The new gift bot factors in the gift giver’s relationship to the recipient, plus the recipient’s age and interests. 7% of consumers who use the gift bot add a product to their carts.

UncommonGoods LLC has souped up its gift guide with artificial intelligence.

Late last month the online-only retailer of niche home decor products launched a bot called “Sunny” to help shoppers find gifts.

Shoppers on can find a banner ad marketing Sunny as an expert gift finder. They can click on the link to “Find a gift they’ll love.” Once on the landing page, shoppers enter who they are shopping for, such as a brother, friend or mom and select an age group from the choices of baby, kid, teen, adult or senior. The shopper then selects interests for that person from a list of 20 interests, such as cooking, dogs or beer. Shoppers can also input other interests that are not listed. Shoppers then click “see my suggestions” and can browse a list of products factoring in these inputs.

On the product results page, a shopper can hit a “star” on an item to indicate that she likes that product and UncommonGoods will populate more product results for consumers to scroll through based on shopper preferences.

This is a rich set of gifting data we didn’t have before.
Zack Notes, senior product manager at UncommonGoods

The bot is powered with the technology of UncommonGoods’ search results, which BloomReach Inc. generates, and the retailer’s product recommendation algorithm, which Kibo generates, says Zack Notes, senior product manager at UncommonGoods. The algorithm, which UncommonGoods built in house, uses artificial intelligence, which means it learns over time the more consumers use it.


For example, if a consumer is looking for a gift for her adult niece who is interested in coffee, traveling and entertaining, and stars a coffee mug with a map on it, that same shopper will then see more products based on a different consumer’s star of that mug, such as products the other consumer starred or eventually purchased.

Although the gift-finding bot has only been live for a few weeks, 50,000 consumers have used Sunny, and the retailer is pleased with the results, Notes says. Consumers who land on the Sunny page and make a purchase have a 45% higher average order value than a consumer who visits the website at a different entry point and makes a purchase, Notes says.

Overall, the conversion rate for consumers who use the bot is similar to its average conversion rate, which Internet Retailer estimates is 1.96%, according to UncommonGoods is No. 746 in the Internet Retailer 2017 Top 1000.

UncommonGoods has a few metrics it is hoping the gifting bot will achieve this holiday season: A bounce rate, which measures the number of consumers who leave the page without clicking to anything else on the site, less than 10% (the bot’s bounce rate is currently at 15%), 40% of consumers to star an item (currently about 30% do), and 10% of consumers to add a product to their cart (currently at 7%).  For comparison, about 15% of UncommonGoods shoppers who are on a product detail page add a product to her cart.


A benefit of the gift bot is the free form input box, Notes says. This box, where shoppers input other interests of the shopper that were not previously listed, is proving to be useful for shoppers to get in the gifting mentality and to find a gift unique for a particular person, Notes says.

“We got them in the mindset of thinking about that person and thinking about their interests, and jogging their thoughts,” Notes says. “We’re seeing more niche interests and all sorts of interests in search terms that we weren’t seeing in our normal search.” For example, consumers are adding interest such as tea, unicorns and Texas, which consumer don’t normally input into the website’s search box, he says.

These niche interests, and the other inputs, such as “I am a mom buying for my baby,” gives UncommonGoods valuable insights into its shoppers that it would otherwise wouldn’t have, Notes says. One surprising thing the retailer found, Notes says, is that there are many teens using the gift guide buying for other teenagers, even though the retailer’s typical customer demographic is a woman in her 30s. The retailer plans to incorporate more of this information into its bot over time, Notes says.

“This is a rich set of gifting data we didn’t have before,” Notes says.


UncommonGoods has had a gift finder on its site before, however, it required a lot of manual tagging on the back-end and the gift finder didn’t learn over time, as its new bot does, Notes says.

“We wanted to give shoppers different ways to give gifts, move faster with gift giving and give gifts in smarter ways,” Note says. “For us to be a leader in giving, it only makes sense for us to have the best gift finder we can.”

The retailer started developing the bot in February of this year with a team of  six employees and launched the bot on its site in July. “It was up for 36 hours and then it crashed the site because the processing required was too heavy,” Notes says.

The retailer had to re-engineer the backend technology to store data in servers and the cloud instead of saving information to the user’s session. Sunny has not crashed the site yet, Notes says.


UncommonGoods is promoting the gift bot on its home page, at the bottom of every product page and via Google search ads and Facebook ads, he says.