The Traditional Method of Segmentation

In this series of articles, I will investigate the current market landscape for BI, Analytics and Data Science vendors with the goal of answering a few important questions.

First, how should vendors in this market segment and evaluate who they should target? In other words, how should they develop their list of target accounts?

There are two traditional methods for market segmentation. The worst-case scenario involves marketers segmenting based on gut feel. Where is the market headed? What companies seem ripe for targeting? And so on. Guessing, in other words.

The better scenario involves reviewing historical sales data and targeting companies that have a profile (Industry, Revenue, Employee Count, Geographic Location, etc.) resembling that of past customers. In the case of BI and Analytics vendors, I imagine that the target accounts thus selected tend to come from the Software and Technology, and Finance and Insurance industries.

With targets selected, the second question becomes, how should these vendors market to their target accounts? In most cases, marketing messages and positioning echo whatever major conversation is being had within the market. Nowadays, content would therefore include references to phrases such as, “Citizen Data Scientist,” and emphasize how organizations need to empower these folks by moving away from Microsoft Excel, for example, and adopt more robust analytics solutions.

We advocate a different approach. Rather than segmenting based on gut feel, or even historical sales data, we recommend leveraging intent data and segmenting the market based on the observed behavior exhibited of potential customers. We also recommend using this same behavioral data, rather than the latest buzz word, to inform marketing content, messaging, and strategy.

Behavioral Segmentation

In the example that follows, I examine intent data related to a tool that first debuted in the mid-1980’s (1987 for Windows users) and has since become ubiquitous across every organization, regardless of industry, size, or geographic location: Microsoft Excel. I chose it as my starting point because so many solutions today market themselves as an alternative to “managing by spreadsheet.”

The specific behavior I chose to focus on was consumption of Excel tutorials and visits to Excel user forums. Individuals engaging in these behaviors were, according to my hypothesis, encountering the limits of what Excel was capable of and were, therefore, good potential prospects for BI and Analytics vendors.

To collect data, we monitored 45 Excel tutorial and user forum sits over the course of 52 days (1/1/19 – 2/21/19). In that time, we detected roughly 1.13 million active users across roughly 172,000 webpages (Table 1).

Table 1

So, a lot of people visited a lot sites related to Excel, which may be interesting but is not very useful from a targeting perspective.

To get there, we needed to clarify what we were looking for. In a sense, we were looking for “Future Citizen Data Scientists,” or individuals well versed in all of facets of Excel who were reaching the upper threshold of its capabilities.

The skills that most readily identify “Future Citizen Data Scientists” are:

  • Organization – The ability to quickly organize and understand data.
  • Evaluation – The ability to statistically/analytically evaluate organized data.
  • Iteration – The ability to create a repeatable process that can iterate upon the former.

All of these skills can be mapped back to and identified within Excel:

  • Organization – Usage of Pivot Tables
  • Evaluation – Usage of Statistical and Analytical Functions
  • Iteration – Visual Basic (VBA) programming

With this set of skills clearly defined, we were able to segment our 172,000 webpages into three distinct categories (representing the Excel skills we cared about), and monitor user activity for those pages specifically (Table 2). As you can see, there were many more pages and active users associated with the VBA category than with the other two. (For the sake of this analysis, we ended up combining the Statistical/Analytical and Pivot Table categories into a single category for our final user segmentation.)

Table 2

Who are the Future Citizen Data Scientists?

In the final step, we sought to classify the 1.13 Million users according to our skills categories. We did so by considering how many pages each user viewed from the two categories, and/or how many days they were actively engaging with each category. Those individuals that met the criteria for both a VBA Power User and an Analytics/Pivot Power user were then classified as a “Future Citizen Data Scientist (Figure 1).

Only .10% (1,113) of all identified users could be considered “Future Citizen Data Scientists” (Green). If that number seems small, it merely reflects how few Excel users have attained this high level of proficiency.

Figure 1

What Does this Mean?

So, what might a marketer do with all this fairly high-level information? After all, these are just a bunch of dots without names, titles, emails, phone numbers, or even an organization to associate them with. (Though, to be fair, at this point we are only one step away from appending at least some of that information to them.)

Well, each of these users (30,672 in total) represents a potential customer (or influencer within an organization) who could be approached with a specific, hyper-targeted marketing strategy:

  • “Citizen Data Scientists” (Green): Target with content focused on elevating data analysis skills to the next level. Topics could include data exploration, data investigation, data cleansing, statistical analysis and modeling, report generation/automation, and so on.
  • VBA Power Users (Yellow): Target with content focused on report generation/automation, data wrangling, data cleansing, and the integration of multiple data sources.
  • Analytic/Pivot Power Users (Red): Target with content focused on data exportation/investigation, data cleansing, outlier detection, statistical analysis and modeling.

With the data provided by my analysis, you would have enough to place ads (reflecting one of these strategies) directly in front of these users precisely at the point when they are the most highly engaged with relevant content. In this case, the ads themselves are content that will now be closely associated with these activities.

We must remember that primary goal is to engage with these users and this engagement can only happen if they visit your website. While ad performance is notoriously low for click through rates (CTR), it is not the only metric that should be used to evaluate their performance. A more complete metric is the percentage of users that visited your website after being served an ad, regardless if they arrived via a click through. It’s still a win for your marketing efforts if they visited, no matter the terms of their arrival.

In part two of this analysis, I’ll demonstrate the next logical step in this analysis. How to resolve these users to their respective companies, aggregate their activity, and begin to create a profile for each company. In mapping their DNA, based entirely on their users, we will have the ability to create hyper targeted strategies for each company.

Do you know which specific companies are currently in-market to buy your product? Wouldn’t it be easier to sell to them if you already knew who they were, what they thought of you, and what they thought of your competitors? Good news – It is now possible to know this, with up to 91% accuracy. Check out Aberdeen’s comprehensive report Demystifying B2B Purchase Intent Data to learn more.