A previous article provided a of conceptual frameworks – analytical structures for representing abstract concepts and organizing data. Data scientists use such frameworks in a wide range of contexts, from use case ideation and validation of machine learning models to productization and operation of user-facing solutions. The framework type (e.g., hierarchy, matrix, process flow, relational map) and framework dimensions (e.g., categorical, ordinal, continuous) largely determine the feel and look of a conceptual framework. While the previous article devoted more room to a discussion of framework types, will place the highlight on framework dimensions. With the assistance of a real-life case study, we’ll see how modifying the framework dimensions can yield a perceptual shift that may unlock latest insights. This deep dive goals to raised equip readers to make use of and construct conceptual frameworks more effectively.
Note: All figures in the next sections have been created by the creator of this text.
A Primer on Framework Dimensions
Whereas the framework type defines the structure of what you are attempting to represent, the framework dimensions determine the content. The scale generally fall into three classes: categorical, ordinal, and continuous. The next sections examine this classification of framework dimensions in additional detail and go over some facets that it is best to consider when including multiple dimensions in a framework.
The Big Three
Allow us to start with the category of dimensions, which is possibly the only class of dimensions. Because the name suggests, the dimension consists of a finite set of discrete categories that needn’t be in any particular order. As an example, if the dimension represents an organization’s markets, it could possibly be divided into geographic categories corresponding to “USA,” “Germany,” and “China.” Similarly, you possibly can have a categorical dimension that breaks down the corporate’s products into different product segments (e.g., by ingredients, relevance to customers, and so forth). It is usually idea to maintain the MECE principle ( and ) in mind at any time when you might be breaking down a dimension into smaller categories; in spite of everything, you wish the categories to totally cover the scope of the dimension and avoid redundant categories.
dimensions are much like categorical ones, with the extra feature that the categories making up the dimension are also ordered in a roundabout way. The ordering means that you can say that one category is “greater than,” “lower than,” “equal,” or “unequal” to a different. Suppose you took an organization’s set of markets and ranked them by a criterion like profitability. The rating would impose an ordering on the set of markets, thereby producing an ordinal dimension representing the profit-based (ascending or descending) ordering of markets. Nonetheless, the rankings needn’t imply that the profitability values of nations are evenly spaced; the profitability gap between the top-ranked and second-ranked country could possibly be different from the gap between second- and third-ranked countries. Ordinal dimensions are also often used to construct survey questions, taking the shape of a Likert scale (e.g., “disagree,” “neutral,” “agree”). The ordering allows responses across the survey participants to be analyzed when it comes to where they lie on the size for every query.
Finally, a dimension gives a quantitative measure of something. Unlike categorical and ordinal dimensions (which consist of discrete categories or values), continuous dimensions can potentially tackle any value (nevertheless tiny) inside a given range. For instance, the probability, in percentage terms, of some event occurring can lie anywhere between 0% and 100%; values corresponding to 5%, 10% and 10.00123% would all be permissible. The values of a continuous dimension are also inherently ordered.
Selecting Dimensions Properly
It will be significant to think about the strengths and limitations of every dimension class before applying them to your framework. As an example, you possibly can take a look at the knowledge content of every dimension class. The presence of an ordering and the flexibility to tackle increasingly fine-grained values inside a given range contribute to the depth of the knowledge content. Based on information content, ordinal dimensions must be favored over categorical ones, and continuous dimensions must be favored over the opposite two at any time when they might be measured in a granular, quantitative manner. Nonetheless, the knowledge richness comes at the associated fee of the resources needed to acquire and analyze the info underlying the size. Also, presenting and explaining information-rich dimensions to an audience might be hard, since there’s lots of content that should be unpacked and digested. As such, even if you happen to use continuous dimensions to perform the evaluation, it could make sense to “bucket” the continual data into ordinal and even categorical data to simplify what’s shown to an audience.
Moreover, since frameworks can involve multiple dimensions, it’s important to attain an optimal interplay between the size. There are at the very least two basic decisions that you’ll need to make on this regard – what number of dimensions, and what kinds, to incorporate within the framework. Especially within the early stages of analyzing an issue, the tendency is to be generous with the variety of dimensions considered, because the problem will not be well-understood at this point and there’s a risk of eliminating potentially beneficial dimensions prematurely. But as your evaluation progresses, a handful of dimensions will typically stand out from the remaining as being especially key; these dimensions stands out as the ones that designate the answer most completely and succinctly, or those that unlock novel insights. The variety of dimensions may depend upon the framework type that you wish to use. For instance, whereas a two-by-two matrix can only handle two dimensions, a hierarchy can potentially handle many more.
When deciding on the sorts of dimensions to incorporate within the framework, you may select either dimensions of the identical class or of various classes. Each class comes with a singular way of excited about the underlying data. Using dimensions of the identical class has the advantage of letting you transfer a method of considering across the size within the framework. As an example, if you happen to know that the framework only uses continuous dimensions, you then can potentially apply the identical quantitative way of considering – and the associated machinery, corresponding to arithmetic operators and statistics – to all of them. You may thus also compare dimensions of the identical class more easily (think “apples to apples” versus “apples to oranges”). Nonetheless, using dimensions of various classes also has its merits. In a hierarchical framework, using different dimension classes for every level within the hierarchy may help distinguish the degrees from each other more clearly. For instance, the top-level concepts in a given hierarchy could also be categorical, while the sub-concepts could also be ordinal or continuous; on this case, going deeper into the hierarchical structure would even be paralleled by a rise within the information-richness of the size involved, which can help your analytical thought process.
Ultimately, the alternative of framework dimensions when it comes to quantity and variety will most probably be a part of an iterative process. The scale that you just start off with originally of the framework-building process may not necessarily be those you find yourself including in the ultimate framework. Also, as with most things, there’s likely no “perfect” dimension, just dimensions which can be roughly suitable on your framework objective. Being aware of the strengths and limitations of the size and seeing framework-building as an iterative process should help take the pressure off on the outset and assist you to give attention to constructing a useful conceptual framework.
Case Study: Sales Performance at SoftCo
The sheer number of framework dimensions, and their strong coupling with the framework objective, implies that hand-picking “a very powerful” dimensions (or choosing based on another criteria) might be difficult. Yet, changing the size while maintaining the identical framework type can result in very different interpretations of the framework. In the next anonymized case study, we’ll see how even slight modifications to the size could make a giant difference and yield latest insights.
SoftCo is a mid-sized technology company that gives marketing-related software services and products to businesses. The corporate operates within the US and has about two dozen sales reps opened up nationally across different territories. The sales reps are accountable for growing the business of their territory, which incorporates every part from identifying prospective customers to interacting with them and shutting the sale. At the tip of each month, Sally, SoftCo’s veteran Head of Sales, reviews the performance across all territories and reports her findings to the CEO. She also gives feedback to the sales reps to acknowledge achievements and suggest ways to enhance. Over time, Sally has identified several aspects that may influence the performance of individual sales reps, including the quantity of customer interaction (typically phone calls, with a couple of field visits). Figure 1 shows an easy scatter plot (a matrix framework with two continuous dimensions) that compares sales performance to customer interactions for individual sales reps.
The alternative of dimensions in Figure 1 guides the interpretation of the framework in some ways, beyond the proven fact that Sally has chosen specifically to look at customer interaction as a key predictor of sales performance. Using continuous dimensions lends itself naturally to quantitative measurement. Sales performance is thus measured by the sum of money each rep generates per 30 days, while customer interaction is measured by the variety of sales calls made per 30 days. In fact, these measures alone are probably not sufficient to totally capture the 2 framework dimensions. As an example, the variety of calls doesn’t tell us anything concerning the quality and distribution of the calls across customers, and the dollar value of the deals a sales rep generates in a month doesn’t tell us much concerning the strategic nature of the deals (e.g., whether the deals were about growing the business with existing customers, or “door openers” for a brand new stream of business with latest customers). Nevertheless, by taking a look at the scatterplot in Figure 1, we will derive several interesting insights:
- There have been 23 sales reps working for SoftCo throughout the observed month. In total, the sales team made about $858,000 on this time period.
- On average, each sales rep made about $37,300 value of sales in the observed month. The best and lowest individual sales were about $50,000 and $14,000, respectively.
- Probably the most efficient and least efficient sales reps (when it comes to $/calls) made about $2,000/call and $160/call, respectively; that may be a roughly 12x difference in efficiency.
- There appears to be a non-linear relationship between customer interaction and sales performance. As much as about 75 calls, each additional call appears to be correlated with a giant boost in sales performance. But beyond 75 calls the link with sales performance is less strong.
Figure 1 thus results in a variety of insights which can be derived by taking a look at the performance of individual sales reps and the performance of all the group. A number of the insights are fairly straightforward (e.g., the variety of sales reps, average sales performance), giving us a general understanding of the size of SoftCo’s sales operation and the character of the business. Other insights, corresponding to the gap between probably the most and least efficient sales reps, and the non-linear relationship between sales performance and customer interaction, are potentially more thought-provoking; besides highlighting possible gaps between the skills of various sales reps and diminishing returns from too many calls, the insights also suggest that other aspects beyond customer interaction may be good predictors of sales performance. The scatterplot representation also makes it easy to discover the outliers among the many sales reps, which might be useful for further evaluation of what sets these outliers aside from the remaining of the sales reps.
Now, to point out how changing the category of the size can result in a distinct perspective, Figure 2 presents a two-by-two matrix that relies on the identical information because the previous scatterplot. The 2 continuous dimensions of the scatterplot have been transformed into ordinal dimensions by splitting them along certain threshold values. Sales performance figures below $25,000/month are considered “low,” while those above are “high.” Similarly, customer interaction figures below 75 calls/month are “low,” and people above are “high.” The alternative of the edge value is clearly essential and must be based on reasonable argument. For instance, the sales performance threshold could also be based on a minimum sales goal that every sales rep is required to hit, and the shopper interaction threshold could possibly be related to the purpose at which the curve in Figure 1 starts to flatten (indicating a shift within the marginal value of additional sales calls).
Whereas the scatterplot in Figure 1 drew our attention to the performances of individual sales reps and the general trend in the connection between sales performance and customer interaction, the two-by-two matrix in Figure 2 enables a more simplified view that lends itself to a segmentation of sales reps into different groups. In line with conventions, the bottom-left quadrant of the two-by-two matrix shows the group of sales reps that could be in an undesirable position; these reps are making relatively few calls and generating few sales. The highest-right quadrant accommodates “star performers” that evidently appear to interact extensively with customers and ensure that that this tough work translates into actual sales. The dynamics in the opposite two quadrants seem less clear. The reps within the top-left quadrant seem to attain high sales despite making relatively few calls – what’s the secret behind their efficiency and is it sustainable? The reps within the bottom-right quadrant have the alternative dynamic, making lots of calls that don’t appear to repay – if these reps are essentially working as hard because the star performers, why are they not achieving similarly high sales figures?
By drawing attention to different segments of the sales team, the two-by-two matrix might be used to develop tailored strategies that address the unique characteristics of every segment. For those within the bottom-left of the matrix, it’s important to search out out why each customer interaction and sales performance are relatively low. Do these sales reps should take care of difficult customers, do the reps need more training, or are the reps allocating a few of their time to other beneficial activities that should not captured by this month’s sales performance (e.g., training other staff, strategic planning, and private development)? Armed with these additional insights, Sally can develop measures that higher capture the true value that the sales reps within the bottom-left quadrant of Figure 2 create for SoftCo.
Similarly, for the bottom-right quadrant, a brand new strategy could also be needed to extend efficiency by translating the relatively high level of customer interaction into actual sales; this may occasionally involve prioritizing certain leads over others, training the sales reps to be more tenacious in closing each sale, and motivating them to proceed hustling. For the remaining two quadrants, achieving sustainability could possibly be the important thing objective. It’s value understanding what makes the sales reps within the top-left quadrant so efficient and what the opposite sales reps can learn from them. At the identical time the reps within the top-left also need a technique for reducing the chance of slipping down if their customer interaction doesn’t consistently pan out. Finally, a technique is required to maintain the reps within the top-right quadrant motivated (e.g., by social recognition, monetary rewards, opportunities for promotion) to maintain them performing consistently at a high level.
To shut off, here’s a helpful video by Mike Gastin that expands on a number of the considerations discussed above when selecting dimensions for two-by-two matrices:
Reflection Questions
This section consists of three sets of reflection questions that can prompt you to think more deeply concerning the material covered above. The aim is to enable you to quickly understand the essential principles and get you excited about how you should use them in your personal work.
Set 1:
Set 2:
Set 3:
The Wrap
While the framework type determines how the framework will say something (the structure), the framework dimensions define what specifically shall be said (the content). Three classes of framework dimensions are especially common in practice: categorical (unordered, discrete categories), ordinal (ordered, discrete categories), and continuous (a number line inside a given range). It is feasible to remodel a dimension from one class to a different by changing the depth of the knowledge content (e.g., bucketing continuous data to yield an ordinal dimension). It will be significant to think about the amount and variety of dimensions a framework must have to attain the overarching objective. Include only as many dimensions as are truly needed, especially when presenting the framework. Limiting dimensions to a single class can have some advantages, although the interaction of dimensions from different classes also has its merits.
