Aging Three Surveys Before the January Cycle
Helio Systems, a 900-employee medical device maker in Minneapolis, is refreshing pay ranges for its software engineering job family before the January merit cycle. You have three survey sources with different effective dates: Survey A effective January 1 with a median base of $132,000 and 18 incumbent matches; Survey B effective July 1 with $129,500 and 9 matches; Survey C effective October 1 with $126,800 and 4 matches. Today is November 15, and the company budgets a 4.0% annual increase. Policy is to lead the market at the 60th percentile for critical roles and match the 50th for others. Software engineering is designated critical. The VP of HR proposes averaging the three raw medians and setting the midpoint at $129,433, then placing the range at plus or minus 15%. Two sources define 'total cash' differently: one includes on-call pay, the other excludes it. Incumbent compa-ratios range from 0.81 to 1.19.
1. Calculate and critique the VP's $129,433 midpoint. What is the correct aging and weighting approach?
Read the worked answer
The VP's $129,433 is the simple average of three raw medians: (132,000 + 129,500 + 126,800) / 3 = 129,433. It is wrong for two reasons. First, the surveys have different effective dates, so they measure different labor markets. Age each to a common date, January 1. At 4.0% annual, Survey A ages 12 months: 132,000 × 1.04 = 137,280. Survey B ages 6 months: 129,500 × 1.04^0.5 ≈ 132,064. Survey C ages 3 months: 126,800 × 1.04^0.25 ≈ 128,049. Second, the samples differ. Weight by incumbent matches: (137,280×18 + 132,064×9 + 128,049×4) / 31 ≈ 134,574. That is the aged, weighted market median. The 60th percentile still needs percentile data or a documented premium; do not derive it from medians.
2. Recommend a midpoint and range, addressing data cuts and the total cash definition.
Read the worked answer
Use $134,574 as the aged median for the 50th percentile. Because software engineering is critical and policy is to lead at the 60th, obtain the published 60th percentile from at least one source or apply a documented competitive premium, for example 3-5%, which gives roughly $138,600-$141,300. Do not simply average. Treat Survey C cautiously: four matches is thin, and it may be a different scope. Check whether the sources use base salary or total cash; on-call pay should be included consistently or excluded from all. If the sources cannot be harmonized, use base salary only and track on-call separately. Set a range around the midpoint with a spread wide enough for the job family, often 40-50% for engineering, not 30%. Then place incumbents by compa-ratio and performance.
3. What governance should be in place before the January cycle?
Read the worked answer
Before January, document the survey selection, effective dates, aging factor, weighting, and percentile choice. Get compensation committee approval for the midpoint and range. Run a compa-ratio distribution and identify outliers below 0.85 and above 1.15; decide whether to adjust or freeze. Communicate the methodology to HR and managers, and train them not to promise ranges. Keep an appeal process and a review date. If the data are thin, set the range with a wider spread and revisit in six months rather than overfitting to four matches.