In eCommerce, the limitless options for online shoppers have organized into a limited, top-heavy landscape. To keep up with Amazon, traditional retail chains like Walmart and Target have upgraded their digital storefronts. These retailers commit big budgets to staying at the top of the heap, springing for designer AI services from the likes of IBM Watson.
A new solution from Chicago-based firm Avatria, called Avatria Convert, has launched in the new year, offering affordable machine-learning services for midsize businesses and small-enterprise players. Although this is the first product from Avatria, the company has been working with major retailers like Discount Tire since its inception in 2014. Founding partner Harry Thakkar assembled his team from colleagues who worked for years with similar brands at Accenture Interactive. According to Thakkar, the Convert product will likely expand into a suite of offerings for etailers and other digital outfits.
The combination of human experience and data-based expertise at Avatria has yielded an effective gameplan for merchants looking to optimize the buying experience for their customers, although some of the facts-based insights about consumer behavior might leave marketers scratching their head.
A major factor born out by all the data Avatria has analyzed is the consumer’s limited time. (And this also applies to the B2B space – which the company also services – perhaps even more so.) Because consumers are short on time, they are quick to either purchase or move on. They also are unlikely to click through multiple pages of items, or to set search parameters to improve the selection of items to look at. Instead, web shoppers – 73 percent of them – tend to look at what’s presented on the first page, and either buy something they see, or move on. Eight-eight percent refuse to use search filters on the site, according to client data from Avatria. The mission, then, for Avatria’s clients, is to be more effective in showing what visitors are looking for immediately. Machine learning and algorithms that analyze previous customer behavior make it possible to refine what merchants show to their visitors in future experiences.
While an Avatria solution can be up and running for a client in a month, much faster than larger big-enterprise AI offerings, in Thakkar’s experience, Avatria also looks to lengthen the duration of the relationship with clients.–>
“We’ve put a heavier focus on long-term relationships with customers,” Thakkar said. “Our goal is to continue to add value two, three, four years down the line. Secondly…based on the twenty-plus customers we’ve helped, we have a very good understanding of gaps and pain points in the marketplace, and a good understanding of what did and didn’t exist in that space previously.”
One problem Avatria came across repeatedly was the dismal performance of the merchant’s search tool. Also, search filters, seen traditionally on the left side of the screen in Amazon’s layout, were rarely used, according to the data.
“We saw that the site search wasn’t working, and [visitors] weren’t able to find the products they were looking for,” said Thakkar. “With the analytics data customers already have on the website, we then pass it into machine learning algorithms we’ve developed, specifically geared toward metrics and data points related to eCommerce.”
From 40 different metrics, Avatria Convert tracks and analyzes what’s looked at in the product inventory on a site, in addition to customer data, how long visitors spend on a product details page, and how many times a product is looked at before the purchase. With a focus on the “ready-to-buy” stage of the consumer journey, merchants with many categories of products can save time by zeroing in on which categories are more active and need further adjustment.
Thakkar adds, “Ultimately the algorithms assess those different metrics and normalize the data so that the client is comparing apples to apples. We can then provide recommendations on how to display products in a way that will be most appealing to customers.”
For midsize businesses on the fence about utilizing this kind of data for their digital store, the retail segment isn’t as important as sales volume. Thakkar suggests that 50,000 monthly visits or purchasing “sessions” is a minimum for the sample size to be large enough for the machine learning to make an impact. This also translates into revenue above $1 million.
According to Thakkar, Avatria’s A/B testing demonstrates that the algorithms improve conversion rates for clients of this size, 30 to 70 percent.
“A lot of marketers are pretty savvy and have a good sense of what’s happening on their site,” Thakkar stated. “But they don’t always have enough data points to make decisions on how to focus their time.”
AI can save time for both consumers and eCommerce sites. If merchants have the time and data resources to optimize their display, visitors are more likely to find what they want and make a quick purchase.