Discovering the Hidden Gems: Dimensional Analysis for eCommerce Beginners
Today, we're going on a hunt in eCommerce. No, we're not looking for hidden discounts or secret sales. We're searching for something much more valuable: real insights about our customers that can help us make their shopping experience better from the moment they step into our (online) store. This is what we call "dimensional analysis" . It sounds a bit technical, but really it’s just about getting to know our customers better by looking at their browsing habits from different angles.
Why the Beginning Matters
Imagine walking into a store. What makes you stay? A friendly greeting, easy-to-find products, and maybe a sale sign or two, right? For online shops, the "hello" is our website, and the friendly greeting is how quickly and easily customers can find what they're looking for. That's why we focus on the top of the shopping funnel, which is all about making a great first impression.
Rate Metrics: Our North Star
In part one we took a look at the northstar “rate metrics” related to website engagement. These metrics will also serve as our northstar while looking at deeper analysis. To recap we have:
- Engaged User Rate: It tells us what percentage of visitors are really interested in what we have to offer. The higher this number, the better. Think of it as customers not just walking by, but actually stopping to look at our products.
- Product View Rate: This one's about how many visitors are looking at individual product pages. If lots of people are checking out a product, it's like a crowd gathering around a display in a real store—it's popular!
- Category View Rate: And this tells us how many visitors are browsing through different categories. It's like customers walking through aisles in a physical store, showing interest in different sections.
The Many Dimensions of Shopping
Now, dimensional analysis might sound like rocket science, but it's really just looking at our shoppers from different angles. We have all these different "dimensions," like whether they're new or returning, what country they're in, what time they shop, and what pages they visit first, second, and last.
Let’s take a look at some common dimensions we use in analysis with stores today:
New vs. Returning
If we see a lot of new faces, that's great—it means our ads or our search engine game is strong. But if we see a lot of the same folks coming back, that's even better—it means they liked us enough the first time to return.
Fast Loading Pages: Cache Hit Rate
We also look at something called the "Cache Hit Rate." This is like a fast lane for our website. When it's high, it means our pages load super fast, and who doesn't love that?
Shopping Around the World: Country
Then, there's the country. It's pretty useful to see where in the world people are shopping from. If we have a lot of shoppers from a particular country, maybe we could start showing them products that are popular there, or there is an untapped market for ads as we were unaware so much engagement was coming from organic international traffic.
Timing is Everything: Time and Date
We also look at what time and date people shop. Maybe we get a late-night crowd that loves to shop at midnight sales, or maybe we have early birds who like to browse at dawn.
The Shopper's Journey: Pages
And let's not forget about the pages they visit. The first page is like the store's front door. The second page is where they go next—maybe straight to the sales rack? And the last page is like the checkout counter. Where they leave tells us a lot about what they liked or didn’t like.
Finding the Needle in the Haystack
Dimensional analysis of the data can often be a hard process to master as the amount of data and combinations make it difficult to find where to start. Often it may require testing and retesting combinations of dimensions across to find something that sticks out.
Another avenue is to have a machine do the work for you, just like Edgemesh Analytics can do. Instead of looking for those outliers manually, Edgemesh can try almost every combination of dimensions on demand to find the ones that are most influential for a northstar metric, like engagement rate. Below is an example of what the machine found for us.
- Android users who clicked on ads: 3467 out of 117662 people
- Engaged User Rate for this group: 38.27%
- Engaged User Rate for all visitors: 26.06%
So, what does this tell us? It says that when Android users find us through ads, they like our store—a lot. This is super helpful because it means one area we should think about to improve bottom line store performance is in showing more ads to Android users since they appear to engage at a higher rate than the average visitor.
The Story Behind the Numbers
So, what's the big deal with all these numbers and dimensions? They tell us stories—like which ads bring the friendliest visitors, what products are like magnets, and what times are like shopping festivals.
Imagine if we knew that people who land on our new sneaker page tend to buy more than those who land on the homepage. We'd probably want to make those sneakers the first thing everyone sees, right?
Or what if we found out that shoppers from Canada love our winter jackets, and they usually shop on snowy days? We might start showing those jackets to all our Canadian visitors as soon as the forecast says snow.
In the end, dimensional analysis is like being a shopping detective. It's about finding clues that help us make our store the friendliest, most interesting place it can be. And when we do it right, not only do our customers have a great time, but they also bring their friends, and our store becomes the go-to spot for a great shopping experience.
This process will be a common practice in optimizing any store and conversion rate. We’ll continue to do this as we continue further down the sales funnel…
