How can you tell if you've made progress on a trip from A to B without knowing where either A or B is, or where the intermediate points are? You can't. It's as simple as that. Without an effective measurement system, one that clearly indicates where the goal is and where matters stand now, real progress is all but impossible. Unfortunately, measurement remains the most abused component of many quality processes. Too often it's seen either as a weapon ("Gotcha!") or religion ("Measure everything!"). Both approaches are counterproductive. The moment when measurements become weapons precedes the moment when people begin lying to the measurers by about 47 seconds (according to careful measurements). And the rumor spread by quality control veterans that "the one who dies with the most charts wins" never has been subjected to objective measurement. Only two legitimate reasons justify taking measurements for quality improvement. The first is to use measurement data as a source of ideas; the second is to check progress against predefined goals. In the second case, when progress lags, revert to the first reason. It's usually not the data itself, but how the data is used that causes problems. For example, imagine that Jones, the boss, tells Smith, the subordinate, on the first day of the month, "I want you to take measurement x every hour, record your findings and send me the results at the end of each day." Smith, being a competent, diligent subordinate, follows instructions. Along the way, he notices that the numbers aren't what he and Jones originally had agreed upon as a goal, and he begins making a list of possible improvements. Now, it's the middle of the month, and Jones has data for two weeks in hand as he approaches Smith. Before Smith can say anything, Jones announces, "Now I know why things are wrong. The problem lies here with you. You're screwing things up, and I've got the numbers to prove it!" What are the odds of Smith sharing improvement ideas in the face of this attack? But imagine if, after two weeks, this had been Jones' approach: "When you have some time, we need to talk. I think you'll agree that the numbers you've been recording aren't what we both hoped they'd be by now. As the person on the scene, you're in the best position to think of ways we can get these numbers going the right way. Can you help?" The second approach undoubtedly will yield far different, and far better, results. The difference is one of attitude. Another stumbling block to effective measurement is fear -- of numbers and jargon. It all sounds so complex, but it needn't be. As one wit put it, "If you have to take a square root, you've gone too far!" Even the dreaded Pareto analysis, the crown jewel of statistical process control, is not that complicated. Consider this story from our book, Quality in Action (Wiley, 1992) about a paperboy named Pareto: Young Pareto noticed that every so often his bundle of papers was delivered late. Since he arrived at the dark, cold corner each morning at precisely the same time, not having the papers there when he arrived was both unpleasant and inconvenient. He decided to keep track of actual delivery times for a few weeks. Because he didn't care how early the papers came as long as they were waiting for him, he marked all such occasions as OK on his score sheet. But, when he arrived at the corner before the papers did, he recorded the exact time they finally were dropped at his freezing feet. He soon confirmed what he suspected: The papers were rarely late by more than five minutes on Monday through Friday and always 30 to 40 minutes late on Saturdays. Concerning the five-minute, weekday variations, Pareto decided it was likely that traffic and weather variables were at play. The one late weekday delivery outside the five-minute variance was explained by a late-breaking news story the night before that had justified holding the presses. He thought it was interesting that the ratio of on-time papers to late papers was 5 to 1. A phone call to the newspaper office downtown revealed that there was a different driver on Saturdays. The Saturday driver left the central office at the same time as the Monday through Friday driver -- the dispatcher made sure of that -- but he had a slightly different route. The budding statistician realized he had two choices: either sleep an extra 30 minutes on Saturday mornings or ask the Saturday driver to change the route and get to his corner earlier. The deciding factor was whether his customers cared; as far as Pareto was concerned, the extra sleep was an attractive option. While making his collections that week, Pareto asked his customers if receiving their paper 30 minutes later on Saturday mornings presented a problem. They were unanimous in agreeing that it did not. Pareto benefited from his analysis. He slept a little longer on the weekends from then on, and his weekly tips were bigger. Customers were delighted to have a paperboy who asked them when they wanted their paper delivered. It hadn't happened before. The scenario is written like a fairy tale, but in reality productive use of measurement is immeasurably easier and more effective than incorrect use. About the authors Pat Townsend and Joan Gebhardt have written more than 200 articles and four books: Commit to Quality (John Wiley & Sons, 1986); Quality in Action: 93 Lessons in Leadership, Participation, and Measurement (John Wiley & Sons, 1992); Five-Star Leadership: The Art and Strategy of Creating Leaders at Every Level (John Wiley & Sons, 1997); and Recognition, Gratitude & Celebration (Crisp Publications, 1997). Contact them at ptownsend@qualitydigest.com. |
