Link a Six Sigma project to organizational goals

Imagine a team celebrating faster order handling while customer support receives more complaints about wrong orders. A faster local measure is not enough. An improvement project needs to say what value it seeks, for whom, and which side effect is unacceptable. Six Sigma seeks to reduce errors and variation by understanding and measuring a process and testing causes, rather than merely telling staff to be more careful. DMAIC is a suitable road map for an existing process that misses performance expectations. A substantially new service calls for a different design decision, covered in the design-for-quality lesson.

SMART means specific, measurable, achievable, relevant and time-bound. Do not present achievable as a guarantee before studying the data; treat it as an initial estimate to revisit at the measure tollgate. An operational statement is: “Reduce orders with incorrect delivery addresses from 12% of 500 orders in the baseline month to 5% or less within four months, without increasing late deliveries above 8%.” This states numerator, denominator, baseline period, target and deadline. The primary measure is address error and the balancing measure is lateness. If “order” is defined differently in the two periods, the comparison is invalid even when the arithmetic is right.

Work through a fictional example. Among 500 orders, 60 required an address correction: 60 ÷ 500 = 12%. After an initial trial, 24 corrections were logged among 480 orders: 24 ÷ 480 = 5%. The numerical target is met, yet late deliveries rose from 6% to 11%. The whole project cannot be called successful. The new check may have fixed addresses while slowing flow. Break cycle time into steps, locate the wait, and test a change to the check that retains accuracy. The before-and-after comparison also does not establish causation: season or order mix may explain some change. A measurement and follow-up plan is needed before attributing the effect to the intervention.

Quality leaders contributed different tools for this reasoning. Shewhart distinguished routine variation from unusual signals. Deming emphasized management of the process and repeated learning. Juran linked quality planning, control and improvement. Ishikawa made teams consider a network of possible causes instead of blaming one worker. These are practical distinctions rather than a request to memorize biographies. Ask what decision changes. When an error rate varies within routine process behavior, punishing one bad day is unlikely to improve the system. When an unusual break follows replacement of the address-entry software, investigate that event first.

A useful scorecard names each metric, its operational definition, owner, update frequency, baseline and target. Relate the error count to workload: the count can rise while the rate falls if order volume doubles. Show the balancing measure alongside the primary outcome. Escalate to the sponsor when the team needs authority to change a system; send daily operating signals to the process owner. No single local indicator proves that profit or market share improved. Those more distant outcomes need separate evidence.