Market Segmentation is one of the most important parts of putting together a content strategy, along with personas. Who is buying your product or using your service? Where do they come from and how can you best market yourself to them?

Without market segmentation, you’re like a fisherman casting your net at the beach and coming up with nothing 99% of the time – except that one time you got lucky.

But how do you put together market segments? If you’re an established business, then you’ve probably got a lot of data on past customers. Geographic, Demographic, and lifestyle segmentation can easily be achieved with that data, but what about Benefit Segmentation?

Benefit Segmentation is the division of consumers based on their wants and needs. The question you should be asking yourself is: how does my customer benefit from my product?

Here, I’ll offer a simple technique for using comments to construct benefit segments yourself. Whilst this can be repeated with any type of business, I’m using a sample from an internal travel blog. Here, the question I need to ask myself is “What is my audience getting out of the content I create?”

Why Don’t I Just Read the Comments Myself?

You’re probably wondering why you can’t just do this yourself? Read through a couple of comments, look at what people are saying, and go ahead and use that as your rationale. Well…

You can. There’s nothing stopping you from doing that. However, if you’re looking to put together an airtight market segmentation strategy based on data, you’re going to need quantity, and reading through 1,000 or 10,000 different comments just isn’t a productive use of your time – especially if you can automate most of the process in the first place.

Content Collection

As you have to do with any data-based research task, you’ll need to start by collecting and organizing your comments. It’s up to you how you do this, but you’re going to want to end up with a .txt document.

As a small example, I’m going to use comments from a travel blogging site I’ve worked on, and use comments located on the site’s ‘Asia’ section. In total, I’ve collected 127 comments from 5 different articles, with a total of 7,059 words.

When you come to performing a comment analysis yourself, it’s a good idea to separate different sections of your website into different text documents. Each of these will serve as a base for your analysis, with the rationale that different people come to your site for different reasons and so visit different sections. Whilst you can divide your comments in any way you wish, make sure that your system makes sense.

Removing Comment Boilerplate

Before we get to the analysis itself, we’re going to start by removing any boilerplate from our collection phase. This includes things like dates and times. We also want to remove any of our own comments and replies from the data – we already know what we sound like (unless we’re new and performing an audit).

The easiest way to do this is with the Microsoft Word Replace tool and some careful scrolling and deleting. If you have a lot of data that you’ve collected and you can’t be bothered to trawl through it all, you can use TextCrawler to make changes a little faster. Just be careful that you don’t delete data which you actually want.

For my content, I wanted to delete responses from the site owner, dates, times, and the ‘REPLY’ button text which came with each comment. There were also several names of commenters which I don’t necessarily need, but which would have been too time-consuming to delete.

In the end, the structure of each of my individual comments went from this:

[Name] [Date] [Time] [Comment] [REPLY]

To this:

[Name] [Comment]

Analysis

Populous Keywords

Popping the ‘Asia.txt’ file into AntConc, I’m going to start by looking at the frequency list and seeing if I can spot anything interesting.

Firstly, look for any topics which have been mentioned a lot. The sample I’m using is to small to create any real conclusions, but if you’re dealing with comments from over 200 different posts, you’re going to start seeing patterns emerge. In this small article, I’m seeing a lot of mentions of ‘border’. This is probably because one of the articles selected is about border crossing. I’m also seeing frequent mentions of particular destinations – the destinations which the articles were about.

Finally, I’m noticing that ‘Visa’ is being mentioned a lot. None of the articles I’ve selected were exclusively about Visa’s – although the border crossing article would concern it. The frequency of ‘Visa’ may mean that a lot of commenters want to know more about visa issues and processes. We’ll explore this later.

Looking Deeper

As we’re dealing with comments, we’re particularly interested in finding personal pronouns which refer to the author of the comment. This includes “I” and “we”.

With these, we’re going to want to take a deeper look. If you’re using Antconc, you can click on a particular word and it will open up concordance lines (basically, lines of text which include your selected word).

Selecting “we”, I’m greeted by a lot of content about similar experiences which readers have had or plan to have:

We are looking to be there late January…
we arrived at the border crossing…

And some about how the reader had discovered something new in the post, which they didn’t know about when they visited.

we could have enjoyed it.
we didn’t know enough about this city…
We didn’t stop in…
we were not interested in tubing, but…

We can then go back to the comments and look at the context within which these statements fall. From that we can gather the full extent of the comment’s meaning.

For instance, the content of ‘we didn’t know enough about this city’ was within a comment praising the article for its detailed information. This means that there is an opportunity to create even more detailed content about that city – especially if multiple commenters have mentioned similar things. Moreover, if several commenters have stated that they plan on visiting a location, we can assume that further content related to that topic would be useful.

Looking back at the ‘visa’ keyword, a deeper look at concordance lines reveals a lot of advice on how to obtain a visa:

it is possible to get the visa on arrival
we were exempted from a visa
visa situation is getting stricter

This definitely affords is a future content opportunity, as it is something which hasn’t been covered in enough detail in the original article.

Analyzing Spread

The final step of this analysis is analyzing spread. It’s great being able to identify new content opportunities through frequency, but frequency doesn’t mean a lot if it’s all located in one article. Consequently, we’re going to have a look at the concordance plot to see how our keywords are dispersed throughout the comments.

Firstly, looking at the keyword ‘visa’, you’ll notice that it is dispersed throughout three major areas. This indicates that visas are actually a common issue with the Asia section of the blog, and a topic which this audience desires more information about.

With ‘border’, we’re also seeing a fairly even spread across the second half of the corpus. However, as one of our articles is about a border crossing, it might not be as useful a topic to create new content on.

What About Everyone Else?

You might be aware of one glaringly obvious problem at this point… What about the people who don’t comment?

Well, I never said that this is a 100% airtight and exclusive technique for ensuring you’re capturing the right market segments. It is, however, a way to capture those who are likely passionate about your content. After all, it takes a lot more effort to leave a comment on something than just click away (unless it’s spam).

Market Segmentation is Plural

It’s important to remember that there are multiple ways of ‘splitting a market’. This is just one, and the segmentation itself largely relies on having clear-cut content topics in place. This analysis should help you to find opportunities.