Competitive Analysis Mapping: How to Visualize Expertise Networks and Find Strategic Gaps

In this tutorial, we're going to show how you can use the InfraNodus marketing tool for mapping your competitors' expertise as a network and extracting interesting insights based on the clusters and gaps identified. You can also use this approach for building the social networks of expertise skillsets within an organization.

As an example, we will use a table with the list of the software tools that operate in the field of visual text analysis that we generated using the workflow described in our other help article: Beat Your Competition: Target Their Content Gaps with this n8n Automation Workflow 

Here is the link to the competitive analysis table for text visualization tools which you can also use as a template: https://docs.google.com/spreadsheets/d/1NO54_uyoCR4Tb2cvnbAc53AQg8o8pWJCR0ZGBl1tNBM/edit?usp=sharing

In that table, we have a column with the name of the companies and several columns with the descriptions of their CTA, USPs, and various other insights we obtained using the InfraNodus n8n workflow above: 

The easiest way to build the graph is to link the values in the first column (the company names) to the USP column. We will then see the main terms the companies use to describe their unique selling proposition. 

Another way to visualize the data (if you want to be more precise) is to create a social network that links each company to a specific skill, but you will need to prepare your data for it. 

We will demonstrate both approaches below. 

Mapping Companies to their Expertise Skillsets

In this approach, we are importing the data as a CSV file to InfraNodus. The data contains information about the companies in the text visualization field, their USP (unique selling proposition) and the keywords that identify their areas of expertise. 

There are multiple ways to import this data, so we need to clean and process it first to adjust it to the objectives we have. There are multiple ways of doing that:

Option 1: Automatic Data Import via InfraNodus CSV Import Settings

In this scenario, we clean the data using the InfraNodus CSV import itself. 

When asked for which columns we want to analyze in the step 2, we select column A (Name) and column E (USP).

Note, that in the additional settings in the 2. Columns to Analyze section, we need to specify:

1. Superlinked column — we tell InfraNodus to link the "Name" column to all the values in all other columns we select. It will have an @mention sign added to it and therefore show how the company names are linked to all the entities extracted from the related USP column
2. Entity detection (optional) — we tell InfraNodus to perform entity detection on the USP column, so that instead of visualizing each word it extracts the main skills of each company from that column and builds a graph from it. 

As a result, we are converting the table into something like this:

Note the @mention sign added to the names of the columns and the [[wikilink]] syntax used to extract the most important entities from the USP column text. 

You don't have to use the Entity detection setting, but in this case each company name will be linked to more words from the USP column, which is actually quite good for subsequent AI analysis and topical clustering (because the models will have more data) but the social network skills graph might become a bit saturated. 

Option 2: Better Control — Cleaning the Data with an AI Tool

If you don't want to bother with the InfraNodus settings or something is not working, you can also modify your original CSV file to replace plain text fields with specific syntax. You can do this manually when you gather the data or ask an AI tool to do that for you.

There are two ways to do that:

1. Transforming the freeform text data to comma-separated list of #hashtags — in this scenario, you list comma-separated values in each column instead of using freeform. For instance, instead of writing a full description of what MaxQDA can do, you can simply write: "#qualitative_data_analysis, #ai, #visualization, #transcription" in the USP column. Then, when importing the CSV to InfraNodus, you need to specify that you want to treat each comma-separated value in the columns you analyze as a separate entity node. This way you tell the system to only process those comma-separted values you identify as important.
2. Adding the [[wikilink]] syntax to the freeform text — in this scenario, you use [[double squarebrackets]] to show Infranodus which entities in the text you want to analyze, so you manually create the table that we specify in the example above rather than relying on InfraNodus to do that for you.

You can also do this automatically: simply upload the CSV file above to ChatGPT, Manus AI, or Claude and use the following LLM prompt where you ask it to do the same things and generate a new table for you:

Generate a new CSV file from the file I uploaded with only two columns:
Name - Extracting each value from the original and adding the @mention sign in front and replacing spaces with underscores
USP - listing only the most important expertise categories you identify from the original column, comma separated

Option 3: Maximum Control — Transforming a Spreadsheet into a Matrix

You can also create a much more precise representation of relations you want to analyze. In this instance, you need a matrix where every row represents the relation you want to analyze. 

For instance, in the table above we have another column with keywords assigned to each company based on its expertise. We can ask an AI tool to convert it to a table where the final result will look something like this:

In this case, InfraNodus will extract every relation: @MaxQDA will be connected to the "qualitative_data_analysis" node, which will also be connected to @NVivo etc. This lets you control precisely what kind of relations you want to visualize.

If you need to emphasize the strength of the connection, you can include the same relation several times. For instance, 3 rows for @MaxQDA to "qualitative_data_analysis" but only one for the same relation to @NVivo.

You can perform this conversion using an AI tool with the following prompt:

Convert the CSV file provided in the following way:
take the value from each "USP" column and for each comma-separated value 
create a new row that connects the value in the name field to that value 
in this value, replace spaces with underscores and prefix with a hashtag

Extracting Insights from the Company Expertise Social Network

Based on the settings we provided, InfraNodus will build a graph like this:

In this graph, the company names are the nodes with the @mention prefix. The skills are the nodes with the [[wikilinks]] syntax (we performed entity detection, that's why the wikilinks are added).

The nodes that have a bigger size have more influence in the network (using betweeness centrality measure). The nodes that are closer to each other and have the same color belong to the same cluster. The analytics panel on the right shows what those clusters are.

Let's take a closer look at the most important players and expertise sets:

• @MaxQDA — because it links to both the "graph visualization" (green) cluster and to the "qualitative software" (pink) cluster. We would identify this company as our main competitor in the visual text analysis field because it both perform qualitative analysis and visualize the results. 
• "graph" and "visualization" — because a lot of companies in our list have this expertise listed can perform this task
• "qualitative data analysis" — because a lot of companies also have this expertise
• "customer feedback analytics" — another field of expertise that a lot of companies have
• "maps and mind mapping" — another important expertise cluster
• "@Kineviz" — interesting competitor, because they provide both graph visualization capabilities and customer feedback analysis
• "@Open_Knowledge_Maps" and "@Heuristica — another important competitor because they are leading in the field of "concept mapping"

We now have a really good overview of the competitive field and identified the companies we should collaborate with (or compete with) in order to take over this market. 

For instance, if we want to compete in the field of visualization for qualitative analysis we could point our strengths in relation to MaxQDA by connecting our product to the "customer feedback analysis" cluster, emphasizing how it could be used to also analyze customer feedback and bringing the advanced techniques from qualitative analysis to surface-level sentiment analysis. 

At the same time, we could engage into a collaboration with Kineviz — a niche tool that is strong on data visualization and customer feedback analysis but is not strong on qualitative analysis and concept mapping capabilities.

Finally, we could use these insights to identify a niche that we want to focus on that is not occupied by too many competitors: e.g. the concept mapping one that seems to be smaller than others (because it's not linked to the other fields yet).

Finding the Innovation Gaps

Another advantage of using network analysis to map competition is your ability to see what is currently missing in the market. 

As the clusters of nodes representing companies and their USPs are clustered based on the density of connections, you can use the Content Gap feature of InfraNodus that shows you which clusters of nodes are not sufficiently connected. 

For our example, we can go to Analytics > Content Gaps, and we'll see this:

There is a gap between the cluster on Mind Mapping and Feedback Analysis, meaning there are no companies whose USPs include both. They are only connected with AI, but nowadays everyone's doing AI, however, no company in the market offers to use mind maps for customer feedback or market analysis. This could be an interesting niche to fulfil. 

We can reiterate through the content gaps to identify more interesting niches like that.