Small and medium-sized eCommerce companies may think applying artificial intelligence (AI) to retail is only for large retailers. They are envious of the ability of retail giants to define individualized customer journeys, send users relevant products and information, and tailor offerings to specific client needs.
Veer AI, a Toronto-based company, has a mission to bring these enterprise-level AI capabilities to retailers of almost any size.
Fatima Khamitova, one of the founders of Veer AI, is a longtime retail consultant with a Masters degree in analytics. A former Deloitte consultant, Khamitova has worked with a “who’s who” of larger retailers, exploring “the intersection of machine learning-driven marketing.”
Khamitova loved her job at Deloitte’s AI Lab but ended up leaving it for the excitement of a large VC-funded AI startup. But even in this high-energy world, something was missing. “One of the challenges,” she noted, “is that we could only really service companies that could afford to spend hundreds of thousands and often millions of dollars on these types of projects.”
The result was, with co-founder Jiachen Yao, the birth of Veer AI and a mission to “democratize access to data-driven insights.”
The value proposition was aimed at the small and medium-sized business market. But Veer AI quickly discovered that it wasn’t only the cost of programs they needed to address; Khamitova saw “a big gap in the tools that were out there. A key failing was that most tools, even if they were affordable, were descriptive but not necessarily prescriptive in terms of what to do with that data.”
Retail is driven by direct links from action to result. With small and medium-sized businesses, the issue of results is mission-critical.
Khamitova acknowledges this, and sympathizes with these companies. “Small businesses have to operate with the best they can get, the easiest tools that are cheapest; otherwise they wouldn’t be able to survive.” Yet the cheapest tools often don’t get the results needed.
Khamitova felt that Veer AI could address this issue. “We wanted to be able to not only provide dashboards and data and such but also tie them in together with the type of marketing tactics you need for customers to take advantage of data insights.”
Veer AI has taken a different approach. Much of the spotlight in retail has been on dashboards and “recommendation engines” which have brought incredible results for companies like Amazon and Netflix. However, according to Khamitova, that approach doesn’t always work in the SMB niche her company serves.
“Traditional recommendation engines work by looking at historical data,” she said. “They are useless when a brand introduces a new product. So a lot of apparel and fashion businesses wouldn’t necessarily benefit from something like this since their products are replenished every few months.”
“Most of them (customers) will never go back. Most of the types of clothes would never go back to the shelves as that is just how their business operates. A traditional recommendation engine would have insufficient data for all new purchases, which makes it a not very useful thing.”
Veer AI has instead concentrated on identifying customer segments, and finding often unseen patterns in customer behaviour.
Khamitova noted the example of one company which, from the data Veer AI gathered, found that their customers absolutely loved turtlenecks. Using this data, the company was able to do more targeted campaigns where they looked at the potential value of customers who never bought turtlenecks before, and offered them a voucher for 20 per cent off. The results were astonishing. “The lifetime value of these customers doubled,” said Khamitova.
Veer AI’s data-driven approach has shown that “one strategy does not fit all customers.” In another example, Veer AI demonstrated that discounting for loyalty doesn’t always work. “People who buy gift cards are seen as 30 per cent more loyal to a brand,” said Khamitova. “You really like it, and will spend more money.”
That’s the traditional wisdom; but with analysis Khamitova discovered that at least one of her customers who makes baby clothes found that their customers were buying gifts not because they especially care about the brand. This type of gift card buyer was least likely to rebuy. Offering an incentive was not necessary, and unlikely to succeed.
Hyper-targeting to customer segments is an idea that has proven itself in large retailers and eCommerce companies. It turns out the same results can also be achieved with smaller retailers. Good data and analytics leads to better targeted offerings.
Veer AI’s approach allows a retailer to target “the segment that will most likely respond to the offer, and if they respond to those types of offers, they become better customers.” To these retailers, a “better customer is a more loyal customer, one with a better and more intimate relationship with the brand, who comes more often and buys more every time they visit.”
For many stores this is the difference between success and failure. Khamitova noted that for her customers, “70 to 80 per cent of their money comes from existing customers. If you don’t have high return rates, there’s a big opportunity to invest more into return customers. We help by providing those very prescriptive insights, saying this is exactly how you should be working with these types of customers.”
The proof is in the data. “Being able to hyper-target your customers based on their characteristics clearly works,” said Khamitova. “Campaign emails we’ve been sending on behalf of our customers have been opened at a much higher rate – 100 to 200 per cent higher than previous campaigns. The clickthrough rates are higher, and we have a much higher rate of conversion than they’ve ever seen.”
Through examples like this, Veer AI has been able to prove the relevance of their offering to even small and mid-sized retailers.
What do you know about my business?
Despite these successes, and even with a promise of direct results at an affordable price, there was some reluctance. Even the qualifications that Khamitova brought to the table were not as valuable as she might initially have thought. She would quote a series of brand names of large retailers to her potential clients only to hear them ask, “What do you know about my problems?”
Her answer is to speak the language of her customers. Khamitova will give concrete examples of how much money can be lost by unnecessary discounts. She’ll take, for example, a segment with high disposable income, and show why discounts for this segment are counterproductive, and can even lead to substantial losses.
“When someone has the discretionary income to be able to spend $20,000 on clothes, they don’ t care about you giving them an extra 20 per cent off sale prices,” said Khamitova. “A lot of people may still give it to them, thinking it will keep them around, but it’s actually not the case. How do you differentiate the marketing when it comes to those different types of customers? If (their purchase) was $20,000, and you gave them 20 percent off, they end up saving the customer $4,000. But that’s $4,000 the brand ultimately lost on people that would have purchased regardless of the special discount.”
Khamitova has also learned to tailor her language to her clients. Buzzwords don’t impress small and medium-sized retailers. “Terms like segmentation and data and machine learning are so overused, and they’re thrown at people in the hopes of getting them to spend money,” said Khamitova. “I understand the hesitancy behind speaking to someone who talks about such things. Sometimes the people using these buzzwords don’t know a lot about what they’re talking about. I’ve come across many data people who, in my opinion, have no real idea what they’re doing.”
That straight talk, and great examples, have won over some key customers. With that, Veer AI has discovered the power of concepts from eCommerce and “as a service” models. They have taken on companies on a “freemium” or trial basis, and the results are winning over customers. Many are now in the process of converting to the paid version of the product.
Measurable results count, but even more powerful is the power of “peer reviews.” Khamitova notes that while naming past customers like Roots may not impress, her current customers are attracted by the fact that “some of the brands that use us are relatively famous Queen West brands.” (Queen West is a highly trendy retail area in downtown Toronto.)
“It’s very interesting to see the interest from them,” said Khamitova. “The moment we started having some of these brands was the moment things really changed. It definitely made a lot of sales easier.”
The power of the cloud – the future of retail Khamitova acknowledged that none of this success would be possible without leveraging the cloud. Veer AI has the classic benefits of being able to attract talent from anywhere. In addition, the cloud allows the company to focus not on infrastructure but on the unique elements of their value proposition. For clients that already have an online store or marketing, Veer AI is able to work on standard interfaces to a relatively small number of cloud “e-stores” that support retailers.
Because these e-stores have their own standards, Veer AI has also been able to have relatively standardized information from all their clients. Without compromising any confidential information, they are able to benchmark their clients against a larger pool of data.
However, Khamitova does not believe that “the store of the future” is just an eCommerce store but, rather, a hybrid “where you have the ability to shop both in-store and online.” She has been seeing examples of this during a global pandemic that accelerated eCommerce shopping. In her mind, “all the best-performing stores have both a brick-and-mortar store and a theory of robust, well-built websites.”
Khamitova is optimistic about the future of Veer AI. She said the company’s main focus is on providing more to customers.