Supply Chain by the Numbers

Supply Chain by
the Numbers

A very sharp process improvement program manager — he was, admittedly, not a supply chain expert — recently asked me what seems like a simple question: “Is SCOR a noun or a verb?” Being a habitual SCOR user and proponent, I approached the answer in much the same way I would answer the question, “How do you put on your pants — right leg or left leg first?” In other words, I needed to mentally retrace some steps.

SCOR project steps

Step 1
First, pick and define your project's SCOR metrics; second, define a data collection plan including appropriate data sampling, data segmentation requirements and defect analysis for each of your chosen SCOR metrics; third, collect your benchmark comparative data; fourth, achieve consensus on your channel competitive requirements prioritizing delivery reliability, flexibility and responsiveness, supply chain management cost, and asset management efficiency using one superior, one advantage and two parity ratings; fifth, assemble your SCORcard and calculate your competitive parity, advantage and superior requirements opportunities.

Step 2
First, complete an AS IS Geographic map and SCOR process Thread Diagram; second, using the chosen SCOR metrics above, segment the data by physical location, including the defect analysis; third, conduct a disconnect analysis using the SCOR Level One Metrics as categories that includes disconnect brainstorming, affinity diagrams and fishbone analyses; fourth, identify leading practices that support your competitive requirements priority above; fifth, identify TO BE Material Flow changes (including eliminating disconnects and adding appropriate leading practices); sixth, conduct an opportunity analysis calculating savings opportunities for each TO BE change; seventh, prioritize changes using an effort-impact matrix and identify two to three quick hit changes; eighth, initiate detailed design, pilot and rollout phase of the quick hits.

Step 3
First, complete an AS IS Process Flow (a combination of transactional analysis and swim diagrams); second, using transactional productivity calculations, segment the data by business process (SCOR Level Three), including the defect analysis; third, map the leading practices (identified in Step 2) using SCOR Level Three process elements; fourth, identify TO BE Work and Information Flow changes using the SCOR TO BE process Blueprint (including eliminating disconnects and adding appropriate leading practices); fifth, conduct an opportunity analysis calculating savings opportunities for each TO BE change; sixth, aggregate work and information flow changes with material flow using the same effort-impact matrix.

Step 4
First, aggregate the supply chain changes identified in Steps 2 and 3 into a project approach; second, assemble the final opportunity calculation (using the defect analysis) for each project, essentially assembling an ROI; third, assemble a project charter based on the preferred implementation approach, i.e., Six Sigma, lean, software implementation, class A, etc.; fourth, initiate detailed design, pilot and rollout phase.

Supply chain delivery performance

Those who know me can accept the fact that my mind does, in fact, go through those mental steps — and those who've worked with me on a project know that I practice what I preach. After retracing the steps, I concluded that SCOR is a noun. I used one guideline in coming to that conclusion: improvement (savings) needed to be initiated to be considered a verb (Step 2-point 8 and Step 4-point 4). While SCOR metrics have a huge impact on identifying what to improve and estimating the impact, by themselves they will not save any money (point 2 of Steps 1 through 4).

The following is an example of a SCOR project team's journey of turning Delivery Performance data and analysis into real supply chain improvement. The team followed seven basic steps:

1. collect raw data;
2. derive descriptive statistics to characterize the sample;
3. assemble a histogram to isolate the defects;
4. conduct a pareto analysis of the defects to generate top causes (problem statements);
5. utilize fishbone analysis to identify root causes for top problem statements;
6. assemble action plan to eliminate root causes;
7. execute action plan.

Raw data

There are three questions to answer in gathering raw data. First, what is the operational definition that you will use to query the data? Second, what is an appropriate sampling plan, or how much data is enough? Third, what are appropriate ways to segment the data?

The project team chose one of the literal definitions of SCOR Delivery Performance — “On Time and In Full Delivery to Customer Commit Date.” The challenge that most of the company's current delivery measures focused on line item fill rate at order shipment. Ideally, the team wanted to collect data for all sales orders in the last 12 months, but after some test queries concluded that all sales orders in January and February would be sufficient to represent known variability.

To expedite the data extraction, the team identified several different ways to segment the data (sorting and grouping): by sales order, customer number, customer ship to number, planned shipment date, actual ship date and internal shipping location (source of supply).

Descriptive statistics — characterizing the data

While we have a number for the metric, we have little understanding. There are three questions to answer in assembling descriptive statistics: First, which segments require statistics? Second, what “average” should be used? Third, what is the standard deviation (variability) of the sample?

The project team chose to apply descriptive statistics independently to each facet of the On Time and In Full segments of the measure.

In the case of On Time Delivery, the team concluded that the Mean was sufficient to represent the average and that the data was highly variable, ranging from 53 days late to 51 days early. For the In Full Delivery segment, the team concluded that the Median characterized the average better than the Mean, with this data segment ranging from 552 units short to 17 units over.

Histogram

This step really initiates the “why” questioning process, and the focus changes from the entire data set to only those data points considered defective. There are two questions to answer at this step: First, what is considered a defective data point? Second, how should the defective data points be grouped and labeled to begin the problem-solving process?

The project team defined as defective any sales order that was either not on time or in full. Using the histograms for both On Time Delivery and In Full Delivery, the team was able to begin the problem-solving process.

Pareto charts

While the histogram tells us what is defective, the pareto analysis starts to identify the “why.” There are two questions this step attempts to answer: First, what are the primary categories of defects? Second, what are the biggest problem areas?

The project team started by analyzing and sorting the On Time Delivery defects (using an affinity diagram approach). They then completed the process on the orders delivered On Time but not In Full.

Fishbone analysis

The fishbone analysis focuses on generating a cause and effect summary for each bar on the pareto chart. The results will ultimately lead to the action-result step. There are two questions this step attempts to answer: First, what are the primary causes contributing to a specific defect category? Second, what are the secondary causes to a specific defect category?

The project team divided up the bars on the pareto chart and created 11 fishbone diagrams.

Each primary (and secondary) bone provided the focus for action planning and execution and ultimately performance improvement results. For example, the bone labeled Planning Master Data - Customer Lead Time Settings impacted five orders that were not delivered On Time. Resolution of this issue would have immediate positive impact on about 5% of all orders. Eliminating the Data Settings are Inaccurate bone would have immediate positive impact on 17% of the total orders.

Plan and execute improvements

This step focuses on an action plan required to eliminate the primary and secondary causes (and ultimately “kill” the whole fish). There are at least nine aspects to a good implementation plan:

  • Issue and root cause analysis;
  • Recommendation;
  • Action plan;
  • Responsibilities/timing;
  • Payoffs;
  • Implementation resources;
  • Implementation leader;
  • Implementation sponsor(s);
  • Charter status.

The project team assembled project charter summaries for each fish (one for each bar on the pareto chart) and initiated immediate action on two. The implementation approach was based, in large part, on the complexity of the change.

The intent of this article has been to provide a summary roadmap of the basic steps of taking a metric through to implementation and results. While SCOR provides a world class cross-industry framework, it does not do the heavy-lifting analysis and problem-solving, nor does it make change happen — that part is up to you. LT

Peter Bolstorff is president and CEO of SCE Limited (www.scelimited.com), which supports “do-it-yourself” supply chain performance through education, coaching and process expertise. He is the co-author (with Logistics Today's Bob Rosenbaum) of Supply Chain Excellence: A Handbook for Dramatic Improvement Using the SCOR Model (AMACOM, 2003). As a member of the board of directors with the Supply-Chain Council, he has been involved with the development of the SCOR model since its inception. He can be reached at [email protected].

Logistics Today logo
July, 2004

Feedback on this article?

© Want to use this article?
Click here for options!

Copyright© 2004 Penton Media, Inc.

at a glance

This article offers a step-by-step example of a typical SCOR supply chain implementation

Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish