Is Visual Search a hype? - 4 questions to answer before implementing

TL;DR: No, it is not a hype. But you need to do your homework before deciding to invest in Visual Search.

In recent years, with giants like Google, Amazon, Bing, and Pinterest doing heavy marketing for their visual search engines, end users are adopting the technology and recognizing its value.

We see visual search emerging as one key marketing trend for digital commerce. Large retailers in fashion, home and furniture, and tools and appliances have launched visual search engines to aid their customers.

  • 21% of advertisers believe visual search is the most important trend for their business in 2019. 

  • 45% of retailers in the UK use visual search in 2019.

  • By 2021, early adopter brands that redesign their websites to support visual and voice search will increase digital commerce revenue by 30%.

What values does Visual Search bring to your business?

There have been reports showing economic benefit of visual search. But the technology add-on is more valuable to certain business types and industries than to others.

The first factor is product type. Visual search is more valuable to products with extensive visual variance and to products for which visual attributes is the main distinction among items. Fashion, furniture, plants, groceries, machine components and tools are among the most popular product types to apply visual search.

The second factor is how digitized your marketing and sales activities are. Visual search is primarily beneficial to online marketing, digital media, digital after-sales and maintenance processes. It means that if your sales, marketing and customer support team have not invested much on digital applications and platforms, your business is not ready to reap the benefits of visual search.

Use case examples are:

How do you solve the data problem?

Visual search is powered by artificial intelligence. It requires sample images of the products to train machine learning models. To ensure high accuracy, sample images should resemble real images of products that customers take or upload. Therefore, model images taken in studio are not ideal for the purpose of training recognition AI. Your business may have already have sample product images in place or not. Either way, collecting, organizing and reviewing sample images are complicated tasks that may require consultation from computer vision experts and will consume large amount of time if not done right. Data strategy plays an important role in implementing your visual search.

Should you build your own team for Visual Search or use a service?

Most businesses do not have in-house machine learning experts. Therefore, when you consider applying visual search, you have the options to add this expertise to your own team, or use the service of visual search or computer vision companies.

Since machine learning team requires large financial and technical effort to build and maintain, the former option is more reasonable for businesses whose strategy is to invest heavily in an in-house IT team. It would make sense for companies who want to make visual search their core competency and have resources (money and time) to build up the technology.

Meanwhile, the option to seek external service providers would help your business acquire the market-standard visual search capability at more competitive price and time. However, this is not an one time cost, but will require frequent model customization and updates. Fortunately, barrier to switch service providers is not high for visual search service. And as the visual search market expands, there will be more and more service providers for you to select depending on your budget, use case and requirements.

How do you select a service provider?

If you decide to go for the external service option, do not be overwhelmed by all the technical details while choosing one service providers. Generally, there are two types of service providers: the giants and the emerging companies, i.e. the startups.

The giants enjoy better investment in infrastructure, so they are better for large businesses with lots of customization, varied use cases and complex requirements. The startups usually focus on certain use cases and provide more consultant to small and medium businesses. Do not mistaken that visual search by startups is not as reliable in terms of accuracy and performance as compared to the giants. Usually, it only offers less in terms of applications or use cases, while all of startups have the tendency to expand to more applications eventually. If your business is cost sensitive, it also makes sense to look for the services by startups, while again, not all startups provide cheaper service.

If you wish to analyze cost and benefit of visual search for your business, contact us for support and demo.

SHARITEK provides advanced visual search technology to enable quick search with camera for products or spare parts.

436 views0 comments