ViSenze Blog

6 Things about visual search technology all ecommerce Product Managers need to know

Posted by ViSenze 02-November-2015

Welcome to the wonderful world of visual search technologyIf you think you’re in need of a crash course on what visual search technology is, its various capabilities and use cases, benefits and more, you’ll be in for a nice surprise. This introductory guide is exactly what you need to know about the software and how to implement it onto your e-commerce platform.

Let’s hit the road.


Before implementation

1. What does visual search technology mean?

One of the biggest points of confusion for Product Managers is the differences between image search, reverse image search and visual search. The terms sound so similar on paper, but they mean a world of a difference.

To clarify:

Image search Reverse image search Visual search
Simply means that images are returned as search results - regardless of the process, method or the initial search query input.                                                                                                                                                    A subcategory of image search, this allows for an image (or its URL) to be used as input query and returns results related to the image.

Again, the process and method of how search results are generated are varied and diverse - where the search process could be carried out by using metadata and keywords or by pure computer vision.
An even smaller subcategory within reverse image search, visual search specifically refers to the technology that carries out pixel-to-pixel comparisons between images in order to generate visually similar results.

2. What visual search use cases can be applied to your platform?

Visual search as a core technology is able to power different applications on an e-commerce platform.

For example, here’s a simple breakdown of the use cases that ViSenze’s visual search technology can offer:

Use case What it does Benefits
Search by Image Allows your consumers to search by using an image instead of keywords. The query image can be uploaded or inserted via an URL. The results return visually similar items in the product catalog based on attributes such as color, shape, patterns and more.
  • Increase consumer satisfaction by surfacing relevant results easily.
  • Decrease time spent by consumers who are searching for a specific type of item.
Find Similar

Presents visually similar items when the consumer would like to do a deep dive on a particular product.

  • Increase engagement by allowing consumers to narrow down their search by easily refining preferences.
You May Also Like Recommends products alongside the one your consumer is currently viewing, based on visual similarity and metadata.
  • Increase conversion rates by making sure consumers don’t hit any dead end and always find something they like.

Shop The Look

Allows your consumer to pick a photo from a curated gallery populated with user-generated content, crop to zoom in on a particular item in the picture, and search for visually similar products available on your site.
  • Increase conversion rates by making the path from inspiration to purchase smoother and shorter.
Automated Mix & Match Suggests products from other categories on top of the item that your consumer is viewing to complete the look, thus acting as a personal stylist.
  • Increase basket size by recommending complementary items and suggested pairings.

To see these use cases in action on various web and mobile e-commerce platforms, visit ViSenze’s Youtube Channel.

3. When can visual search technology be applied to your database?

The technology works best when your image database is large and varied. As a rule of the thumb, we would suggest 3,000 unique SKUs per vertical in a database as a prerequisite for the software to be effective.

The reason being that as the software is applied to a single, private database such as retail and e-commerce platforms with product catalogs, the search results generated depend not only on the accuracy of the technology, but also on the size of the database.

So if the volume of images in your database is low (thereby limiting the number of matching and similar images), implementing visual search technology to your platform probably would not achieve optimal results.

4. What do you stand to gain from using visual search technology?

Offer a tightly integrated offline to online experience

The online world is being merged back into the offline one, and you need to find ways to incorporate technology into different stages of the customer journey - starting with their offline inspiration to online product search.

By offering a visual search application that users can access on their smartphones, they can easily take photos of products around them in real-life and find similar ones on your e-commerce platform. The offline to online transition is now made seamless and smooth.

See your sales figure go up, up, up

Visual search reduces friction - your consumers just take a picture of a product and let our technology find the corresponding item on your platform.

As compared to the existing keyword search function on your website, visual search takes less time and provides better results. It speeds up processes online, increases basket size and boosts your engagement and conversion rates.

Avoiding lost sales due to stockouts

When consumers hit a dead end on your site because an item is out of stock, product recommendations will give them visually similar alternatives and  “Check out” rather than abandoning their cart in disappointment.


During implementation

5. Resources required

In some cases, certain technology vendors make available their cloud-based software APIs for integration with clear documentation so you would only need a few in-house engineers and user experience specialists to carry out the integration.

It is also possible to reach out to third party system integrators to get the job done. Depending on the complexity of the use cases you wish to use for your site, the process would take a couple of days to a few weeks to complete.


After implementation

6. Measuring ROI on visual search technology for your business

Often, when it comes to measuring the costs and benefits of implementing visual search technology, most people are not sure of what to measure and how to measure its ROI.

We’re here to tell you that the process is not as complicated and tough as you may think.

To start, determine exactly what the technology enhances for your organization in terms of hard metrics and financial gains, such as the following:

  • Conversion rates
  • Click through rates
  • Basket size
  • Product view numbers
  • Bounce rate and exit rate

Next, factor in costs, including:

  • Software cost
  • Implementation cost

Lastly, apply your calculations as you carry out your testing and slowly release the function to more and more of your consumers through different stages.

To further optimize your metrics, also carry out A/B testing - not just with or without the visual search feature, but also the function’s placement and position on your site page. Ask questions like where do your consumers conver? On which part of the page is there better engagement?


So what’s next?

Test our visual search technology for yourself - have a go at our demo.