Blog

Visual search in 2020: Where it stands and where it's headed

Jeff Hunt , founder and CEO of Snap36, offers up insight on why retailers need to start thinking about optimizing the retail site for visual search.

Photo by istock.com

February 26, 2020 by Jeff Hunt — CEO/Founder, Snap36

Visual search isn't just a theoretical concept anymore — it's a growing part of the search landscape. In the year ahead, we're likely to see big companies make plays to improve their visual search capabilities. If you're looking for market indicators, look no further than Pinterest and Amazon, as both sites have seen visual search increase in popularity in the past year.

Back in 2018, Pinterest reported that they fielded over 600 million visual searches a month, and there's little doubt that this number has grown as they continue to upgrade their visual search tool to make it easier for users to move from search to purchase.

Meanwhile, retail giant Amazon's visual search tools remain of limited use today — primarily helpful for housewares and women's shoes — but you can bet that this toolkit will get sharper in the months ahead as visual search continues to gain popularity.

Gartner estimates that by 2021, early adopters that redesign their websites to support visual and voice search will increase their digital commerce revenue by 30 percent. If this sounds appealing to you, here's what you should know about visual search today and how this technology is evolving in the year ahead.

The state of visual search

Every day visual search engines are actively using AI and machine learning to get better at image-based search. They're learning shapes, sizes, colors, and patterns, and they're learning how to use these qualities to identify objects in the same way the human mind does.

But the technology isn't excellent - yet. If you use Amazon's or Google's visual search tools today, you may be able to get into the ballpark of what you're looking for, but it's unlikely that you'll land on the exact object you were hoping to find. When it comes to search, text queries are still the best option available today and consumers know they can rely on it.

According to research from the Intent Lab, only 35 percent of consumers have conducted a visual search. For those whose results don't return the product or information they're looking for, there's currently little incentive to try again. Part of the issue may be a lack of familiarity, but there's also an overall lack of consumer confidence in this technology today. Even when a visual search feature is available, only 20 percent of app users make use of it.

Barriers to visual search adoption

Even apps like CamFind can only take you within so many steps of the object you're looking for. But that's not the real barrier to investment in 2020: the problem is that it's hard to make money using visual search today. Even with an accurate and reliable visual search tool, this technology hasn't been monetized yet, and it's unclear how brands will be able to compete around similar products in the visual search arena.

In some verticals, like fashion, visual search can be tremendously useful, even with categorically accurate results. For example, if you're looking for a pair of shoes that's similar to the one you've seen, a visual search tool may help you on the path to purchase. But brands like Nike and Adidas may not want their customers using this source yet, as similar design elements could cross-promote brands within the search.

The other reason this lack of precision in modeling is an issue is even more technical: detailed pattern recognition requires a lot of lift. For example, the amount of machine learning needed to distinguish between two small similar auto parts and two t-shirts are quite different.

The ML models simply aren't there yet, and it takes a lot of computer power and money to train these machines to provide the experience consumers expect. This all means there is a ton of upside, and industry experts expect to see these programs turn a corner in the near future.

Start thinking about optimizing the retail site

Even though visual search isn't broadly available (or useful) yet, ad revenues for visual search are predicted to grow to $865 billion by 2022. Now is the time to prepare for visual search by optimizing your site for this coming capability. There are a few easy steps you can take:

1.    Build an arsenal of high-quality images. Feed your visual search engine the assets it needs to become an accurate  “visual vocabulary” for visual searchers.
2.    Mark up your images properly. Google's guides are the gold standard.
3.    Work on image text and include images in your sitemaps.

If the first item on this list seems like a whole other topic, that's because it is. As you look at your product offerings and how they're presented, consider which are easy to find through textual search and how they might be recognized (or difficult to identify) based on their attributes.

Because investing in product photography takes time and investment, begin to think about the changes you might want to make now and consult with experts before making an investment in the future of your product imagery.

Consumers are ready for visual search

Consumers may not be using visual search regularly today, but they're ready to adopt this new technology. Research from ViSenze found that 62 percent of Gen Z and Millennial consumers want to have visual search capabilities, and their appetite for the technology means that this trend is worth your attention.

To paraphrase Pinterest's CEO, the future of search will be about pictures rather than keywords, especially in the world of online retail. While text and voice will continue to dominate queries for accurate or abstract ideas, the rise of visual search is assured.

Jeff Hunt is the founder and CEO of Snap36, which provides interactive 360° product photography services.