Upsy Shopping Helper increased its e-commerce clients’ average purchase by up to 41%.

A big moment for a business owner is making a sale on their new product, but even a bigger moment is when they hear that a product has exceeded the expectations set for it. That’s when you know you are on the right path. We have had the privilege to enjoy that feeling for the past few weeks and hoped it would never disappear. At the peak of our product development, Upsy Shopping Helper has been forging up to 40,76% increases in our customers’ average purchase sizes.

Everyone who has ever operated an online store knows how good those numbers taste and everyone, who has ever delivered tools for an online store owner knows, how happy it makes you to have a satisfied customer.

What’s happening now, was a utopia six years ago when we founded Upseller, but our mission was the same back then as it is now. Our job is to make a small or a medium-sized online store to increase its sales. Six years ago, it meant a chat. It felt great having a person to speak to another on a website. Then we introduced AI to support the chat, to fetch answers to simple, but frequent questions. The third phase of the product development were chatbots utilizing NLP (Natural Language Processing). Now, you could take the human entirely off the easiest questions and let the bot talk to customers. However, something was still missing. We hadn’t started a company to improve customer service, but first and foremost, to increase the sales of our clients.

That’s why the recently launched Upsy is a big deal for us, and an opportunity to deliver exactly what we founded the company for. Until now, we have focused on the B2B market and selected B2C cases. Upsy, however, focuses on e-commerce, where chats and bots don’t deliver results. Upsy utilizes machine learning and AI, but not just reactively by searching answers to the e-store customers’ questions, but having learned enough about the customer’s needs, it can proactively offer suitable as well as supplementary products to the customer. This is exactly how a savvy brick-and-mortar sales rep operates. First, the sales rep asks questions to understand the need, then they offer solutions and exciting additional spices and finally they find an appealing price for the package. Accordingly, the basis for developing Upsy has been sales psychology, and only after that, the machine learning mechanisms have been put in place to serve the good purchase experience.

It is intuitively clear, that also in e-commerce a skillful sales rep sells better than a perfectly ok customer service rep, but it hasn’t been possible to build one before. Now that it is, the results still manage to surprise me. We have set up Upsy for several of our e-store clients, and Upsy met or exceeded our expectations in increasing above all the average purchase size. When the sales rep finds the exact right product for the customer and recommends a few supplementary products, in addition, the sales will be better than without a sales rep. Nevertheless, the fact that the customers of our control group increased their average purchase size with the help of Upsy at minimum 24,4% and at best 40,76% surpassed our expectations.

The new factor that enables such astounding results is the utilization of middle-weight AI also in small and medium-size stores. Not long ago, AI was the pricy specialty of the enterprise-sized, and it seemed as only the big, international stores could take advantage of it. But, as in all technology, the prices go down and the availability improves. With correctly thought-out applications it can now be brought to the reach of SME stores and then smooth out the playing field with a whole new force. Now, even the smaller stores can utilize the tools, that were a competitive advantage reserved for the bigger players a moment ago.

In these blogs of mine, I like to draw analogies from world-class sports. Board sports are my wheelhouse: with them, I learned how to win and to coach others to win. When things threaten to get complicated, I like to go back to my boarding experiences to gain clarity in thinking.

In my mind, the arrival of AI for the small e-stores is comparable to when Youtube came to the aid of snowboarders. Before Youtube, we trained by ourselves without seeing or hearing others, and we had to travel to major international competitions to see what the true level was in the world. We trained in the dark and updating one’s skills was expensive.

A few years ago, I took up wing surfing alongside snowboarding as a summer sport. While training, I have witnessed the sport developing five times faster than snowboarding back in the day, because everyone has access to the newest know-how at an instant. On Youtube or Instagram, you can view what sort of tricks someone has developed or what sort of wave someone has ridden. The democratization of information with the help of new, accessible technology has accelerated the sport’s development to a wild speed.

Now, the same is happening in e-commerce. The information and understanding, that was recently accessible only to the biggest, is now available for the small and medium-sized as well. The know-how and results develop rapidly in the whole industry, and especially with those, who take up the new tools.