Corporate Retail's use of Alternative Data: Getting to Why

With market share data, brands and retailers are gaining visibility into "what happened?" (i.e. What is my market share? What brands are outperforming?). However, questions like, "Are we more exposed to older generations than our competitors?" or "Is our market share with new customers to the category the same as our share with all customers?" are tough to answer quickly or reliably. Explaining why performance is positive or negative is not nearly as accessible.

Historically, retailers have analyzed the behavior of customers through primary sources (i.e. data collected through the normal course of business) such as:

• Loyalty programs or membership - i.e. when you go to CVS and type in your phone number at the checkout. This data helps them understand what customers are buying and when/how that is changing. With this information, retailers can connect past purchases to the same household and provide a much richer set of analyses.?
• Web traffic - E-commerce retailers are able to monitor website traffic and see not only what customers bought and connect that to previous purchases, but they can also see what products customers looked at as a proxy for what customers considered buying, how long they shopped, etc.?With e-commerce, there is a massive amount of data that is not available to traditional brick & mortar vendors.

Most large retailers have data teams dedicated to sorting through this internal data to uncover trends and important insight that might lead to better decisions. For example, retailers might look at "conversion" (the percentage of customers that made a purchase from those who opened the product page) to monitor how that is changing across features or price points.

Although these techniques are innovative and valuable, we live in a world with ever-expanding optionality; as e-commerce evolves, consumers are changing shopping behavior continuously. Analyzing internal customer data is only one piece of the puzzle and can not fully explain the drivers of performance. To understand the customer, it is essential to understand their behavior across all retailers and channels. Further, brands do not have access to these primary sources at all, which creates an information asymmetry - retailers have the insight on customers while brands are left in the dark.?

To address this, there are a growing number of Alternative Data consumer panels that retailers and brands can license through third party providers such as:

• Receipt scan - panelists agree to share purchase receipts through email connection, physical receipts scan, or retail account connection.

• Credit/Debit transactions - panelists agree to share transactions through participation in credit card rewards programs or by connecting their credit/debit card feed to a budgeting or financial planning application.?
• Clickstream - panelists agree to share web browsing activity in exchange for rewards or services like a VPN.
• Survey - retailers and brands will partner with research organizations to define the right questions to ask and source a small panel.

Consumer panels can be incredibly powerful vehicles but acquiring the data is just the beginning.?The process in getting to why has a few key steps:

1. First, generating hypotheses on which factors could be contributing elements to your problem.?
2. Second is understanding the bias of the dataset and determining if it can be useful to address your hypotheses (I walk through this in detail here).?
3. Finally, sourcing metrics from the data to test the hypotheses.?

An example is useful here: Let’s say a cosmetics brand has lost 200 bps of market share at Amazon in the last 4 quarters. There’s an almost limitless number of potential variables and contributing factors to why customers decide to buy what they do and when. The magic happens when industry experts partner with data experts to isolate what questions can be answered with alternative data. In this case, let's say some of the hypotheses are:?

• We’re losing share because there are new brands attracting first-time customers.?
• We’re losing share because our existing customers are spending less on cosmetics at Amazon.?
• We’re losing share because competitors are taking customer spend.?

By using an email receipt panel we can aggregate some key metrics to address each hypothesis:

Each metric could then be analyzed to test the hypothesis and ultimately enable teams to take action. For example, if these metrics do confirm that existing customers are spending less on cosmetics at Amazon, then marketing investment on Amazon search optimization can be diverted into broader outreach at retailers like Ulta or Sephora.

With consumer panels, retailers and brands can make more informed, data-driven decisions quicker and more effectively. Teams can migrate from trial and error to action based decisions from data-driven theory. This could be the difference in winning market share in an increasingly competitive marketplace.???

If all these steps sound like a ton of work and experiential knowledge, well…it’s because it is. What we are doing at Yipit is building a team that works for you to answer your key questions on your unique problems. Our team is trained and experienced on the biases of these sources and the dynamics of what's possible - ultimately we want to solve these problems with you and create an environment where you can self-serve to confidently drive business value as part of your day-to-day process.