We know that in search, not all things are created equal. Not every tactic will yield the same results. Not every link is of equal value. Not every keyword you rank for is valuable. We tend to spend an inordinate amount of time thinking about ranking strictly at the domain level.

We're looking at a site, determining its Domain Authority, trying to determine if it's going to be easier or harder for that site to get visibility.

This principle is invaluable and is still considered best practice. But we don't think about the fact that groups of pages on a site can have different ranking opportunities. This is especially true for e-commerce sites with a large number of SKUs. For example, Walmart offers a huge number of products in its store, from clothing to grills to trampolines. It's unlikely that all of these product groups have the same chance of ranking well just because Walmart has high domain-level authority. Some of these products may have more backlinks at the page level, some may belong to less competitive industries. Google may see Walmart as an authority at the topic level for some of these. One of the interesting types of analysis we've done for clients with a large number of page groups is designed to answer the question, "Which of our pages are most likely to rank well?".

The answer to this question can help determine the direction of your SEO or content campaign. It can show you potential quick win opportunities for categories where you can create new content and have a better chance of ranking. It can help you identify low-performing groups and see where more time and investment is needed. So how do you do this? I'll show you our process for answering this question using keyword segmentation. Please note: although we're using e-commerce sites as an example, with proper labeling, a similar analysis can be done for many different types of sites.

What is keyword segmentation?

One of my favorite digital marketer sayings comes from the legendary Avinash Kaushik, who once said, "All data in the aggregate is crap." While he was referring to web analytics, he could very well have been talking about website ranking as well. While it's certainly useful to look at overall ranking data, being able to segment it into more meaningful groups can help provide much more insight. While we often think about Google Analytics segments, we often don't apply the same logic to our ranking data. For this analysis, we need to find a way to segment our keywords into different groups. Fortunately, there are tools that allow us to do this very easily. STAT is one of the most robust ranking tools that allows you to easily create custom keyword segments. Other tools like Moz Pro and Ahrefs can also help you segment keywords. In this article, I will be using examples from STAT.

1. Upload your keywords

The first step is to load keywords into STAT.

Since STAT automatically finds URLs that rank for a given keyword, it should do this for you. Once you see this data, you can get to work.

2. Segment keywords into relevant groups

The next step is to segment your keyword list into relevant groups. Most eCommerce sites tend to want to segment them by product category (clothing, outdoor products, electronics, etc.). That is the example we will use here. However, you can segment them by many other things. Categories, search volume, potential revenue - you can use all of these. Here is an example where we grouped the pages of a bedding website by different keyword segments. This allows us to see the rankings for each specific group. For example, clicking on down pill ow might show us the rankings for down pillow pages: However, clicking on the down comforter group allows us to get rankings data for a separate keyword group: Here we can see that two very different stories are happening on the same site.

The ranking of "down pillows" is rising and "down comforters" are falling. Segmentation allows us to easily see this.

Creating segments at scale

If you work for a large e-commerce site, you might think that creating segments alone would require weeks of work. If you're tracking 10,000+ keywords, creating meaningful groups seems like a monumental task. Fortunately, many ecommerce sites have already properly taxonomized pages using internal breadcrumb links. These internal breadcrumb links have already done the categorization for you, so all you need to do is extract the breadcrumbs from your site and link them to the keywords you want to segment, all of which can be easily implemented at scale using Screaming Frog's custom BreadcrumbList schema extraction. Since the breadcrumbs categorize your site's products, we can use the same categories for STAT segments. The goal is to create keyword segments based on the naming conventions of your site's breadcrumbs. Open the Screaming Frog custom extraction and toggle the drop-down list to "Regex". You'll need to add the following regex (many thanks to Brian Gorman for creating it):

"position":(\d+)(,)"item":\{"@id":"(.*?)"(,)"name":"(.*?)"\}

This should extract the position, element, and name of the breadcrumbs associated with each URL of your site. For example, when looking at the REI site, we would get this example:

You'll have to do a bit of formatting, using text-to-column conversion and concatenation. But you should be able to make your spreadsheet link each URL to the appropriate categories:

However, in order to load this data, you need to link each breadcrumb at the keyword level. This can be done by exporting existing keyword ranking data. Export the keyword data and make sure you export the "ranking URL" of each keyword. You can then use VLOOKUP to associate the correct breadcrumb to each keyword. You should end up with breadcrumb related keywords.

3. Keyword reloading with related tags

The new CSV file should have a list of all keywords in one column and all associated tags in another column. Reload this list of keywords into STAT to import the new tags into the database. I highly recommend using the bulk upload feature in STAT. If you have already added these keywords to STAT, don't worry. STAT will not add them as duplicate entries, but will simply apply the new tags to the keywords you are already tracking. Alternatively, if you've been tracking your keywords in STAT for a while, you can take advantage of the powerful "Backfill" feature.

In this case, STAT will show historical data for all the new tags you just added, and you won't have to wait for the tool to collect new data. We highly recommend doing this if you already have keyword ranking data for your tags.

4. analyzing ratings by category

Now that we have all the data imported and segmented into STAT, we can start analyzing to find out how our product categories are performing compared to each other. Create a new spreadsheet. Label the left column "Category" and add all the product categories you want to analyze. Label the second column "Average Rating." You will need to get the average rating for each category. You can do this by clicking on each individual label and viewing the "Average Rating" metric in the "Dashboard" tab. Finally, you need to calculate the average rating of all product categories. Create a third column called "Difference from Average Rating".

For example, if we look at the product categories within the Home & Furniture section of Walmart and we know that the average page rank is 18, we can get the following result. In this view, we can see that Walmart performs well for products belonging to the categories Bathroom, Decor & Accents, and Kitchen & Dining. Adding more category pages and products belonging to these categories may have a better chance of performing well. Conversely, Walmart does not show high results for Wall Art, Mattresses and Rugs compared to other pages on the site. If these are high-priority SKUs, Walmart may want to rethink its SEO strategy for these pages and prioritize supporting this content. If you're more interested in competitor data, you can look at your product groups through this prism as well. You can load the same keywords and tags into STAT and track rankings on a competitor's site rather than your own. This will provide insight into how each product group ranks compared to the others. For example, here's an example of the same analysis, but taking into account how Walmart's categories rank against Bed Bath & Beyond's categories.

Conclusion

Overall, the purpose of this paper is to help you drill down into product segments to determine what works well and what doesn't work well compared to the rest of your site.

This provides a more comprehensive view of visibility than just a general view of ranking data.

This analysis will be especially useful for sites with large inventory or groups of pages, where segmentation may be necessary to properly evaluate ranking performance.

This kind of analysis not only provides short-term information, but also allows you to make long-term decisions about where to invest in SEO and content.

Read More Press

View All