When tracking your data, it’s important to set yourself up for success. E-commerce and retail companies want to use their data to understand why some customers fall out of their funnels or why customers become repeat buyers. They want to understand the key lifecycle events that lead up to the sale of a physical good, so they can, for example, test whether personalized shopping experiences yield higher conversions or build a multi-channel cart abandonment campaign. But first, they need to make sure those lifecycle events are being captured in their datasets.

That’s where a tracking plan comes in. A tracking plan is a living document that can be used across your organization to record what events and properties to track, where you’ll be tracking them in your code base, and why you’re tracking them.

In this guide, we’ll share with you the core events most relevant to e-commerce companies that can get you started immediately in understanding your customers and driving sales.

Talk to a product specialist today about building a clean, high-quality data spec so you can focus on brand engagement and sales growth.

Identifying your customers

Before diving into specific event tracking, you’ll want to make sure you track who your users are with the the .identify() call. You should call .identify()

Traits are pieces of data that you track about a specific user. Read our guide about selecting traits to learn more.

Here are the most common user traits e-commerce companies include in their tracking:

• first name

• last name

• email

Here are a few examples of other helpful user traits:

• locale

• currency

• phone

• lifetime_value

• order_count

Here is a sample .identify() call for Segment:

In analytics.js:

    analytics.identify({
      first_name: 'Andy',
      last_name: 'Jiang',
      email: '[email protected]'
    });

In analytics-ios:

    [[SEGAnalytics sharedAnalytics] identify:nil
                                    traits:@{ @"email": @"[email protected]",
                                              @"first_name": @"Andy",
                                              @"last_name": @"Jiang",
                                              @"experiment_viewed": @"Coupon" }];

In analytics-android:

    Analytics.with(context).identify(new Traits().putValue("first_name", "andy").putValue("last_name", "jiang").putValue("experiment_viewed", "Coupon").putValue("email", "[email protected]"), null);

The main benefits of using the traits listed above are:

• You can use the traits to personalize content in your email or push notification tools by inserting their information. For example:

• You can create cohorts based on the traits in any of your tools. For example, you could use order_countto determine how many of your users are first or repeat, or experiments to determine how many have seen this particular experiment variant

Given the power you have in your downstream tools to create cohorts based on these dimensions, you may be tempted to throw more contextual data into the traits , such as UTM params, IP addresses, and userAgents. But if you’re using Segment’s client-side analytics.js library, then all of these contextual pieces of data are automatically collected.

Selecting key e-commerce and marketing events

Now that you are tracking who your users are, you can work on what they’re doing on your website or in your app with a .track() call. Tracking customer events lets you learn about your customers, measure the impact of your marketing efforts and product decisions, and proactively engage your customers in a meaningful way that drives sales.

E-commerce businesses, unlike SaaS or consumer apps that optimize for product engagement, focus on directing users down their funnels to a conversion goal, like purchasing a product. To best understand why customers convert, it’s important to track and measure all key funnel events.

• Product Viewed

• Product Added

• Checkout Started

• Order Completed

There are also auxiliary actions to track to measure your customer’s engagement with your site. These actions are good to track so you can better understand their intent on your website.

• Products Searched

• Product List Viewed

• Product List Filtered

Lastly, we have a set of semantic campaign events that are automatically collected so you can understand the conversions within these particular channels:

• Email Opened

• Email Link Clicked

• Push Notification Received

• Push Notification Tapped

• Deep Link Clicked

• Deep Link Opened

You may notice a pattern in the event names. We’ve selected the “Object Action” naming convention to ensure that all of our event data is clean and easily analyzable, while choosing “snake_case” for the traits and properties. It doesn’t matter what you choose, so long as it’s consistent. Without a uniform and enforced naming framework to guide developers that add tracking code later, your data could get marred with conflicting naming structures. Learn more about the importance of naming conventions.

We recommend tracking core checkout activity on the server-side. Learn more about tracking on the client vs. server.

Selecting your properties

Properties are similar to traits, but they’re associated with specific actions, rather than with an individual user. Each .track() call can accept an optional dictionary of properties, which can contain any key-value pair you want. These properties act as dimensions that allow your end tool to group, filter, and analyze the events. They give you additional detail on broader events.

For e-commerce, since most events are customers choosing, browsing, and checking out products, all of the traits must contain key information about the products themselves:

• product_id

• sku

• price

• quantity

• currency

These traits must be included because many tools rely on them for analysis. If there was one of them missing, the call would be ignored.

Using a specific tool and want to see how Segment handles sending calls to it? Check out the documentation.

Here is an example .track() call:

In analytics-node:

    analytics.track({
      userId: '019mr8mf4r',
      event: 'Order Completed',
      properties: {
        order_id: '50314b8e9bcf000000000000',
        total: 20,
        currency: 'USD',
        products: [
          {
            product_id: '507f1f77bcf86cd799439011',
            sku: '201',
            name: 'Folsom',
            price: 10,
            quantity: 1
          },
          {
            product_id: '505bd76785ebb509fc183733',
            sku: '204',
            name: 'Brennan',
            price: 10,
            quantity: 1
          }
        ]
      }
    });

In analytics-ios

    [[SEGAnalytics sharedAnalytics] track:@"Order Completed"
                          properties:@{ @"order_id": @"50314b8e9bcf000000000000",
                                           @"total": @"20",
                                        @"currency": @"USD",
                                        @"products": @"Coupon" }];

In analytics-android:

    Analytics.with(context).track("Order Completed", new Properties().putValue("order_id", "50314b8e9bcf000000000000").putValue("total", 20).putValue("currency", "USD").putValue("products", "Coupon"));

It’s important that these events contain particular properties, such as sku , otherwise the downstream tools won’t be able to create out-of-the-box revenue and sales reports. Learn more about semantic properties in our e-commerce spec.

Using data to understand why your customers don’t convert

The most successful e-commerce businesses not only efficiently move their customers through the funnel towards conversion, but also have the infrastructure to collect and leverage customer data. Having a tracking plan focused on key funnel events can help e-commerce businesses get a sense of the health of their funnel.

Aside from funnel health, having these key pieces of customer data gives companies the ability to tailor and personalize each interaction, as well as build marketing campaigns around actions taken or omitted.

Without taking this critical step of mapping out key customer events, businesses often spend too much time revisiting their data model or analyzing impartial data sets. Instead they could spend that time understanding and addressing customers’ needs.

Talk to a product specialist today about building a clean, high-quality data spec so you can focus on brand engagement and sales growth.

If you have any questions, or see anywhere we can improve our documentation, please let us know!