SAS and 80/20

© 1998 by SPIKEware, Inc. All Rights ReservedLast Updated Wednesday, June 26, 2002

SAS, The SAS Software System, and its components are trademarks of The SAS Institute, Inc. in Cary, NC

Rapid Decision Support (RDS) and its components are trademarks of SPIKEware, Inc. in Schaumburg, IL

Microsoft Excel and its components are trademarks of Microsoft Corporation in Redmond, WA

This material is provided "as is" by SPIKEware, Inc. There are no warranties, express or implied, as to merchantability or fitness for a particular purpose regarding the materials or code contained herein. SPIKEware is not responsible for errors in this material as it now exists or will exist, nor does SPIKEware provide technical support for it. Questions or problem reports concerning this material may be addressed to SPIKEware by electronic mail: [email protected]

SAS™ and the 80/20 Principle

An Added Bonus to Rapid Decision Support™

Presented by Paul McDonald


SAS and the 80/20 Principle

RDS--What is it?

• Rapid Decision Support

• The ability to make useful decisions quickly



The 80/20 Principle

• Other names– “The Rule of the Vital Few”– “The Pareto Principle”– “The Peter Principle”

• The principle states that most of your results come from a small proportion of your efforts



80/20 Examples

• Business– 20% of salespeople make 80% of sales– 20% of investments make 80% of return

• Sports & Leisure– 20% of at-bats make 80% of the hits– 20% of movies make 80% of profits– 20% of picnickers eat 80% of food



Key Points of the 80/20 Principle• Rarely is the “split” exactly 80/20

– The concept is that there are vital few and trivial many

– 80/20 is a rule of thumb to tag the concept

• Most 80/20 decisions are very easy to make in time and efforts

• All 80/20 decisions do not require complex analysis



How to Make a Quick 80/20

• The RDS Base Macro %PARETO will prepare a basic 80/20 or PARETO chart for analysis– SAS/GRAPH package recommended– Microsoft Excel is also very useful

• PROC PARETO in the SAS/QC Product– SAS/Graph package recommended– This is the ideal approach



Submitting PROC PARETO

• Syntax: – PROC PARETO < options >;– VBAR (variables) / < options >;/* required */– SYMBOL < n > <options>; /* available with graphic */– NOTE <options> 'text'; /* devices only */– BY variables;

• Can utilize the SAS/Graph procedures to make high-resolution output or otherwise will create low-resolution output

• More versatile, easier to use, and more efficient than %PARETO macro



SAS/QC Output(High Resolution)



SAS/QC Output(Low Resolution)

Pareto Plot of GROUP vs PROFIT

„ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ† 70 ‰ +++++*‡ 100 ‚ +++++*++++++ ‚ ‚ +++*++++++ ‚ ‚ ++++++ ‚ ‚ +++*++ ‚ ‚ ++++++ ‚ ‚ ++*++ ‚ 60 ‰ ++++ ‚ ‚ ++++ ‚ ‚ ++*+ ‚ ‚ ++++ ‡ 80 ‚ ++++ ‚ ‚ ++*+ ‚ ‚ ++++ ‚ 50 ‰ ++++ ‚ ‚ ++*+ ‚ ‚ ++++ ‚ C ‚ ++++ ‚ u ‚ +*+ ‚ m ‚ +++ ‚ u ‚ +++ ‡ 60 l 40 ‰ +++ ‚ a P ‚ *+ ‚ t e ‚ ++ ‚ i r ‚ + ‚ v c ‚ ++ ‚ e e ‚ + ‚ n ‚ ++ ‚ P t 30 ‰ + ‚ e ‚ ++ ‡ 40 r ‚ * ‚ c ‚ ++ ‚ e ‚ + ‚ n ‚ + ‚ t ‚„ƒƒƒƒƒƒƒƒƒƒƒ† ++ ‚ 20 ‰‚ ‚ + ‚ ‚‚ ‡ƒƒƒ+ƒƒƒƒƒƒƒ† ‚ ‚‚ ‚ ++ ‚ ‚ ‚‚ ‚+ ‡ƒƒƒƒƒƒƒƒƒƒƒ† ‚ ‚‚ * ‚ ‚ ‡ 20 ‚‚ ‚ ‚ ‚ ‚ ‚‚ ‚ ‚ ‚ ‚ 10 ‰‚ ‚ ‚ ‚ ‚ ‚‚ ‚ ‚ ‡ƒƒƒƒƒƒƒƒƒƒƒ† ‚ ‚‚ ‚ ‚ ‚ ‡ƒƒƒƒƒƒƒƒƒƒƒ† ‚ ‚‚ ‚ ‚ ‚ ‚ ‡ƒƒƒƒƒƒƒƒƒƒƒ…ƒƒƒƒƒƒƒƒƒƒƒ…ƒƒƒƒƒƒƒƒƒƒƒ† ‚ ‚‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‡ƒƒƒƒƒƒƒƒƒƒƒ† ‚ ‚‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‡ƒƒƒƒƒƒƒƒƒƒƒ…ƒƒƒƒƒƒƒƒƒƒƒ† ‚ ‚‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‡ ƒƒƒƒƒƒƒƒƒƒƒ†‚ 0 ‰‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚‡ 0 Š‹ƒƒƒƒƒ…ƒƒƒƒƒ‹ƒƒƒƒƒ…ƒƒƒƒƒ‹ƒƒƒƒƒ…ƒƒƒƒƒ‹ƒƒƒƒƒ…ƒƒƒƒƒ‹ƒƒƒƒƒ…ƒƒƒƒƒ‹ƒƒƒƒƒ…ƒƒƒƒƒ‹ƒƒƒƒƒ…ƒƒƒƒƒ‹ƒƒƒƒƒ…ƒƒƒƒƒ‹ƒƒƒƒƒ…ƒƒƒƒƒ‹ƒƒƒƒƒ…ƒƒƒƒƒ‹ƒƒƒƒƒ…ƒƒƒƒƒ‹ƒƒƒƒƒ…ƒƒƒƒƒ‹Œ Dav Joe Ann Nic Sam Ben Dan Rob Pat Amy Jea Bob

Sales Group



The %PARETO MacroFor Those Without SAS/QC• Syntax: %pareto (data=, x=, y=, file=)– data = SAS_dataset– x = x-axis_variable (character or numeric)– y = y-axis_variable (numeric)– file = output_file-specification <optional>

• Data does not have to be summarized by the X value



Sample SAS Logs

8388 %pareto (data=year, x=group, y=profit, file=pareto)NOTE: SAS/ACC-ODBC License Not ActiveNOTE: Creating Comma-Separated file=PARETO for import into MS-Excel template pareto.xlt.NOTE: SAS/GRAPH License Not ActiveNOTE: Creating Pareto Plot using Base SAS ProceduresNOTE: Using a second SAS Session in background processing to generate reportNOTE: SPIKEware SAS macro PARETO used 11.92 seconds.NOTE: (c) 1997 by SPIKEware, Inc. All Rights Reserved.

7386 %pareto (data=year, x=group, y=profit, file=pareto)NOTE: SAS/ACC-ODBC License ActiveNOTE: Creating MS-Excel Spreadsheet file=PARETO for import into MS-Excel template pareto.xlt.NOTE: Load completed. Examine statistics below.NOTE: Inserted (12) obs into pareto.xls.NOTE: Rejected (0) insert attempts see the log for details.NOTE: SAS/GRAPH License ActiveNOTE: Creating Pareto Plot using SAS/GRAPHNOTE: SPIKEware SAS macro PARETO used 7.86 seconds.NOTE: (c) 1997 by SPIKEware, Inc. All Rights Reserved.

%PARETO macro with BASE SAS procedures and BASE SAS output file.

%PARETO macro with SAS\GRAPH procedures and SAS\ACCESS output file.



%Pareto Output(High Resolution)



%Pareto Output(Low Resolution)

Pareto Plot of GROUP vs PROFIT

„ƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒ† ‚ ‚ 22.5 ˆ ˆ ‚ + ‚ 100 ‚ + ‚ ‚ ‚ + ‚ ‚ ‚ ‚ 20.0 ˆ ‚ + ˆ ‚ ‚ ‚ 90 ‚ ‚ + ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ 17.5 ˆ ‚ ‚ + ˆ ‚ ‚ ‚ ‚ 80 ‚ ‚ ‚ ‚ C ‚ ‚ ‚ + ‚ u ‚ ‚ ‚ ‚ m 15.0 ˆ ‚ ‚ ‚ ˆ u T ‚ ‚ ‚ ‚ + ‚ 70 l o ‚ ‚ ‚ ‚ ‚ a t ‚ ‚ ‚ ‚ ‚ t a ‚ ‚ ‚ ‚ + ‚ i l 12.5 ˆ ‚ ‚ ‚ ˆ v ‚ ‚ ‚ ‚ ‚ 60 e P ‚ ‚ ‚ ‚ ‚ r ‚ ‚ ‚ +‚ ‚ P o ‚ ‚ ‚ ‚ ‚ e f 10.0 ˆ ‚ ‚ ‚ ˆ r i ‚ ‚ ‚ ‚ ‚ 50 c t ‚ ‚ ‚ ‚ ‚ e s ‚ ‚ ‚ ‚ ‚ n ‚ ‚ ‚ ‚ ‚ ‚ t 7.5 ˆ ‚ ‚ ‚ ‚ ˆ ‚ ‚ +‚ ‚ ‚ ‚ ‚ 40 ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ 5.0 ˆ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ˆ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ 30 ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ 2.5 ˆ +‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ˆ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ 20 ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ 0.0 ˆ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ‚ ˆ ‚ ‚ ‚ ‚ ‚ ‚ Sales Group ‚ ‚ ‚ ‚ ‚ ‚ ‚ ŠƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒƒƒƒƒƒƒƒƒƒƒˆƒƒŒ Dav Joe Ann Nic Sam Ben Dan Rob Pat Amy Jea Bob

Sales Group



%Pareto Output(Export to Microsoft Excel)

Pareto Graph

0%

5%

10%

15%

20%

25%

Category

Pe

rce

nt

of

To

tal

0%

20%

40%

60%

80%

100%

120%

Cu

mu

lati

ve

Pe

rce

nt

Percent 21% 18% 15% 8% 7% 6% 6% 6% 4% 3% 3% 1%

Cum Percent 21% 40% 55% 63% 70% 77% 83% 89% 93% 96% 99% 100%

Dav Joe Ann Nic Sam Ben Dan Rob Pat Amy Jea Bob



How to Find the Vital Few

• Draw in several best-fit lines (eyeball method works well)

• Circle the points of intersection of the best-fit lines

• Draw vertical lines intersecting each circle to declare the Pareto boundaries



The Pareto Boundaries

• Data to the left of the first best-fit line intersection point are the vital few

• Data to the right of the last best-fit line intersection point are the trivial many

• Data in between the outer-most best-fit line intersections may be trivial or vital



How to Find the Vital Few

Vital Few

Trivial Many

Your Call



IMPORTANT!!!

• Not every distribution will fit the model of the vital few

• The analysis is only as accurate as your eyeball

• A radical change in slope of the cumulative frequency line is a good indicator of the vital few model



Recommended Reading

• The 80/20 Principle – by Richard Koch– ISBN 0-385-49170-0

The 80/20 principle -- that 80 percent of results flow from just 20 percent of our efforts -- is one of the great secrets of highly effective people and organizations. Did you know, for example, that 20 percent of customers account for 80 percent of revenues? That 20 percent of our time accounts for 80 percent of the work we accomplish? Now available in paperback, "The 80/20 Principle" shows how we can achieve much more with much less effort, time, and resources, simply by identifying and focusing our efforts on the 20 percent that really counts

Available at SPIKEware Books , your favorite bookstore, or your local public library.

http://www.vstore.com/cgi-bin/pagegen/vstorecomputers/spikeware/page.html?vcontext=T&mode=itempage&file=/page/itempagev4/itempage.spl&prodID=966540



About the Author

SPIKEware, Inc.115 ½ West Main StreetWest Dundee, IL 60118Phone (847) 428-6250Fax (847) 428-6270

SPIKEware, Inc. is a consulting firm dedicated to helping other people solve their problems through technology. We specialize in providing solutions for our customers through the SAS Software System.

http://www.SPIKEware.com/

Paul D. McDonald is the CEO of SPIKEware, Inc. and has been a SAS programmer since 1993.

Paul has an A.A. in Electrical Engineering from Cloud County Community College, a B.A. in Physics from Southwestern College, and an M.B.A. in Finance from Keller Graduate School of Management.

Paul can be reached by e-mail at [email protected]

SAS and 80/20

Documents

pareto note

pareto principle

pareto data

pareto macro

pareto syntax

pareto chart

pareto output export

pareto boundaries data