Discover 3.0 Fundamentals: Basket Analysis

https://blog.adobe.com/media_c30416273c6743354fa37bda81f75bb2c2292a99.gif

The retail industry experts in Adobe Consulting continue to share a series of analysis quick wins for retailers, using Adobe Discover 3. For a limited time, Adobe SiteCatalyst 15 clients can inquire with their account team and ask to take part in a free trial of Adobe Discover. We’ve made it easier than ever to try Discover, and we’re showing some great Discover analysis opportunities specific to the retail industry. For more information and to request trial access, contact your account manager or account executive.

Adobe Discover – Retail Quick Win #6 Basket Analysis

Discover is a powerful segmentation tool that allows you to easily slice and dice your data in new and revealing ways. It allows web analysts to easily create new segments on the fly and see the effect on any SiteCatalyst report. This type of data exploration allows business users to answer advanced analytics questions almost as quickly as they can be asked.

One interesting type of analysis you can do with Discover is basket analysis. Basket analysis allows you to understand the product affinities for different types of customers and what they are purchasing together, so that you can optimize and target your marketing more effectively. You can also identify opportunities for cross-selling and upselling that you may not have seen in the SiteCatalyst reports.

Discover easily enables this type of analysis because you can create a segment for a type of customer or type of order and then layer onto a product or division report. For example, what percentage of sales for a given division is driven by email vs. display advertising? For customers who purchase accessories, what other types of products are they purchasing? The answers to these questions can help drive marketing and optimization strategy on your web site.

Let’s take a look a few examples of basket analysis using Discover:

Basket Analysis Example 1: Mobile Customer Order Mix

Question: Is the mix of products different for customers who purchase on their mobile device? What products are they more or less likely to purchase?

Potential Action: Based on the results, you may want to consider displaying different marketing or product mix to mobile customers, or modify the mobile web experience in order to improve mobile conversion.

Analysis approach:

Identify/ segment mobile customers: You may have a separate experience for mobile customer, or even a different site where you redirect customers. You can create a segment in Discover where “mobile device type equals mobile phone”. You could also create a separate segment for tablets if you wanted to break that out separately.

https://blog.adobe.com/media_5a8d395617b7bd500b47e65367401ddb63837f0e.gif

Run the desired report you want to segment. For example, the Products report or Division report. Then add in the mobile segment.
Look at the percentage of mobile sales for each product or division. Are there any outliers?
You can also drill into a given division to identify specific products that are affecting the results, or trend the results over time to see if the results are consistent over time. To get a quick view of the trend, just click the green magnifying class next to the number in any cell in the report. See the image below for an example:

https://blog.adobe.com/media_1ba66587208a4edc6e0720d96be4145ebad4cfca.gif

In the sample data below, you can see that the average % of mobile orders is 11%. However, you can see that Electronics is 22% of mobile orders. Therefore, there may be a stronger affinity for mobile shoppers to purchase electronics than other types of products. You may want to explore that theory by testing Electronics creative for the mobile visitors.

https://blog.adobe.com/media_ee2a2bec08594148509fb3ac5a17c24aaec899bf.gif

Basket Analysis Example 2: Product Affinity Analysis for Jewelry

Question: How many orders contain Jewelry? Is Jewelry purchased alone or with other products? What other types of products are being purchased with Jewelry?

Potential Action: Identify opportunities for cross-sell and up-sell for Jewelry that can be delivered via existing email segmentation, on-site marketing or even through product recommendations engines.

Analysis approach:

Create a segment for “orders where Division equals Jewelry” (or whatever Division or product type you want.

https://blog.adobe.com/media_691796e614a6a8612e4e587e2bdfd5412254c83b.gif

Run a products or division report and layer on the segment.
Copy the data to Excel where you can create calculations such as percentage of orders which contain Jewelry and percentage of other Divisions which also contain Jewelry.
At orders, revenue and units, you can calculate average order value, units per transaction and average unit retail for each of the segments.
In the sample data below, customers who purchased Jewelry also purchased Women’s items 13% of the time. Surprisingly, they also purchased shoes 14% of the time. This may be an opportunity to test a merchandising message that features shoes with jewelry.

https://blog.adobe.com/media_573a4d6771626c8c9810f6a9eb6e5edd275b4c8c.gif

Tips and Tricks:

The results of Analysis are often best shared with others via Excel. My favorite button in Discover is the one-touch-copy button that copies the data which you can paste in Excel. I recommend keeping the format on one sheet then just referencing the data in another sheet where you do your formatting. That way, it’s easy to refresh the data—just copy and paste.

Once you’ve identified a segment you may want to share, it’s easy with SiteCatalyst v15. Just create a global segment in SiteCatalyst where you can create a dashboard that contains the info you want to share. Additionally, Report Builder can also use segments to create automated reports that are emailed out to your business partners.

Next Steps:

I have outlined just a couple of ways to do basic basket analysis using Discover. Hopefully, this will give you some idea of the power of Discover to explore product affinities and customer behavior on your site. Another idea could be to look at what products customers view, but don’t necessarily purchase before buying other items. For example, what other brands did a customer view before purchasing a given brand. This could again lead to potential cross-sell and upsell opportunities for the marketing or merchandising team.

https://blog.adobe.com/media_32f6b6a3605b47d21d42522ff1ece5f4cbb8b8e4.gifDavid Yoakum is a Senior Consultant in Adobe Consulting, focused on digital strategy, analytics & optimization for retail & travel clients. He tweets at @davidyoakum.

If you’re an online or cross-channel retailer using Adobe SiteCatalyst 15, you should try these Retail Quick Wins in Adobe Discover. We’ve made it easier than ever to experience a free trial of Discover. For more information and to request trial access, contact your account manager or account executive.