The Energy of Framework Dimensions: What Information Scientists Ought to Know

A earlier article supplied a of conceptual frameworks – analytical constructions for representing summary ideas and organizing information. Information scientists use such frameworks in all kinds of contexts, from use case ideation and validation of machine studying fashions to productization and operation of user-facing options. The framework kind (e.g., hierarchy, matrix, course of movement, relational map) and framework dimensions (e.g., categorical, ordinal, steady) largely decide the appear and feel of a conceptual framework. Whereas the earlier article devoted more room to a dialogue of framework varieties, will place the highlight on framework dimensions. With the assistance of a real-life case examine, we’ll see how modifying the framework dimensions can yield a perceptual shift that may unlock new insights. This deep dive goals to higher equip readers to make use of and construct conceptual frameworks extra successfully.

Be aware: All figures within the following sections have been created by the writer of this text.

A Primer on Framework Dimensions

Whereas the framework kind defines the construction of what you are attempting to signify, the framework dimensions decide the content material. The scale usually fall into three lessons: categorical, ordinal, and steady. The next sections look at this classification of framework dimensions in additional element and go over some features that you need to take into account when together with a number of dimensions in a framework.

The Massive Three

Allow us to begin with the category of categorical dimensions, which is presumably the best class of dimensions. Because the identify suggests, the dimension consists of a finite set of discrete classes that needn’t be in any explicit order. As an illustration, if the dimension represents an organization’s markets, it could possibly be divided into geographic classes reminiscent of “USA,” “Germany,” and “China.” Equally, you possibly can have a categorical dimension that breaks down the corporate’s merchandise into totally different product segments (e.g., by components, relevance to prospects, and so forth). It’s also a good suggestion to maintain the MECE precept (mutually unique and cumulatively exhaustive) in thoughts every time you might be breaking down a dimension into smaller classes; in any case, you need the classes to totally cowl the scope of the dimension and keep away from redundant classes.

Ordinal dimensions are much like categorical ones, with the extra characteristic that the classes making up the dimension are additionally ordered in a roundabout way. The ordering means that you can say that one class is “better 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. Nevertheless, the rankings needn’t suggest that the profitability values of nations are evenly spaced; the profitability hole between the top-ranked and second-ranked nation could possibly be totally different from the hole between second- and third-ranked nations. Ordinal dimensions are additionally usually used to assemble survey questions, taking the type of a Likert scale (e.g., “disagree,” “impartial,” “agree”). The ordering permits responses throughout the survey members to be analyzed when it comes to the place they lie on the dimensions for every query.

Lastly, a steady dimension provides a quantitative measure of one thing. Not like categorical and ordinal dimensions (which include discrete classes or values), steady dimensions can doubtlessly tackle any worth (nonetheless tiny) inside a given vary. For instance, the likelihood, in share phrases, of some occasion occurring can lie wherever between 0% and 100%; values reminiscent of 5%, 10% and 10.00123% would all be permissible. The values of a steady dimension are additionally inherently ordered.

Selecting Dimensions Properly

You will need to take into account the strengths and limitations of every dimension class earlier than making use of them to your framework. As an illustration, you possibly can take a look at the data content material of every dimension class. The presence of an ordering and the power to tackle more and more fine-grained values inside a given vary contribute to the depth of the data content material. Primarily based on info content material, ordinal dimensions ought to be favored over categorical ones, and steady dimensions ought to be favored over the opposite two every time they are often measured in a granular, quantitative method. Nevertheless, the data richness comes at the price of the assets wanted to acquire and analyze the information underlying the size. Additionally, presenting and explaining information-rich dimensions to an viewers may be exhausting, since there’s loads of content material that must be unpacked and digested. As such, even for those who use steady dimensions to carry out the evaluation, it might make sense to “bucket” the continual information into ordinal and even categorical information to simplify what’s proven to an viewers.

Moreover, since frameworks can contain a number of dimensions, you will need to obtain an optimum interaction between the size. There are not less than two fundamental selections that you’ll want to make on this regard – what number of dimensions, and what varieties, to incorporate within the framework. Particularly within the early levels of analyzing an issue, the tendency is to be beneficiant with the variety of dimensions thought of, for the reason that drawback will not be well-understood at this level and there’s a danger of eliminating doubtlessly precious dimensions prematurely. However as your evaluation progresses, a handful of dimensions will usually stand out from the remaining as being particularly key; these dimensions would be the ones that designate the answer most utterly and succinctly, or those that unlock novel insights. The variety of dimensions might also rely on the framework kind that you just wish to use. For instance, whereas a two-by-two matrix can solely deal with two dimensions, a hierarchy can doubtlessly deal with many extra.

When deciding on the sorts of dimensions to incorporate within the framework, you may select both dimensions of the identical class or of various lessons. Every class comes with a novel mind-set concerning the underlying information. Utilizing dimensions of the identical class has the benefit of letting you switch one mind-set throughout the size within the framework. As an illustration, if you realize that the framework solely makes use of steady dimensions, then you may doubtlessly apply the identical quantitative mind-set – and the related equipment, reminiscent of arithmetic operators and statistics – to all of them. You possibly can thus additionally examine dimensions of the identical class extra simply (assume “apples to apples” versus “apples to oranges”). Nevertheless, utilizing dimensions of various lessons additionally has its deserves. In a hierarchical framework, utilizing totally different dimension lessons for every degree within the hierarchy might help distinguish the degrees from each other extra clearly. For instance, the top-level ideas in a given hierarchy could also be categorical, whereas the sub-concepts could also be ordinal or steady; on this case, going deeper into the hierarchical construction would even be paralleled by a rise within the information-richness of the size concerned, which can assist your analytical thought course of.

Finally, the selection of framework dimensions when it comes to amount and variety will almost certainly be a part of an iterative course of. The scale that you just begin off with originally of the framework-building course of could not essentially be those you find yourself together with within the closing framework. Additionally, as with most issues, there’s seemingly no “excellent” dimension, simply dimensions which can be roughly appropriate on your framework goal. Being conscious of the strengths and limitations of the size and seeing framework-building as an iterative course of ought to assist take the stress off on the outset and can help you give attention to constructing a helpful conceptual framework.

Case Research: Gross sales Efficiency at SoftCo

The sheer number of framework dimensions, and their sturdy coupling with the framework goal, signifies that hand-picking “a very powerful” dimensions (or choosing primarily based on another standards) may be tough. But, altering the size whereas sustaining the identical framework kind can result in very totally different interpretations of the framework. Within the following anonymized case examine, we’ll see how even slight modifications to the size could make an enormous distinction and yield new insights.

SoftCo is a mid-sized expertise firm that provides marketing-related software program services and products to companies. The corporate operates within the US and has about two dozen gross sales reps unfold out nationally throughout totally different territories. The gross sales reps are liable for rising the enterprise of their territory, which incorporates every part from figuring out potential prospects to interacting with them and shutting the sale. On the finish of each month, Sally, SoftCo’s veteran Head of Gross sales, opinions the efficiency throughout all territories and reviews her findings to the CEO. She additionally provides suggestions to the gross sales reps to acknowledge achievements and recommend methods to enhance. Through the years, Sally has recognized a number of components that may affect the efficiency of particular person gross sales reps, together with the quantity of buyer interplay (usually telephone calls, with a number of subject visits). Determine 1 reveals a easy scatter plot (a matrix framework with two steady dimensions) that compares gross sales efficiency to buyer interactions for particular person gross sales reps.

The selection of dimensions in Determine 1 guides the interpretation of the framework in some ways, past the truth that Sally has chosen particularly to look at buyer interplay as a key predictor of gross sales efficiency. The usage of steady dimensions lends itself naturally to quantitative measurement. Gross sales efficiency is thus measured by the amount of cash every rep generates per 30 days, whereas buyer interplay is measured by the variety of gross sales calls made per 30 days. After all, these measures alone are in all probability not enough to totally seize the 2 framework dimensions. As an illustration, the variety of calls doesn’t inform us something concerning the high quality and distribution of the calls throughout prospects, and the greenback worth of the offers a gross sales rep generates in a month doesn’t inform us a lot concerning the strategic nature of the offers (e.g., whether or not the offers had been about rising the enterprise with present prospects, or “door openers” for a brand new stream of enterprise with new prospects). However, by trying on the scatterplot in Determine 1, we will derive a number of attention-grabbing insights:

• There have been 23 gross sales reps working for SoftCo throughout the noticed month. In whole, the gross sales staff made about $858,000 on this time interval.
• On common, every gross sales rep made about $37,300 price of gross sales in the noticed month. The very best and lowest particular person gross sales had been about $50,000 and $14,000, respectively.
• Probably the most environment friendly and least environment friendly gross sales reps (when it comes to $/calls) made about $2,000/name and $160/name, respectively; that may be a roughly 12x distinction in effectivity.
• There appears to be a non-linear relationship between buyer interplay and gross sales efficiency. As much as about 75 calls, every extra name appears to be correlated with an enormous enhance in gross sales efficiency. However past 75 calls the hyperlink with gross sales efficiency is much less sturdy.

Determine 1 thus results in a spread of insights which can be derived by trying on the efficiency of particular person gross sales reps and the efficiency of all the group. A number of the insights are pretty simple (e.g., the variety of gross sales reps, common gross sales efficiency), giving us a normal understanding of the dimensions of SoftCo’s gross sales operation and the character of the enterprise. Different insights, such because the hole between essentially the most and least environment friendly gross sales reps, and the non-linear relationship between gross sales efficiency and buyer interplay, are doubtlessly extra thought-provoking; apart from highlighting attainable gaps between the skills of various gross sales reps and diminishing returns from too many calls, the insights additionally recommend that different components past buyer interplay might also be good predictors of gross sales efficiency. The scatterplot illustration additionally makes it straightforward to determine the outliers among the many gross sales reps, which may be helpful for additional evaluation of what units these outliers other than the remainder of the gross sales reps.

Now, to indicate how altering the category of the size can result in a distinct perspective, Determine 2 presents a two-by-two matrix that’s primarily based on the identical info because the earlier scatterplot. The 2 steady dimensions of the scatterplot have been remodeled into ordinal dimensions by splitting them alongside sure threshold values. Gross sales efficiency figures beneath $25,000/month are thought of “low,” whereas these above are “excessive.” Equally, buyer interplay figures beneath 75 calls/month are “low,” and people above are “excessive.” The selection of the edge worth is clearly necessary and ought to be primarily based on cheap argument. For instance, the gross sales efficiency threshold could also be primarily based on a minimal gross sales goal that every gross sales rep is required to hit, and the shopper interplay threshold could possibly be associated to the purpose at which the curve in Determine 1 begins to flatten (indicating a shift within the marginal worth of extra gross sales calls).

Whereas the scatterplot in Determine 1 drew our consideration to the performances of particular person gross sales reps and the general pattern within the relationship between gross sales efficiency and buyer interplay, the two-by-two matrix in Determine 2 allows a extra simplified view that lends itself to a segmentation of gross sales reps into totally different teams. In step with conventions, the bottom-left quadrant of the two-by-two matrix reveals the group of gross sales reps that could be in an undesirable place; these reps are making comparatively few calls and producing few gross sales. The highest-right quadrant comprises “star performers” that evidently appear to work together extensively with prospects and ensure that this tough work interprets into precise gross sales. The dynamics within the different two quadrants appear much less clear. The reps within the top-left quadrant appear to realize excessive gross sales regardless of making comparatively few calls – what’s the secret behind their effectivity and is it sustainable? The reps within the bottom-right quadrant have the alternative dynamic, making loads of calls that don’t appear to repay – if these reps are primarily working as exhausting because the star performers, why are they not reaching equally excessive gross sales figures?

By drawing consideration to totally different segments of the gross sales staff, the two-by-two matrix can be utilized to develop tailor-made methods that tackle the distinctive traits of every section. For these within the bottom-left of the matrix, you will need to discover out why each buyer interplay and gross sales efficiency are comparatively low. Do these gross sales reps should cope with tough prospects, do the reps want extra coaching, or are the reps allocating a few of their time to different precious actions that aren’t captured by this month’s gross sales efficiency (e.g., coaching different workers, strategic planning, and private improvement)? Armed with these extra insights, Sally can develop measures that higher seize the true worth that the gross sales reps within the bottom-left quadrant of Determine 2 create for SoftCo.

Equally, for the bottom-right quadrant, a brand new technique could also be wanted to extend effectivity by translating the comparatively excessive degree of buyer interplay into precise gross sales; this will likely contain prioritizing sure leads over others, coaching the gross sales reps to be extra tenacious in closing every sale, and motivating them to proceed hustling. For the remaining two quadrants, reaching sustainability could possibly be the important thing goal. It’s price understanding what makes the gross sales reps within the top-left quadrant so environment friendly and what the opposite gross sales reps can be taught from them. On the similar time the reps within the top-left additionally want a technique for lowering the chance of slipping down if their buyer interplay doesn’t constantly pan out. Lastly, a technique is required to maintain the reps within the top-right quadrant motivated (e.g., by social recognition, financial rewards, alternatives for promotion) to maintain them performing constantly at a excessive degree.

To shut off, here’s a useful video by Mike Gastin that expands on a number of the issues mentioned above when selecting dimensions for two-by-two matrices:

Reflection Questions

This part consists of three units of reflection questions that may immediate you to assume extra deeply concerning the materials coated above. The goal is that will help you shortly perceive the essential ideas and get you excited about how you should use them in your individual work.

Set 1: Take an present framework (e.g., one that you’ve used or in-built an actual information science undertaking) and analyze the framework dimensions in additional element. What number of dimensions does the framework have and which of the three lessons we’ve mentioned do they belong to? Does altering the category of any of the size have an effect on your interpretation of the framework and the insights which can be produced?

Set 2: When you have seen and/or produced a number of frameworks to date, it might be an excellent time to take inventory of the size you are likely to see most frequently. What lessons do these dimensions fall beneath? To what extent is the recognition of those dimensions an excellent factor or a foul factor when it comes to reaching every framework’s goal?

Set 3: Are you able to consider another lessons of dimensions past the three we’ve checked out on this article? To what extent are these different lessons totally different from those we’ve mentioned?

The Wrap

Whereas the framework kind determines how the framework will say one thing (the construction), the framework dimensions outline what particularly will probably be stated (the content material). Three lessons of framework dimensions are particularly widespread in apply: categorical (unordered, discrete classes), ordinal (ordered, discrete classes), and steady (a quantity line inside a given vary). It’s attainable to rework a dimension from one class to a different by altering the depth of the data content material (e.g., bucketing steady information to yield an ordinal dimension). You will need to take into account the amount and variety of dimensions a framework ought to have to realize the overarching goal. Embrace solely as many dimensions as are actually wanted, particularly when presenting the framework. Limiting dimensions to a single class can have some advantages, though the interplay of dimensions from totally different lessons additionally has its deserves.