If you are anything like me, you love a good cup of coffee/tea and your contribution to Starbucks’ last quarter (Q1FY17) report of $4 Billion USD net revenues1 is more than you want to tell your spouse. So let me give you a few excuses to visit your local Starbucks in order to study some examples of data quality.
Before we get into it, let me say that I don't usually use names of companies when illustrating poor data quality, but in this case I had to make an exception, because it includes both good and less than optimal traits. Besides, the title is just so catchy. The following examples are meant to provide real-life meaningful connections between our daily lives and data quality, per the goal of this blog.
First, is an example of good operational data quality practices that I found. As seen in photo "a", taken of a Starbucks paper bag. As you can see, the letter “C” is underlined in order to clarify its representation as “C” rather than an “L”, which could be confusing if written quickly. I wondered about this and confirmed with the barista that indeed this was the reason for the underline. As discussed in the last blog post on Representation, the Conformed Dimensions of Data Quality calls out the Underlying Concept of Easy to Read and Interpret as essential to data quality. Clearly, Starbucks has identified the need for careful representation in order to make sure we get the right order.
Next, we’ll take a look at two areas where Starbucks could improve operations thereby increasing consumer confidence. First you can see in the screenshot “b” below, that as you purchase items (and gain loyalty stars towards free food and drinks) the results of your transactions may not be reflected in their Android App for 24 hours. Now I may be impatient, but does it really take 24 hours in order to sync up financial transactions with their mobile app?
The reason that I call this out is that the screenshot “c” below shows my current balance of $16.72 USD, but when I walk into a Starbuck and reload my account for $20 it doesn’t reflect that in my APP, however the associate at the register allows me to purchase against that account immediately. This is an example of Timeliness and Currency (both experienced by the customer and the associate).
 |  |  |
| Photo a | Screenshot b | Screenshot c |
In order to process my purchase in a Timely way, the associate has to process my $20 purchase with my credit card company (within seconds typically) and then allow me to purchase items against that credit. They are able to do that because the credit is cached locally on the Point of Sale (POS) machine or transacted via a high-speed financial network. That means that the information reflected in near real-time (approximately 10 seconds) is Current (reflecting where the money sits- as credit on my card or as a sale by Starbucks and revenue generated). The problem is that the financial balance isn’t timely relayed to me, but is to the cashier. “Timeliness is a measure of time between when data is expected versus made available.” (Conformed Dimensions r3.4) In summary, although the Starbucks financial systems have Current data, I am unable to access that in a Timely way via the Android Application.
In conclusion, we can see that when defining time related aspects of data quality it's important to review who the customer is (cashier to process transaction versus consumer) and what their expectations are. However much I may want my balance to be reflected in the app immediately, Starbucks knows that I want to pay for my drink and get on with my life more. Therefore my guess is that the financial transact-ability was prioritized over the customer's knowledge of the current account balance. Perhaps in the not so distant future both will be available in a Timely manner to both audiences.
References:
1. Starbucks Financial Release: Starbucks Reports Record Holiday and Record Q1 FY17 Results, 01/26/2017 (http://investor.starbucks.com/press-releases/financial-release/2017/Starbucks-Reports-Record-Holiday-and-Record-Q1-FY17-Results/default.aspx, accessed 4/3/2017).