Some websites and apps let shoppers take pictures of products and launch an online search for similar items. While visual search is not yet as accurate as text search, recent developments make good product photography more important for online retailers.

Jeff Hunt, founder and CEO, Snap36

Jeff Hunt, founder and CEO, Snap36

Pay attention to ecommerce industry leaders, and you’ll hear that visual search—taking a picture with a phone and searching for matches as you would with text in Google—is “the next big thing.” Because of that, you’ve probably also heard that it’s essential to optimize your site for visual search.

While I agree that images of products are more important than ever to today’s online customers (and therefore to online retailers), I think it’s worth slowing down a minute and examining the state of visual search so that we’re creating visual assets that will both offer meaningful ROI in the present and position us for future growth. Here’s a look at how internet retailers can do that.

The Future of Visual Search (And Why We’re Not There Yet)

The future state of visual search seems tantalizingly close. Already, Amazon and Pinterest have functions that let you snap photos and search their databases for similar items. And the app CamFind functions as a kind of Google-for-pictures—a search engine that uses images instead of words, as does Google’s own app via Google Lens.While we are seeing new apps and advancements with visual search, it’s far from replacing good old text search.

Retailers must mark images up properly so that search engines can easily interpret them and serve them to searchers.

This is true for three main reasons:
1. The tech’s not there yet. Right now, Amazon’s visual search function is fun and interesting but nowhere near as accurate as its text search. If you actually want to find and buy something, it isn’t yet reliable. Behind the scenes, powerful AI engines must be trained to recognize the objects you want to search for with hundreds of thousands—if not millions—of images. And even with CamFind, which lets you get general information (like identifying the species of houseplant you just inherited from your aunt) instead of just purchase options, results aren’t always accurate.
2. It’s too hard to make money. Even if the tech were perfect—meaning it had 100% accuracy in identifying images—monetizing digital search is still a mystery. So for example, if I snap a photo of a piece of my dishwasher that’s damaged to find out what it is, where do I end up? What if the best match for this crusty old part is from someone’s art photography blog? Is the artist who owns the site charged for that traffic? And if I then return to Google and conduct a text search, based on what I learned from the artist, to order the piece from Amazon or Home Depot, how would either site know that my purchase started with an image search? Without solving for that, would it matter whether or not Amazon or Home Depot were optimized for visual search?
3. It’s useless for idea-based queries. What if you want to know the weather in New York right now? Or you’re trying to remember the five stages of grief? Or you’re wondering what this building you’re hearing about on a podcast looks like? You’d never use visual search for those queries or millions of others. So even when we have the capability to conduct effective and accurate visual searches, it will never fully replace text (or perhaps one day, voice) search.


Still, the expectations of mobile users are changing. More and more – especially in retail settings – we expect to see multiple, high-quality images of the things we’re thinking of buying.

How to Thrive in the Present and Future of Visual Search

We know that images play a powerful role in improving conversion rates in ecommerce – and that image type matters. Recent research shows that buyers who encounter 360-degree photography are 14 percent more likely to make a purchase than those who see still images. Those with a virtual try-on option are even more likely to buy.

So it’s clear more images help drive more conversions. The question is, how many images? And of what kind? What constitutes “enough” to be competitive today and remain competitive in the years to come?


To answer that, let’s consider two scenarios that online retailers may well to encounter in the near future:

Scenario 1: You receive an offer to have your products on the shelves of an Amazon Go-type store. The retailer offers you shelf space and you want to jump at the opportunity. Here’s the catch: they require dozens of high-resolution photos of your product.

Why? Because frictionless shopping (that is, the no-cashier model where shoppers take what they want and leave, while the retailer charges everything to their account) relies on more than shelf sensors. It also depends on a network of cameras tracking everything customers touch to capture as much data as possible (as well as information about mis-shelved items).

To participate in this exciting future of retail, then, you must have the quantity and quality of visual assets required for a visual search application to identify your products.


Scenario 2: You get an offer to do product placement—in an existing movie or TV show. How? The tech wizards behind the program have figured out how to replace every x item on screen (say, a chocolate bar) with one of your products (the Scrumdiddlyumptious bar, for example).

What do they need to make this possible? Images capturing your product from every angle, of course.

The beautiful thing here is that investing in high-quality images isn’t prohibitively expensive, and it’s just as easy to commission many as it is to commission one. For online retailers who haven’t yet made this investment, now is the time; waiting for the hype means you will be unprepared for opportunities that are just a few months or years down the road.

Beyond the Images: Be Visible to Search Engines

Of course, succeeding in visual search today is just as important as being prepared for its future, and while high-quality images are essential, they aren’t enough to stand out. Retailers must also mark those images up properly so that search engines can easily interpret them and therefore serve them to searchers.


That means including necessary labels (alt tags, title tags, file names, and captions), building an image sitemap, using Google schema markup for assets like recipes and gifs, and choosing the best size and resolution for your users (which means balancing the demands of load speed and quality).

So, yes: it’s time to optimize for the image-heavy search results of today and prepare for the visual search behavior to come. Those who do will be best situated to meet shoppers’ shifting demands and seize the opportunities that new search techniques introduce.

Snap36 provides spin photography technology and services.