Amazon's Monthly Active Users (MAU) have been steadily increasing for the past six months, indicating robust user engagement. Nevertheless, the average sales per user has consistently declined during this period. What might be the reasons behind this trend?

Context

Before we dive in, let’s first take a step back and make sure we understand the context surrounding the problem. Amazon is an online ecommerce behemoth where users can shop for whatever they might desire online and have it shipped to them swiftly. Amazon’s mission is to be the world’s most customer centric company, which is why if sales are dropping we should be concerned.

Clarification

In order to figure out what is happening we need to accurately understand the situation.

1. How are we defining MAUs? What is the minimum threshold of activity a user must complete to qualify? Does simply loading the Amazon.com website and then bouncing immediately still count? I’m assuming we are talking about shoppers here.

2. Similarly, what exactly do we mean by monthly average sales per user? Are we looking at the number of transactions or the value of those transactions?

3. What is also missing here is a sense of magnitude. To what degree did MAU increase and sales decrease? Something like a 2% fluctuation might be attributable to natural variance but for the sake of this question let’s assume the changes we are observing are statistically significant. 

4. I’m also just going to make the assertion here that our metrics collection and reporting are accurate and bug free.

Honing in on the issue

1. Has this been a gradual change over the past six months or do we see any sharp spikes or drops for either metric?

2. Ecommerce is very impacted by seasonality. If we look at the same time period for the year prior do we see the same trend? What about the prior 6 months?

3. Amazon is a huge platform selling all kinds of goods and serving all kinds of people. Is there a particular product category or buyer profile for which the trends are either more or less apparent? 

4. Are the observed trends consistent across technological variables like device type, OS, browser, or version?

5. Is this specific to any particular geography or region?

Potential Causes

We’ll need some answers to the above questions to help us drill down further and brainstorm some potential causes. For the sake of this mock, I’m going to take some of the context I saw in another answer: Home Category and US Market based. Here are some potential causes: 

1. Interest rates have been steadily climbing over the last 6 months and the housing market in the US has cooled off. If people aren’t purchasing homes there is going to be less of a need to purchase goods for that home. The only caveat here is that we would still need to explain why engagement has gone up despite that? Perhaps users daydreaming or just window shopping for what they will / would put in their house when they do buy one someday. 

2. The increased engagement but decreased purchases tells me people are coming to the Amazon Home goods category looking to buy something but then not buying it. 

   1. We could have increased our prices in the last six month and given a lot of user a little bit of sticker shock when they were expecting something affordable.

   2. Additionally, competitors like Wayfair could be offering discounts or better quality goods at a better price as of the last six months.

3. If we or our vendors are having supply chain issues that could be an explanation for what we are observing. Users could come to Amazon Home goods, find what they are looking for and at an agreeable price then go to check out and see it won’t arrive for 2 months then say nevermind.

4. We could have done a Home Goods and US centric beta release testing a new feature or UI design that could result in the observations we are seeing. For example, what if we are beta testing a new UI design that prioritizes the Browsing & Discovery portions of the user experience over the Checkout process.

Next Steps

We would want to go about validating our list of potential causes above to figure out which one, or combination, is the culprit. Here is how we could validate the above potential causes.

1. Get market research on US based Home goods purchases in the last six months to see if the decrease in purchases is endemic to just us or not.

2. Price check our offerings with competitors’ offerings. As well as look at the historical pricing for our offerings to see if our prices have gone up in the last six months.

3. Examine product availability rates and delivery times then compare them to the prior six months.

4. Check in with other product teams to see if any of them have any beta tests in progress or new releases which look like they might be impacting this area.

Summary

Through inquisition we’ve been able to establish that the trend we are seeing is endemic to the US and the Home category. We then proposed several potential causes for why we might be observing what we are and how to validate whether or not each potential cause is the culprit.