Product recommendations can boost average order value, or even produce a sale when a shopper originally was just browsing with no definitive plans to buy. The idea is simple: Based on what a retailer knows about a shopper, or shoppers overall, through past purchase history and e-commerce site behavior, a system displays products it believes a shopper might like. But do product recommendations work?

In recent research it conducted using results from more than 100 of its retailer clients, product recommendations and personalization technology vendor MyBuys Inc. says they do-and then some.

On average, the retailers, 40 of which are in the Internet Retailer Top 500 Guide, realized a 915% increase of the overall site conversion rate on products recommended in displays in the shopping cart. There was a 474% increase on recommended products displayed on category pages, 411% on product pages, 307% on order confirmation pages, 248% on home pages and 192% on search results pages.

When product recommendations are based on solid customer data and show the most relevant products, they can be an effective tool for boosting conversion, average order value and sales, says Lisa Joy Rosner, vice president of marketing at MyBuys. The shopping cart is an especially effective page on which to display recommendations, she adds, because it showcases other products next to what a customer already has chosen to buy and because it is the perfect place for impulse buys.

When you’re looking at a pair of pants on a product page, it’s good to see other pants possibilities. But in a shopping cart, once you’ve selected your pants and a shirt, it’s enticing to see belts and shoes that go with the pants and shirt. It’s like candy at the checkout counter, Rosner says. It’s not that people simply buy products from the cart display, it’s when you put the right products at the right price point in a personalized manner and in the context of what they’re buying-that’s why we are seeing a 915% conversion rate increase in the shopping cart.