Artificial Intelligence and Machine Learning are increasing both in use and awareness. As computers learn more, they become faster in learning new things, detecting patterns, and identifying shapes, objects, textures, and behaviors. On our previous post we talked about how voice search is revolutionizing ecommerce, and now comes the time for visual search in ecommerce.
As you know, Content2Sell specializes in crafting compelling pieces of product content for ecommerce, and we couldn’t stay out of this matter. Because if buying products online is a little easier every day, selling them requires being aware of all the variables involved.
Machine Learning is exponential
Computers can now process and understand in real-time what’s in front of them, make suggestions, and take user behavior as proper feedback to gain accuracy. This is possible through algorithms especially programmed for a specific task: from filtering emails, to providing search results, developing their own algorithms, and even teaching other computers.
So, imagine Captcha, the “I’m not a robot” system created to tell whether a user is a human or not. Every click you make (or you don’t) tells the algorithm whether a given image is or has a certain object. Over time, the computer will have its own idea of what that object looks like, and what context it is commonly seen in.
How users interact with the results they see is key to the continuous improvement of Machine Learning.
Rather than a trending technology, Machine Learning is already within reach of many companies thanks to the faster learning of computers, the rapid spread of its use, and its decreasing cost.
Google’s leadership
As time goes on, Google keeps adding and integrating functions to become the superapp the West still doesn’t have. Moreover, in the evolution from being just a search engine, Google has progressively included new functions through integrations and acquisitions around the globe. From Youtube to Waze, Motorola, Android, Behavio, Boston Dynamics… And the list goes on.
Thus, Google has been able to develop a virtual ecosystem that integrates nearly all dimensions of our daily lives. And among them, is Google Lens.
Google Lens
Google Lens was first launched in 2017, and since then, more than 1 billion visual searches have been run. In 2018 alone, Google Lens could already detect over 1 billion objects.
The 2019 update integrated it in Android phone cameras, and extended its capabilities by including product search, recognizing bars, restaurants, and businesses (as long as they exist in Google My Business) and can be shown as Knowledge Graphs. Knowledge Graphs are the search result format Google provides to get precise information at a glance (the first answer Google displays, usually seen on the right on PC, and as the top result on mobile).
And since then, its use has only grown.
Among its functions you will find:
- Translation from and to any language supported by Google.
- Detection and identification. From dog breeds to dishes, buildings, and surroundings
- Smart text selection. Point, select, and copy text in any language to use it on your phone.
- Smart text search. Instant Google search made from your phone camera.
- Reading written questions (and providing answers like they were a regular search)
- And of course, shopping.
Google may have an advantage but is not the only one driving revolution. Pinterest had, in 2019 only, 600 monthly visual searches, and partnerships are on the rise to provide image search with Target among others. Instagram announced last year the implementation of both Camera Search and Visual search to provide exactly the same functions in a very similar interface as Google Lens. Meta has the means, for sure, but a different approach – which is no worse.
Benefits of using visual search
- Works as a complementary function to vocal search and typed search.
- Fast answers to questions, which get a little more accurate every day
- If you use Google Lens, it’s ad free!
- It allows your brand to be recognizable. Being there is already promotional if your product or brand elements (logo, design) are recognizable. And if that weren’t enough, every person using your product is a walking promoter.
- It’s multilingual. Products don’t speak languages, and images don’t them. Ads do. There’s no better promotion for a product than seeing it live and having your phone telling you all about it.
A few decisive stats
- 36% of consumers had used visual search. And half of them said that visual information is more important than text when shopping online
- 53% of respondents used visual search from their phones
- 37% said lack of trust was the main reason preventing them from using it more. 26% chose privacy instead.
- 24% linked this to the ads that appeared then when using a visual search (that was not as developed as it is now). (DSI Study, Intent Lab 2018)
- Google Lens was used about 3 billion times per month in In 2021.
Most searched product categories
- Clothes 86%
- Furniture 85%
- Cars 58%
- Groceries 56%
- Vacations 55%
- Household Items 46%
- Electronics 41%
- Wine and Spirit 41%
How to make the most of visual search in ecommerce
In the same way that being on Google’s first results page is a sign that you are someone, imagine how appearing on the top results in the “Shopping” tab can increase your sales in a vocal or visual search. Some products are not related to a specific brand and using your USPs as a pitch line can make a difference.
There are a few essential requirements to be discoverable via visual search. The first one is having an image search feature installed on your website, so user search has a catalog linked to it. The second one is, if you’re going for Google, is having a registered Google My Business account.
That said, there are a couple of things you can do to make the most of visual search in ecommerce.
Optimized images
We always insist in the fact that image quality is the most important thing to sell products online. People expect to have a clear view of what they intend to purchase, as close as it gets to a realistic touch & feel. As we said above, image search must be enabled, and you can do that with a simple plugin.
- Use the highest resolution in your images. Use a general shot picture, and then focus on each one of your products and features.
Mobile-centric content
The world went mobile a long time ago, and people run more searches on their phones than they do from other devices. That’s the main reason why phones include more inherently human capabilities, which in turn power more studies and developments.
- Make sure your product images and your website are mobile optimized to avoid bounces or scrollouts.
SEO
You may think keyword optimization has little to do with image processing, and that’s not entirely wrong. But the hybrid world we’re living in has pushed brands to use Omnichannel strategies to adapt to an evolving user behavior. This is what ALT text is for: allowing images to be recognizable by algorithms. But that depends solely on using keywords to match search intent and search results.
- Adapt your SEO if you want to be found seamlessly across different channels, be it via typed, vocal, or visual search.
Categorization
Algorithms may learn fast, but they still need you to categorize your products properly, so they know exactly what you are selling, and they can match it precisely with the image they process. The more details you provide, the easier it will be for algorithms to process your product’s identifying traits: colors, sizes, varieties,…
This is just the beginning
These are interesting times for online businesses. As most trend reports announced, this is the year when Artificial Intelligence and Machine Learning take-off visibly for end-users. Despite the disruptions and surprises we’ve seen so far, it is happening at an increasing pace.
The irruption of Artificial Intelligence and visual search in ecommerce is long-awaited and has been imagined in many forms, but usually remarking how much time it will save, and how it will enable more users in any part of the world to acquire products they believed to be out of reach.
And this, friends, is going to change a lot of things.
