Cross-Selling and Cross-Promoting Products

The idea of cross-selling basically means offering people alternatives and additions to the current product they are considering. The most common way of doing this is by presenting manually-assigned “related products” somewhere on the site’s product details pages. These may be noted as “related products” or “similar products”, but might also be manually-assigned “accessories” or even more generic recommendations under the titles “May We Suggest” or “You May Also Like”.

But there are many other ways to cross-promote and cross-sell products to your customers. If you have a large number of products, it’s often easier and can be even more useful to use algorithmically-based decisions when cross-promoting products.

Customers Who Bought This Also Bought…

Automatic product recommendations, such as the example above from Walmart.com, are most often based on products that were purchased at the same time by the same customer. You often see this labeled on websites as “Customers Who Bought This Also Bought…” and then a list of the products that were commonly purchased along with the main product being viewed. This type of solution is implemented on an e-commerce site by keeping track of products that are purchased together within the same order. For product A, the site typically recommends the 3-5 most-commonly-purchased-at-the-same-time products. Obviously, your number of orders needs to be high enough to justify reliable statistics for these to work.

Another way to display products from this data is to group items into a bundle and offer a discount if both are purchased. Amazon does this, and although it used to be labeled “Frequently Bought Together”, I noticed today that the headline has changed to Best Value:

People Who Viewed This Also Viewed…

As shown on Kmart.com, this is similar to what we just described, but instead of basing the statistics on the number of times items were purchased together, this is based on how many times items were viewed in detail in a given shopping session (or for a given customer account over multiple shopping sessions, depending on the implementation). The results are somewhat different, because this group of product recommendations includes more similar products that customers may view before deciding on a single product (comparing two pairs of shoes before deciding which one to purchase).

This type of solution is also good for stores with lower order volumes. Because most e-commerce sites have a fairly low conversion rate (2% is the oft-cited average), it can be easier for some merchants to get good statistical data based on product views instead of product purchases.

People Who Viewed This Ultimately Bought

Target’s website uses a middle-ground between browsing data and purchasing statistics. They show which top products were ultimately purchased by customers who viewed the given product. This is a slightly more complicated set of heuristics, as it takes into account both product (and, I presume, “Quick Look”) pageviews as well as sales statistics.

Items in the Same Category

Customers may also be shopping for multiple items by theme (such as party supplies) or for a single item among similar alternatives (such as laptop bags). In both of these cases, being able to see what categories a product belongs to is ideal, as it allows the customer to navigate up the website’s hierarchy one level to view other similar products. This is most commonly done with breadcrumbs, which show the path from the homepage, down through each more specific category, to the category page. The following example shows a basic, clean breadcrumbs trail from Lands’ End:

For sites that have products assigned to multiple categories, there may be more than one breadcrumb displayed, such as on Restaurant Equipment Solutions:

If the site has products in many categories, it might be more useful to display the breadcrumbs for the main category, and links to additional categories in another location, like SmallToys does:

Which Approach is Right for Your Site?

When trying to decide which approach is right for your site, consider the following:

1. Time: Manually-assigned products takes the most time to manage, especially if your product line changes regularly. Do you have enough staff (or enough of your own time) to assign related products to each item in your catalog? Do you have the time to maintain these relationships when new products are added, or old ones deleted? This can be a great approach if you have the resources, but make sure you can keep up. An empty “related products” box looks really bad!
2. Volume: Does the volume of visits and/or orders support “also-bought” or “also-viewed” automatic statistics? And can you obtain this feature within your shopping cart, either natively, via a plug-in, or by integrating a third-party product recommendations solution?
3. Categorization: Do breadcrumbs make sense for your site? If it’s flat (only a single level of categories, with no subcategories) then they can be fairly useless. If you have TONS of categories, it might help to use multiple breadcrumbs, or breadcrumbs in conjunction with “Products In the Same Category As” displays. Try to find the right balance between making it easy for the customer and not overwhelming them.
4. Pricing – Make sure your margins can support discounts on bundled items.

Where to Display?

Typically, products are cross-sold on the product detail page as related products. But two other great locations for this display are on the view cart page (“Customers Who Bought Items in your Cart Also Bought”) and within the customer’s wish list (“Customers Who Bought Items on Your Wish List Also Bought”). These pages are closer to the decision point, and therefore can have a higher impact on either upselling the customer a more expensive product, or convincing them to add a similar product in addition to the ones already being considered. Another location is to make product recommendations on the customer receipt page (after the order has been place) or in the confirmation email. In fact, if your processes and software can support it, give the customer a short window during which she can purchase additional items to ship with the original order, with no additional shipping charges.