Just as any IT implementation shouldn’t be for its own sake—that is, it should serve a business purpose within the sponsoring organization and not simply be a cost center—quantitative analysis within the context of an organization should likewise serve a business purpose. For example, there must be some reason a widget manufacturer commissions a study of its customer base. It wasn’t brought up just to keep the research division busy. There are typically research questions and hypotheses that exist and guide the methodology.
In my own research consulting work, I have often started with broad research questions that then drive more narrow research questions and/or particular segment analyses. At the analysis level, the variables and desired outcomes are examined in order to determine what test to use. From that point, it is easy to get lost in the vocabulary of quantitative analysis and forget that the work is being done to answer a business question.
For example, assuming the National Widget Company commissioned that study of its customer base, I could simply report the measures of central tendency and leave them to interpret why there’s a difference between the mean and median ages. But a true data scientist/analyst helps explain why the numbers mean what they do, and ensures the business users don’t get lost in the lingo. I would take the time to explain that the mean age is 42.5, the median age is 37, and that difference indicates there are more instances of older customers than younger and possibly some outliers bringing that mean age up. I would then turn back to them and ask what this means for their business. Remember that as the analyst, we are not the business subject-matter experts. Offering the numbers to the business and asking them to provide context creates more opportunities for synergy.
Consider another example involving correlation. Two variables, or points of interest as we would call them: widget sales and distance from a major airport. A strong negative correlation (r=-0.49) is found. First we must caution against equating correlation and causation. We would then pivot away from the r-value and put the focus back on the variables of interest: it appears that an individual who lives closer to a major airport is more likely to buy these widgets. Again, we would put the question back on the business to then have a conversation about why these variables might be related and the possible covariates.
In either case, and in any analytics situation, proper use of visualization is paramount. In the latter example it is much easier to see what a high r-value means on a scatterplot as opposed to explaining it verbally. Data visualization bridges many gaps that numbers and words simply cannot fill. These are the languages of dashboards, executive roll-ups, and KPIs.
Overall, the primary thing to remember in keeping an audience engaged in a discussion around quantitative research is this: the variables of interest are the reason for the study, not the numbers themselves. Keep the focus on what matters.