How do you age salary survey data?
A small step that changes every number after it.
Short answer: Pick one date, usually the day your new bands or pay decisions take effect. Move every source to that date at a yearly rate close to how fast pay is rising, about 3% to 3.6% in the US now, scaled to the months between. Skip it and you average figures from different years as if they described one market.
Three sources, three dates
A 160-person payments company in Tampa prices its jobs for bands that take effect on 1 January 2027. One survey describes pay on 1 July 2025. A second describes 1 March 2026. Its live data was pulled in September 2026. The team averages the three as they are. The oldest figure is 18 months behind the band date. At 3.5% a year, that is about 5.3%, or roughly $6,400 on a $120,000 midpoint. Every band built from that average starts behind the market.
The company and its figures are an illustration, not data.
What aging is
Aging moves a pay figure from the date it describes to the date you need, using a rate that stands in for how fast pay moves. Public employers do it in the open. Washington State's 2026 employee pay survey set one common date, 1 January 2026. It aged private market data dated more than 12 months earlier at 3.5% a year, a rate taken from a salary budget survey of Washington employers [1].
The formula is short: Aged figure = reported figure × (1 + yearly rate) raised to the power of months ÷ 12. For gaps under two years, simply scaling the rate by the months gives almost the same answer.
Which rate to use
Two rates are easy to defend. One is a salary increase budget: WorldatWork reported that US budgets averaged 3.6% in 2026, with 3.6% projected for 2027 [2]. The other is a government wage index: Wages and salaries for US civilian workers rose 3.2% in the 12 months to June 2026, with 3.1% in private industry [3]. A budget is what employers plan to spend on raises. The index tracks what the same jobs cost over time. Pick one, use it for every source and write down which one you used.
Use a separate rate only for a job family whose market you can show is moving faster. One rate for everything is less precise but far easier to explain.
Three ways it goes wrong
Aging from the wrong date. The start point is the date the survey data describes, not the date you bought the report or ran the numbers. Many surveys describe pay months before they are published.
Aging past your date without saying so. Aging to the end of your plan year instead of its start builds in a lead over the market. That can be a sound choice, but it is a pay policy decision, not arithmetic. Our guide on which market percentile to target covers that call.
Aging data that is already current. Live data pulled last month needs a month of aging at most. Pushing it forward a full year double counts the market move.
What it means for a company of 50 to 500 people
Record the effective date of every source next to its numbers. Choose one target date and one rate before you look at the results, so the choice cannot bend toward the answer you want. Then age, blend and set midpoints in that order. Our guide Can market data set your salary bands? covers what happens next. When you choose a survey, check that it states its effective date clearly; our guide on which salary survey to buy covers what else to check.
Our view
Age every source to the date your pay decisions take effect, at one stated rate. It takes an hour and it is rarely the step people argue about, yet it moves every midpoint by a few percent. The worst version is the invisible one, where nobody can say what date the bands describe.
Where we help
Every market pricing project ages each source to your effective date and shows the rate we used in the working file. Book a free call.
Sources
- Washington State Office of Financial Management, 2026 Washington State Employee Compensation Survey: Methodologies overview (8 July 2026), section "Aging factors", checked 2026-10-27
- WorldatWork, Workspan Daily: Employers' 2027 pay budget projections point to stability (16 July 2026; 1,799 organizations), checked 2026-10-20
- US Bureau of Labor Statistics, Employment Cost Index, June 2026 (released 31 July 2026), checked 2026-10-27
Changelog: 27 October 2026, first published. The Tampa company and its figures are an illustration, not data.