Euroflorist sees sales bloom through website optimization


There’s an old advertising slogan in the floral business: “Say it with flowers!”

Euroflorist wants its online customers to say it more often, so the Amsterdam-based company is using a powerful new optimization tool to tailor its website in ways that turn more browsers into buyers and increase the average purchase.

At the heart of the program is an artificial-intelligence platform that offers more data about customers and how to adjust the website to increase sales rates.

“I’ve been dreaming about the potential of AI for a long time, and when I came into contact with Sentient Technologies last year, we both knew Euroflorist would be a perfect test case,” says Guido Jansen, chief psychology officer for the online retailer.

Jonathan Epstein, senior vice president of international for Sentient, is just as optimistic about the partnership. “Euroflorist is the leading flower supplier to many countries in Europe,” he says. “They were among our first trials in the European market, and results have been excellent.”

Accelerated experimentation

Sentient’s Ascend tool is a website and landing-page optimization system based on a form of AI called “evolutionary algorithms.” Ascend was designed to help marketers achieve their growth goals much faster by accelerating and automating new design testing and experimentation.

With some 20 sites in 11 countries, Euroflorist is one of the largest distributors of flowers and gifts in Europe, selling more than 2 million bouquets a year. Euroflorist traces its origins back to 1947, though the business really started to take off in 1982, when engineer and entrepreneur Peter Jungbeck launched Svensk Blomsterförmedling in Sweden with his friend and business partner, Lars Hedberg.

Jungbeck drew on his technical background to build a new system that simplified the process by which florists relayed flowers between each other. The company soon had a strong network of florists that delivered flowers across countries; Euroflorist now works with more than 54,000 local European flower shops, making it one of the most extensive and reliable floral networks in the world.

The company has about 200 employees and sends out about 40,000 bouquets weekly. Jansen says the average purchase ranges from $41 to $47.

Euroflorist saw an overall 4 percent increase in the clickthrough rate and a 5.5 percent increase in people ordering more items.

Euroflorist launched its online operations in 1995, becoming one of the first 100 online companies in the world. From the beginning, the site accepted credit cards, although transactions were handled via a Scottish bank since no bank in the company’s home country of Sweden was ready to deal with online shopping.

Euroflorist began using Ascend in 2017 after recognizing that AI could bring significant value to the business; 60 percent of its revenue comes from online commerce, making increased conversion rates a priority.

The company had been doing research on its own, using A/B testing to measure the success of changes to its website while limiting potential damage. The method, also referred to as “split testing,” enables companies to compare the performance of two web pages.

“In the past, only one out of five A/B tests were successful,” Jansen says. “We considered that a pretty good number. But it was taking months to reach a modest 4 to 5 percent lift.

“Our team does a lot of user research and A/B testing, but we were struggling with limited resources. That limited the number of tests we could run.”

The main goal was to lift sales, Jansen says. “Ascend enabled us to test hundreds of combinations in the same amount of time as one A/B test,” he says. “This worked even in some smaller markets, where A/B solutions don’t work due to the low traffic. In our initial test, we saw up to a 17 percent increase in conversions.”

The results come as no surprise to Epstein. “We saw a problem between the speed at which people want to attain improvement and how fast their businesses are operating and the tools that they use to tune their user experiences,” he says.

Epstein says Ascend allows marketers and retailers to test not just one change to the user experience but as many as 50 changes simultaneously. “If you tell it to improve the conversion rate, AI automates the process and displays different designs to determine which will be the best one,” he says. Sentient is working with major airlines and other retailers, he says, and is “seeing rates anywhere from 5 to 50 percent.”

Increasing numbers

For Euroflorist, it was a matter of getting more data on which to base business decisions. “We knew it would be a win for the company,” Jansen says. “We knew AI was an emerging area and we had to experiment with it in order to get ahead of the competition.”

The results of the Sentient system were a foregone conclusion before it was rolled out companywide. “We did a simple test as a trial,” Jansen says. “We tested in three countries, Norway, Sweden and The Netherlands, and we had 256 variants of the product page.”

He says the implementation period was short. “We did hold a workshop to determine what we wanted to do on the web page, based on all the changes that had been made and research that was already done. We picked eight elements on the page, and for each element, we made two variants. Basically, that was our entire investment of time. Then, Sentient implemented it into our website.”

Since then, the company has seen an increase in sales, Jansen says; after the test, the company saw an overall 4 percent increase in the clickthrough rate and a 5.5 percent increase in people ordering more items.

“I think we can increase those numbers as we go along,” Jansen says. “Basically, we’ve done the proof of concept, showing that the system really works. We found out that with Sentient, we can run four times the tests we normally do with fewer users. And we’ll be doing more tests on the entire customer journey, from the time they enter the website to when they leave.”

Len Lewis is a veteran journalist and author covering the retail industry in the U.S., Canada, Europe and South America.


Comments are closed.