July 30, 2020
Quality Control and the Life of Data – Part 3: Grading
In the previous Life of Data
blog installments (Part 1
and Part 2
), we focused on mining for the gold nuggets of information buried in data streams. Specifically, we did a deep analytical dive into a stream of temperature data from a single walk-in cooler located in Miami, FL. We pointed out how historical data can provide early warnings—insights into the equipment’s overall capabilities—and also help to isolate root causes of systemic problems. We even discussed how to determine best temperature settings based on the equipment’s capabilities.
But here’s the deal…all this detailed analysis is perfect to help someone who is already expressing an interest in a certain piece of equipment or seeking ways to improve an activity (or their overall operations). They already know what needs to be analyzed—the realm of responsibilities is a “small” group of data streams from a single restaurant.
But what about the franchisee with 15 or 150 restaurant locations? Or better yet, the brand owner with thousands of store locations? These executive-level folks are just as engaged with improving operations and making sure stores are profitable as the single franchise owner. But staring at a sea of data streams to find those containing the gold nuggets of information is daunting. This is where we step off into today’s blog content. Let’s use this article to talk about sifting and grading your data.
The Need for Grading
Professional experimenters use statistical tools to isolate process inputs that have the most influence on a process’ output. For example, when driving a car on the road, one needs the ability to turn the car in the desired direction. An experimenter would ask, “What factors influence the driver’s ability to turn the car?” The next step would be to list all the candidate factors:
- Turning steering wheel
- Steering linkage condition
- Road surface conditions
The list is longer, of course, but you get the point. Because we know a lot about steering cars, we already know which factors have the biggest influence.
But what if I’m a quick service restaurant (QSR) executive responsible for ensuring smooth operations for hundreds of stores across the country? Imagine all the data streams from candidate factors that contribute to profit, food safety, quality, and customer experience. The numbers are overwhelming. This is when you turn to InfinityQS’ Quality Intelligence platform, Enact®
Data Stream Grading
Enact’s digital food safety platform does the heavy lifting of sifting through thousands, even millions, of data streams to find the vital few that need the most attention. This is accomplished through Enact’s very own, patent pending, analysis method called Data Stream Grading
, or just “grading” for short.
Each day, all data streams in the system are evaluated regarding how they are performing against their standards. For example, freezer temps need to be below 5° F, fryer temps need to be between 350° F and 375° F. Each temperature stream will have its own requirement. Then the system assigns a grade to the stream using patent-pending methods that go beyond the scope of this blog. There are nine possible grades (A1, A2, A3, B1, B2, B3, C1, C2, and C3). In short, a grade of A1 means the process is targeted properly with minimal variation. On the other extreme, a C3 stream is way off target with excessive variation. An A3 grade means that the stream average is way off target, but with minimal variation. The letter and number give clues to the type of corrective actions that need to take place.
The best part about grades is that they can be rolled up across multiple streams to allow the user to quickly see the big picture. Let’s see what grading looks like…
A Dive into Grading
Here’s one way to illustrate grades to help pinpoint areas that need attention. Table 1 below is the Grading Matrix
The Grading Matrix
This grid shows the stream count for each of the nine possible grades. Luckily, most of the streams are A1 and probably don’t need any investigation. However, you can see there are five streams with a C3 grade. At a glance, this simple grid has uncovered the five out of 764 total streams that are the most problematic. No need to analyze all the streams when the Grading Matrix instantly tells you what you want to know—the streams that are causing problems.
Table 1: Grading Matrix shows stream counts for each grade.
To know there are five C3 streams is great, but what are they? No problem, each of the numbers are linked to a details page. Clicking on the “5” brings up the Stream Details
analysis table you see below.
Table 2: Stream Details table lists all streams from the selected grade cell from the Grading Matrix.
You can see from Table 2 that there are three stores with C3 streams. Four of the five streams are cooling equipment issues, and one is a griddle issue. The Expected Yield
is the percentage of time that the stream is conforming. At 6.78%, store 0238 in Seattle is having some major problems with their refrigerator. The Potential Yield
is the best conformance result one can expect from the stream, assuming it’s targeted properly. Using this tool, you will know at any moment in time what streams need the most attention.
Another way to organize the grades is by site across multiple critical-to-quality checks. The Site Summary
table below shows the aggregate grades across different temperature checks from multiple store locations in Ohio.
Table 3 Site Summary table showing aggregate grade comparisons across different types of temperature checks rolled up across multiple store locations within a city.
Scanning Table 3 above, you can see that the worst grade (C2) comes from the Cook Temperature from store 3259 in Ada, Ohio. Just like the Grading Matrix, one can click on the grades on the Site Summary to see the streams that are contributing to the grade. Table 4 below shows stream details from the C2 grade.
Table 4: This Stream Detail table shows what products are having issues with cook temperature.
Table 4 displays all products from store 3259 that are measured for cook temperature. Notice that each of the products receives a grade. Because of the C3 grade, we instantly know that the 1/3 lb. beef patties are having issues. The next biggest offenders are chicken nuggets and fish portions.
With just a few button-clicks, an executive responsible for thousands of stores can quickly know where the most pressing issues are occurring. No more guesswork. No more spending hours buried in spreadsheets. And because all this data is in the cloud, these reports are real-time.
Now that we’ve found the handful of streams that need attention, we’re able to use the tools discussed in the previous blog to further analysis to find root causes. These grading charts will instantly sift through mountains of data, so you can focus on those areas which, if improved, will make your brand more profitable.
So far, the three blogs in this series have focused on data stream results as compared to standards. In the next installment, we’ll discuss how to use these aggregation tools to analyze and isolate problematic compliance areas. So, please come back. It’s going to be fun!
Read the other blog entries in this Life of Data
Take advantage of the technology at your fingertips today: contact one of our account managers (1.800.772.7978 or via our website
) for more information.