Hack87.Use Cross-Sell Data to Sell More Products


Hack 87. Use Cross-Sell Data to Sell More Products

One proven strategy for increasing your average order value is cross-selling products when a visitor is committing to a purchase. Your web measurement application can provide great insight into who best to cross-sell to, provided you know where to look.

The actions that visitors take on your site represent the most valuable data available for determining product offerings and pairing. For example, on an apparel site, certain sweater/shirt combinations may sell particularly well together. By presenting these articles side by side, visitors will be more likely to purchase both, streamlining the shopping process and increasing average order size. By being able to intelligently recommend items that go together, brand owners can provide customers with valuable information at the moment when they are making their purchase decisions, while at the same time working to improve their bottom line.

Before examining the uses of product affinity data, however, it is worthwhile to define the difference between a cross-sell and an up-sell.


Cross-sell item

A complementary product that is purchased at the same time or immediately after a sale. For example, if a customer were buying a video game platform, a good cross-sell would be a few video games, even though he intended to buy only the game console.


Up-sell item

A product that a vendor persuades a customer to purchase in addition to the products that they are already interested in. For example, if a customer were planning to buy a video game platform, a great up-sell would be a plasma television to enhance his new gaming console.

The two definitions are similar, but there are important differences. With a cross-sell item, a customer is showing an interest in two naturally related items and the web site must be able to help them easily find both. With up-sell, however, the customer has shown an interest in a single product, but can be influenced to purchase more based on recommendations he sees on the site.

6.8.1. Quantify the Value of Product Recommendations

Before investing in a product recommendation initiative, it is important to first be able to quantify the value of cross-selling and up-selling products on your site. This not only ensures that you prioritize your projects appropriately, but also provides you with a baseline for understanding where to make future improvements.

6.8.1.1 Step one: Track cross-sell links separately.

The first step to understanding the value of cross-sell begins with tracking. Do you provide cross-sell recommendations on product pages? What about special cross-sell categories or checkout cart links? For each of these different types of recommendation, you should look to track the behavior of visitors that view and purchase product as a result of clicking on cross-sell links. By isolating these behaviors, you can understand the impact of cross-sell and identify areas for improvement.

6.8.1.2 Step two: Measure cross-sell performance.

There are two basic types of measurement to understand for cross-sell: link usage and sales. Link usage measures how easily visitors are able to find a cross-sell link on a page. One way to measure this is by calculating the frequency of link click-through relative to total site visits. Cross-sell sales measures how much your investment in cross-sell is paying off, providing insight into the relevancy of your recommendations.

6.8.1.3 Step three: Take action.

Is your cross-sell usage low? Change the location on the page where cross-sell offers are located to be above the fold, or label cross-sell links more intuitively. Which of our cross-sell categories are delivering the best sales? For underperforming categories, it may be useful to reevaluate the product pairings to ensure that you are providing the right recommendations.

Now that we have defined a basic process for tracking and managing cross-sell results, we can take a look at how to use web data to improve our recommendation capabilities.

6.8.2. Data-Driven Analysis: Find the Low-Hanging Fruit

Cross-sell and up-sell are two of the most data-driven applications for web analytics. Given the sheer volume of data available, companies must begin by focusing their efforts. The best way to do this is to start with the most popular products sold on the site. By matching these products with appropriate cross-sell recommendations, you can quickly increase site sales for a large population of shoppers. To do so requires an understanding of exactly which products are relevant to recommend. The data that can be captured for this type of analysis falls roughly into two categories: cart overlap data and common buyer data.

6.8.2.1 Tracking cart overlap.

When a customer creates an order, the items that she purchases are all available for capture directly from the order confirmation page. By simply taking the top products sold on a site and matching those with items with which they were most frequently sold, data analysts can quickly understand which products are "cross-sold" effectively. These represent items where customers are interested in a bundle, since they are purchasing both products in a single session. Placing these items pairings on product pages can be an easy way to improve cross-sell effectiveness.

6.8.2.2 Tracking common buyers.

Tracking product purchases over multiple sessions provides even more insight into customer affinities. By tracking the top items sold and then tying them to the top products those buyers have purchased over time, data analysts can understand the propensity for "up-selling" a customer who is purchasing a given item. By placing these product pairings on the checkout page or in targeted emails, brand owners can easily increase up-sell performance for a site.

An example of cart overlap and common buyer overlap is depicted in Figure 6-7, which shows that 39 percent of customers who bought a basic dress shirt also bought a reversible belt. From this we can infer that the belt is a good candidate for cross-sell with the shirt.

Figure 6-7. Cart and common buyer overlap


6.8.3. Leverage Cross-Sell Data

Populating cross-sell recommendations for the hundreds or thousands of products would be a daunting task even for a large team of analysts. As a result, extending the value of cross-sell data requires more than just analysis. To deliver intelligent up-sell in a timely and cost-effective manner requires automation. By using a data warehouse to capture and store product purchases and export this information in a structured format to your online retail, call center, or email systems, you can do just that.

To enable automated, intelligent cross-sell, you must begin by thinking about the types of recommendations you are making. Will you be populating recommendations on the product page? Fueling email campaigns to recent buyers? The uses of the data will drive the type of information that is required. During this analysis, it is important to keep in mind the difference between cross-sell and up-sell. For example, when populating product page pairings, the focus should be on items that sell well together in-session (cross-sell), ensuring that all results returned are immediately relevant. With email, however, you may want to focus on non-obvious products that the customer is likely to desire, based on his historical purchase behavior (up-sell).

Brett Hurt and Eric T. Peterson



    Web Site Measurement Hacks
    Web Site Measurement Hacks: Tips & Tools to Help Optimize Your Online Business
    ISBN: 0596009887
    EAN: 2147483647
    Year: 2005
    Pages: 157

    flylib.com © 2008-2017.
    If you may any questions please contact us: flylib@qtcs.net