How to Conduct Keyword Research for Ecommerce from the Ground Up

High performance search engine optimization is built on keyword research, and your keyword research is only as good as your process.

Talk to many online store owners, however, and you’ll often find that keyword research feels like the most mystifying part of the entire campaign. This isn’t helped by the fact that many folks giving advice on ecommerce SEO aren’t applying the rigor necessary for real keyword research.

I do a lot of keyword research, especially for ecommerce sites. In my experience, it’s more about smart planning and reliable systems than it is about rain-dancing to appease our algorithmic overlords. The most valuable first step you can take is to identify what types of terms you should pursue, aggressively expand that list, and then decisively cull it down to what’s most relevant.

Identifying your keyword universe

There are a lot of nuances between new stores vs. existing stores, so for the sake of brevity I’ll assume for this post you’re working on a brand new website.

If your store is more established, you’ll likely already have a nice baseline of data from which you can pull to help determine the direction you want to take with your research. But for a new site, you’re going to need to lean on established competitors.

The right way to do this is to find the major players in the space that aren’t colossal brands; steer clear of Amazon, eBay, Walmart, and other established, generalist ecommerce websites. Don't be too dismissive, however, because you don’t want to necessarily stay away from the big informational brands like Wikipedia or Quora. These sites can actually be treasure troves of keyword terms and topics.

Niche to win

To compete against the 800 pound ecommerce gorillas these days, especially if you’re just getting started, I fully subscribe to the idea that you need to start with a hyper-niche. So a niche within a niche, and sometimes even a niche in that niche (it’s niches all the way down).

Highlighting the importance of specialization, let’s run through the entire keyword research process from start to finish with a real example. Examining something tangible is a good way to make the concepts we’ll cover easier to understand and apply.

Though I’m not currently a store owner, I am a sneakerhead. It’s a not so well kept secret, just ask my wife about the closet I’ve taken over. But niche to win, right? So instead of analyzing sneakers, or even a specific model of sneakers, let’s start our keyword research journey down one more level by examining a common accessory for sneakers: shoelaces.

More specifically, I’m going to look at replacement and “aftermarket” laces, i.e. the fancy standalone shoelaces you might purchase to swap out the default laces on your limited editions.

Our first step is going to Google and doing some basic searching, starting with good old common sense. All we’re looking to do right now is to see what Google comes back with from auto-suggest. For this search I typed in “sneaker laces” just to kick things off (pun regretfully intended).

Many commercial terms are more competitive than ever, so I was admittedly a little surprised to see the #1 ranking wasn’t a household brand but was, at the time of this writing, a smaller niche shop called LacesOut.net. Since a smaller store outside of the generalist ecommerce platforms fits our criteria, we’ll use it as our first example to gather data from.

But before we do that, we need to see what all the Google suggestion keywords look like. To do this, we can use a helpful tool called KeywordKeg. Let's start by entering our root keyword, “sneaker laces.”

Immediately we see that length is a common term used to specify what these searchers are looking for.

Based on the types of products we’re selling, aftermarket and replacement laces for sneakers in this case, not all of these terms are going to be relevant. We’ll want to scan the terms and look for modifiers that we can remove using the negative keywords function.

For example, I’m noticing that searchers use of a lot of brand modifiers such as eBay, Amazon, and Walmart. We want to add those into our negative keywords list to further clean up the results.

I’m also noticing a number of keywords that are material specific and that include products I don’t plan to sell like Kevlar, leather, waxed, and rubber. Let’s remove those, too, to further focus our results.

This leaves us with 137 unique keywords, so from here I want to sort by monthly search volume to get a sense of how popular these keywords are.

Before I go any further, I want to take a moment to explain what each of these columns mean in terms of the metrics they’re showing.

• Search Result: The actual keyword that is being typed into Google.
• Volume: The average number of times this keyword is searched every month on Google.com (specifically targets users in the U.S.).
• CPC: The average cost per click advertisers are willing to pay to show ads on Google when this keyword is searched.
• Comp: The relative level of competition for paid ads via AdWords.
• Value: The approximate value this traffic is worth per month. This is a metric that is calculated by multiplying the average monthly search volume by the average cost per click.
• SEO Difficulty: This is a logarithmic measure of how difficult it will be to rank in the organic results for this keyword based on a scale of 1-100, with 100 being the most difficult.
• CTR Scope: KeywordKeg describes this as “whether the organic results in SERPs get clicked on. It depends on the number of ads, images, product listings that are shown above organic results. The higher the CTR Scope the more traffic you will get from ranking organically for the keyword.”
• Keyword Power: This is a calculated metric comprised of SEO Difficulty, CTR Scope, Search Volume, and CPC that indicates the potential the keyword has for your website. The higher the power, the better.
• Trend: This is the search volume trend for the past 12 months used to judge whether a keyword’s popularity is trending up or down.

The tip of the iceberg

When looking at keyword data as part of your overall research process, it’s important to recognize that the keywords you identify based on your initial run and first attempts at filtering are just the tip of the iceberg. Understanding how your users think with research like a contextual inquiry will help you identify more potential keywords.

Any of these terms with a search volume over, say 100/month, is likely to have a whole universe of related terms around it. For example, let’s look at some of the related terms that come back for a few of the terms in my current list.

💡 Note: I’m using the same list of negative keywords, with a filter set to only show keywords that receive more than 100 searches per month, on just Google.com.

The highest volume term: “no tie shoelaces"

The highest value term: “shoe laces how to”

The most difficult SEO term: “shoelaces knots”

The keyword with the highest power: “boot with laces”

Let the data weigh in

This exercise is exactly why you always pull the data and don’t make guesses when it comes to defining your content map and keyword priorities.

Looking back over the terms generated from the 4 keywords above as seed terms:

1. “No tie shoelaces” comes back with relevant suggestions including 11 other terms.
2. “Shoe laces how to” comes back with relevant suggestions including 139 other terms.
3. “Shoelaces knots” is a dud, with only 1 other term.
4. “Boots with laces” is not really relevant to our product mix, but includes 22 additional terms and makes a strong case for creating a content page targeting these keywords.

Creating a process: rinse and repeat

You can run through the above process around 10-15 times to build substantial lists of relevant terms to work with. Even easier, you can manage most of these steps right in KeywordKeg.

Here’s what my list looks like including terms throughout the shopping funnel i.e. specific shoelace terms, shoe specific shoelace terms (different brands), and top of the funnel terms like “how to” and others.

Armed with a solid list of 200 terms to target, the question remains: now what? Answer: It’s time to group these into topics.

Building topic maps for your store

A topic map is a spreadsheet where you group related keywords together into smaller lists that all roll-up under a representative topic.

For example, some of the topic maps from my data set are:

Topic: “No tie shoelaces”
Keyword	Volume
No tie shoelaces	9,900
Shoelaces no tie	9,900
Shoelaces no tying	8,100
No tie elastic shoelaces	720
Elastic no tie shoelaces	720
Best no tie shoelaces	480
No tie shoelaces for runners	260
No tie shoelaces hickies	260
How to make no tie shoelaces	210
Homar no tie shoelaces	110

Topic: “Shoe laces how to”
Keyword	Volume
Shoe laces how to	5,400
Shoelaces how to	5,400
How to tie shoelaces	3,600
How to put in shoelaces	880
How to bleach shoelaces	210
How to tie shoelaces for running	210
How to tie your shoelaces	170
How to shorten shoelaces	170
How to do shoelaces	140
How to hide your shoelaces	110
How to tie shoelaces fast	100

In both of the examples above, I’m taking general swings at grouping together the keywords from my list into topics. But this isn’t the end of our research process. From here, it’s important to determine if these groupings make sense based on content type.

Google has shown time and time again that it prefers to rank specific kinds of content for specific queries. If you want to rank for a particular term or a set of terms, you need to build out the type of content that Google is showing you it wants to see.

The most common types of content are:

• Informational pages (think Wikipedia)
• Product catalog pages (Category, Sub-Category, Product Detail)
• Blog posts (even these can vary in form, but for the purposes of this post I’m going to lump them all together)
• Video
• Reviews
• Visual / infographics
• Audio / podcasts

Mapping the different content types

It’s helpful to deliberately plan the types of content you need to create to best target your keywords. The most straightforward way to do this is through some old fashion Googling. Create a spreadsheet for yourself using the exported lists of keywords from KeywordKeg as a starting point.

Within that sheet, you can create a new tab with the following columns:

• Keyword
• Topic
• Volume
• Difficulty
• Content Type

The two highlighted rows above will need to be entered manually. Admittedly, this going to be a bit tedious, but I promise it’s worth the effort! Here’s an example of mapping the different types of content for my shoelace examples.

Example 1: “no tie shoelaces”

This one is a great example of a mixed search engine results page, or one that’s clearly evolving.

It looks like the easiest path to break onto this SERP is with a list-form article, and then, over time, invest content and link resources into ranking a product category page. Unless you’re selling direct on Amazon, in which case you’re going to need to amass a few thousand reviews.

Example 2: “shoe laces how to”

I love this SERP. These results are ripe with opportunity to leverage SEO-focused content creation. Specifically, making use of YouTube hacks to climb up into the top rankings (try some of these).

Example 3: “shoelaces Adidas”

Starting to get a sense of what I’m talking about?

In this third example we can immediately see the results skew toward pages with high commercial intent. There’s the product listing ads (PLA) carousel at the very top (before the text ads), and then we’re greeted with two product category pages, two product details pages, and another category page before being offered image results.

Now that you’ve gone though your priority terms and mapped all the content types, it’s time to create a plan for how to build out this content. A smart first step is designing a deliberate “content map.” This asset outlines your requirements (your ideal keywords), your blueprint (what content needs to be produced), and your structures (a roadmap for content production).

I’ve actually covered this topic in-depth before, so if you’d like the A-Z explanation be sure to read How to Design an SEO Content Map.

Developing a content calendar

You have your priority keyword list in hand, you know what types of content you need to create, and you’ve mapped these across your site.

Now it’s time to build a calendar so you can put rubber to road and start getting this content live and ranking for your target keywords. To do this, I find it helpful to create additional sheets in my overall keyword file and lay them out by funnel stage (based on search intent).

View the full image by clicking here.

You’ll also want to bring together all of the other components of this process.

• Topic Focus
• Content Type
• Test title 1
• Test title 2
• Test title 3
• Target keyword
• Additional keywords
• Target length
• URL

Here’s an active example including all of the parameters above.

View the full image by clicking here.

Then, build a sheet specifically for the schedule with the following columns.

• Draft Completed by Date [ACTUAL]
• Publish Date [ACTUAL]
• Draft Needed by Date [ESTIMATED]
• Expected Publish Date [ESTIMATED]
• Author
• Status [Not Started, Assigned, In Progress, Pending, Published]
• Blog post type [If post, Category of Post]
• Post topic and description
• Example Site
• Recommended URL
• Recommended Title
• Recommended H1
• Recommended Meta Description
• Target Keywords
• Notes

You can then bring in your content creation and management resources into this sheet and filter each of these columns to get visibility into your SEO efforts from a content and keywords perspective.

Your process determines your progress

Your search engine optimization efforts are only ever going to be as effective as your process. Most store owners, especially when they’re starting out and are strapped for time, don’t have a process in place and aren’t sure how to create one.

Maybe that was you at one time, but now, you have a game plan. 😃

I hope you found this helpful, and if you have any questions or want to discuss any of the points in more detail, drop a comment below and I’ll respond.