Imagine that you’ve been reading product reviews of dog sofas for your pooch Luna on the Wayfair website. You move a single product, a deluxe couch that includes throw pillows, into your online shopping cart. Then you notice a “frequently bought together” recommendation for a patio umbrella.

Hmm. Is this item meant to be left outdoors, where Luna can sun herself while taking a break under the umbrella if needed, or is something funky going on with Wayfair’s Frequently bought recommendation algorithm?

Or on the Home Depot site, you put a drill bit in your shopping cart and then notice that the frequently bought together (FBT) products section seems a little wacky…it’s suggesting Giant Yard Pong or a rose bundt pan and some other barely related — if related to drill bits at all — items. All you wanted to do was fix a loose shelf, and now you’re focused on having a garden party and baking a cake. And Home Depot has everything you need for that.

Or maybe you’ve had your own frequently bought together ecommerce experience in which seemingly random things were being suggested to go with what you were considering.

You do a double take, and then start to wonder about these supposedly relevant products. Is the ecommerce site’s algorithm for suggested product bundles broken? Has the overworked bot had a mental breakdown? Is the machine that’s supposed to be learning instead snoozing at the switch?

Or maybe you’re actually the poor sap with the intelligence issue — maybe the algorithm is truly, deeply knowledgeable about which seemingly unrelated items are routinely grabbed up by shoppers in the same fell swoop because it has visibility into a ton of mind-boggling data. If that’s the case, then maybe you actually do need some yard pong in your life but just haven’t realized how badly.

Questionable judgment isn’t out of the question

It’s entirely possible that it’s not you. Online stores have been known to pair head-scratcher items in an effort to upsell & cross-sell (a Reddit string vouches for this). Who knows how or why such matching gets perpetuated, but it happens. If nothing else, it makes you start to think and try to guess why the recommendation engine linked the two items.

Regardless of the fact that odd associations can and do happen, if you’re an online retailer, you wouldn’t want to be caught dead making this sort of product recommendation faux pas.

If you have any awareness of how FBT works at all, you might also be smirking and thinking that FBTs are pretty straightforward and hard to mess up. After all, shampoo + conditioner, hiking shoes + backpack? If your shoppers already have green eggs in their cart, they really should consider some ham. How tough could that be?

A better question, though, would be what percentage of your online sales come from frequently bought together recommendations. If the answer is “None” or “I have no idea,” or it’s a very small percentage in any case, read on.

Amazon strikes again

Maybe you’ve heard Amazon’s well-circulated statistic: 35% of the site’s sales are generated from its recommendations created by its advanced algorithm. If Amazon can do that, it’s a good bet that other ecommerce retailers can use frequently bought together recommendations to improve their user experiences and earn more as well. The technology is widely available; you just have to know the best way to implement it for your industry and prospects.

The goal: much higher returns

The frequently bought together product recommendation strategy relies on collaborative filtering. It works within the context of either a single item being viewed or a combination of products (e.g., what the person has stashed in their cart as they perhaps continue browsing other categories and items before reviewing their cart page and going through checkout).

And if you’re a savvy retailer, you know that if you can get people to believe they would be happier if they were to buy not just one associated item but a whole “set” or “kit” of conceptually connected recommended products in order to assuage their FOMO, you can likely move your conversion rate needle considerably higher. 

And that’s the whole point of frequently bought together (a.k.a. Frequently purchased together): big-time increases in ROI. Making the right recommendations to enhance your shoppers’ buying joy can lead to a bunch of upsells (and maybe also some cross-sells). 

Who’s looking at what?

A typical ecommerce frequently bought together section suggests items whose product pages are frequented by other customers. Like other types of online recommendation, upsell features are predicated on the assumption that often-bought-together items complement each other, so offering additional related things to consider may bolster your cart sizes and average order values.

At first glance, a retailer might think all they have to do is sort all those other items according to the number of times people have placed them in their cart with their main item, and then get set up to suggest single-click add-ons from the top of the product list in an easy-to-notice widget.

Logical assumption, but it’s not that simple. 

Want fries with that?

Why is it more complex? Because some everyday items are naturally frequently bought together (think coffee and donuts).

 You want people to buy more than what they might already add to their cart. Without paying enough attention and having a higher-level strategy, on autopilot, you could be suggesting natural matches as frequently bought items. Obvious-item suggestions can hog the limited upselling space on the product detail or checkout pages so that items that might be technically less popular but have a stronger upselling connection to the product can’t push their way in for viewing.

So to get people to click the Add-to-cart button and then buy their accumulated collection of items so you can thereby increase conversion, there’s a solution: an effective frequently bought together app or site feature balances upsell items that are generally popular (and technically most often bought together) with others whose product pages may be less frequented but which are still winning buyables from a compatibility standpoint.

You can address this problem by maximizing your frequently bought recommendations for quality. For instance, Algolia Recommend looks at conversion events on the path to the goal of the user making a purchase. Two or more items are considered to have been bought together if the same shopper registers conversions with them on the same day.

How to frequently get it right

Nailing your frequently bought together recommendations means taking a look at the details and fine-tuning these elements as needed:

• Your understanding of what will make FBT likely to succeed on your particular website. Now that you’re almost finished reading this post, you’re off to a good start.
• Your algorithm. The ability to make intelligent related-item recommendations that resonate with shoppers as promoting high-quality items worthy of their attention and ownership is key.
• Your customers’ shopping experience. Your FBT recommendations must appear in the most optimal places for catching shoppers’ eyes. Your shoppers must find being on your site a pleasure. If you can provide great frequently bought together recommendations while ensuring that the customer journey is enjoyable and stress free, you’re ready.
• Your overall related-product strategy and implementation, which should be oriented toward creating strong user engagement.
• Your developer know-how. Here are some technical tips to help your team successfully apply your frequently bought together approach.
• Your patience levels. It may take some time, effort, and annoyance to find out exactly which FBT techniques work and resonate with your users. Your persistence will be worth it when your AOVs start to skyrocket.
• Your options. You don’t have to go it alone. If you’re worried about getting bogged down in this area in which you may not be well versed, consider the upside of teaming up with an online purchasing recommendations expert.

Frequently good together: Algolia and online retailers

If you determine that you aren’t up for ascending the learning curve on your frequently bought together solution, if it’s too complicated or your developers and support team don’t have time to work on it, we hope you’ll consider the advantages of adopting our proven product, Algolia Recommend, instead.

Our powerful API will let you quickly, seamlessly implement search on your website and mobile applications, including offering excellent frequently bought together functionality for your product types.

Get set up now to start driving higher ROI with frequently bought together recommendations. Contact us and get rolling!