Why You Should Treat Your Customers Like Numbers


There was a story out this week on social media that attempted to draw a correlation between U.S. Olympic athletes failing to medal and the athletes bib number in competitions where bibs are required.  Basically, the argument was that athletes wearing the number “4” were highly likely not to win the competition. The inference being that this was due to the number “4” being an unlucky number in Korea as the number “13” is in the U.S.  The reality is, however, that it is not a coincidence at all, nor is there a supernatural force at play.  The fact is that bib numbers, or starting orders, are assigned based on performance in past events.  In other words, a competitor wearing the number “4” is, on average, expected to be the fourth best competitor in the competition.  Therefore, it would not be entirely surprising if they finished off the podium which recognizes only the top three competitors in each event.

As these athletes probably resent their starting placement, understanding the advantages of starting order and their sense that they can perform better than their past performance may indicate, consumers often resent being treated like a number, as well, because they do not like being fit into a box with other consumers.  However, if properly used, numbers, or more precisely, statistics, can provide for a better customer experience rather than the opposite.  Unfortunately, many businesses attempt to forecast using acquired data without properly interpreting the results.  This can provide for a less than favorable customer experience.

Let’s say that your business produces blue and red widgets.  You want to determine how many of each you should produce, so you conduct an online survey of your customers which results in a 50/50 tie.  The current state of the market informs you that 50% of your customers prefer red widgets and 50% prefer blue.  As a result, you set up a production plan to allocate production, 50/50 by color.   When the widgets are distributed, however, you find that you have a surplus of blue widgets and a shortage of red widgets.  How did that happen?

It happened because you may have misinterpreted the results of your survey.   Perhaps it is that you put too much value into the results of that one survey.  What if you had taken the same survey a month before and demand for blue widgets was at 75% and demand for red widgets was at 25%.  Then, when you took the most recent survey you would have realized a shift in sentiment, where demand for red widgets was on the rise.  Had you had this trended data, you would have predicted that you would need to produce more red widgets than blue widgets to meet future demand.

While this example may seem simplistic, I recently worked with a client who provided me with statistics on client engagement channels, branch, internet, phone, etc.  The discussion we were having was about where the organization would get the greatest results from process and technology improvements.  At the time, the organization was receiving the majority of their customer contact by phone, which seemed odd to me, but we discussed why that might be.  Come to find out it was because their website was a communications blackhole.  Their customers figured out that they got better service from just making a call, rather than trying to use the website.

So, if we just used the raw statistics to determine where to apply resources, the argument could be made that the organization should invest in its telephone equipment and training call center staff.  But, the reality is that the statistics didn’t necessarily demonstrate a customer preference, only a customer reality.  This is why trained data scientists who know your business and understand your data are critical to sound business intelligence.  The term Data Scientist is not just a new term for database administrator; it signifies a person who understands technology, mathematics and, most importantly, your business.  If you have questions about improving your organization’s business intelligence by engaging qualified data scientists, contact us at ds@cubi.pro or visit our website at cubi.pro.

Comments

Popular posts from this blog

How to Estimate the Impact of CECL on ALLL

Are You a Coach or a Referee?

Measurements Matter