts that by 2021, early adopter brands who optimize and design for visual search will increase digital commerce revenue by as much as 30%.

People want to search via image. Recent data finds that 60% of millennials are interested in the ability to search by image as a part of their digital shopping experience, and 74% of consumers say that text-based keyword searches are inefficient in helping them find the right product online (Irvine).

For retailers, it collapses the purchase journey.  As mentioned above, visual search can reduce the number of steps the user takes to make a purchase, reducing sales friction and drastically reducing (or even eliminating) their time in the consideration process.

Pay attention to Pinterest. Pinterest can now target over 5,000 categories through visual search advertising on Pinterest Lens. The company also reports that their users carried out more than 600 million combined monthly searches using their Lens technology in February 2018.

It’s more important for certain industries. While there is potential for visual search in all industries, marketers working in industries where visuals sell, have the greatest opportunity:

  • Fashion
  • Beauty
  • Home Décor
  • Food
  • Travel
  • Retail

Challenges With Visual Search

There inevitably challenges with all new technology. In the case of visual search:

  • Reporting and attribution is not yet possible, meaning you cannot segment search queries to understand if they originated from image or text.
  • It is difficult to monetize visual search in the same way as text search, as targeting mechanisms cannot focus directly on keywords or topics.
  • It is not immediately clear what a user’s intent is when taking a picture with visual search is, so it can be difficult to deliver on their exact intent.

Get Started With Visual Search

So, how can you begin to set your brand up for success with visual search? First, be sure to gauge your industry’s or niche’s potential for visual search. It’s important to evaluate whether or not images are essential to the business and key in driving audiences to convert. Additionally, if the target audience uses visual-centric and product-centric platforms like Pinterest, Instagram and Amazon to find their products, the next step is to audit image assets. Images should be unique (not stock) and free of background clutter. It’s also a best practice to offer multiple photos of your products from different angles or perspectives to give consumers the full picture of the product they’re interested in.

Finally, ensure search engines can understand your image content. If images are missing descriptive alt text and captions, an image sitemap, or ImageObject schema markup, Google will be more likely to misinterpret your product . This, in turn, can negatively affect how often your products are shown, and how accurately they are shown to the correct audience.

Photo 1 Source: Unsplash

Share this article

Share this: