Is an original series the same as raw data?

Not necessarily; it may already be aggregated or statistically estimated.

Can original values be revised?

Yes, according to the producer's release and correction practices.

Does seasonal adjustment remove all economic cycles?

No. Its usual purpose is to handle recurring seasonal and calendar effects; broader business-cycle movements can remain.

Is the adjusted series always preferable?

No. The choice depends on whether the question concerns actual seasonal quantities, underlying movement, or another specific purpose.

AI Glossary / Time series and signal processing / Intermediate

Original time series

Definition

An original time series is a sequence of observations presented without seasonal adjustment. It retains movements associated with the underlying trend, seasonal or calendar patterns, and irregular influences. In official statistics, original describes the adjustment status of a published series; it does not necessarily mean untouched individual records or the first version ever released.

What it means

An original monthly sales series shows the measured sales associated with each month under its statistical definition. A holiday peak remains visible rather than being removed to help compare adjacent months. That peak may be essential for inventory planning even when it obscures the direction of underlying demand.
Original, seasonally adjusted, and trend estimates answer different questions. The original series describes the measured sequence with its seasonal movements. An adjusted series helps separate those recurring influences, while a trend estimate further smooths short-term irregular movements. None should silently replace the others in a comparison.

How it works

A statistical producer collects or assembles observations, applies its measurement and estimation procedures, and publishes values for successive reference periods. Those procedures may include weighting, aggregation, imputation, or corrections. The resulting series can still be original in the seasonal-adjustment sense.
A decomposition may represent an original observation as trend-cycle plus seasonal and irregular components, or use a multiplicative form when proportional effects are more appropriate. Seasonal adjustment estimates and removes relevant recurring calendar effects; it is not simply removal of every fluctuation.
The components are estimated, not directly observed facts. Revisions to source data or seasonal estimates can change published values. Document the release date, measurement units, frequency, and adjustment label so users can reproduce a historical comparison.

Three practical examples

A store uses original December demand to plan actual deliveries because recurring holiday purchases still need to be supplied.
An analyst uses seasonally adjusted employment to examine short-term movement while retaining the original figures for seasonal staffing patterns.
A forecasting team archives each published vintage of an original economic series so a backtest uses only data available at each forecast date.

Think of it this way

Think of a recording before a particular background pattern has been filtered out. The recording already may have been digitized, calibrated, and organized, yet it still contains that pattern. Similarly, original in this context identifies the absence of seasonal adjustment rather than the absence of every processing step.

Worked example

Suppose an original service-volume series records 120 units in one month and 90 in the next. For an original additive illustration, stipulate estimated seasonal components of +20 and -15 units. Subtracting them gives adjusted values of 100 and 105.
The original month-to-month change is (90 - 120)/120 = -25%. The adjusted change is (105 - 100)/100 = +5%. These statements are compatible: the observed volume declined while the illustrative seasonally adjusted measure increased.
The original decline is 30 units. The seasonal component falls by 35 units, and the adjusted component rises by five; five minus 35 reproduces the original decline. The adjusted values can still contain irregular influences, so their increase is not automatically a five-unit change in the underlying trend.
The seasonal components here are supplied for explanation, not estimated from only two observations. Two months do not provide enough evidence to identify a reliable annual seasonal pattern. The calculation also assumes additive effects; multiplicative adjustment would divide by factors instead.

Illustrative case study

Fictional case: Juniper Parcels reports a sharp January fall after a holiday delivery surge. One dashboard uses original counts, while another shows a modest increase after seasonal adjustment. Managers initially think that one dashboard must be wrong.
The analyst labels both clearly and explains that they address different questions. Operations uses the original workload for capacity accounting, while planning examines adjusted movement alongside irregular events and longer history. This invented case illustrates consistent interpretation, not a measured performance improvement or a claim that adjustment always reveals the correct trend.

Common misconceptions and mistakes

Do not interpret original as error-free, unrevised, or exempt from estimation. Published original values can change when reports arrive late or methods are corrected.
Do not compare an original value in one month with an adjusted value in another and describe the result as an ordinary growth rate. Check whether the units, coverage, and reference periods also match.
Do not assume year-on-year original comparisons eliminate every seasonal complication. Moving holidays, trading-day differences, changing seasonal patterns, and unusual events can still affect interpretation.

Business use

Keeping adjustment labels visible prevents contradictory business stories from being built from different series versions. For actual workload and cash requirements, seasonal peaks remain operationally relevant. For monitoring short-term direction, adjusted values may make comparison more informative.
Retain the original series alongside derived versions, identify the revision vintage, and explain which version supports each decision. When forecasting, test whether modeled seasonality should be reapplied to produce the actual quantities that teams need to plan for. A deseasonalized forecast alone may not represent the expected seasonal workload.

Check your understanding

Why can original volume fall while adjusted volume rises?

Answer: The estimated seasonal decline can exceed the observed decline.

What is the original percentage fall in the example?

Answer: Twenty-five percent.

Are the adjusted values pure trend estimates?

Answer: No; irregular influences may remain.

Could the two seasonal components be reliably estimated from these two observations alone?

Answer: No; the example stipulates them rather than claiming an estimation result.

Related terms

Time series
Seasonal adjustment
seasonality
trend-cycle
irregular component
data vintage
revision
additive decomposition
calendar effects

Learn next

compare original and adjusted versions of one clearly documented series. Trace a seasonal peak, check the release notes for revisions, and explain which version would support a capacity decision versus a short-term trend discussion.

Sources

Sources: Australian Bureau of Statistics, Time series data: https://www.abs.gov.au/statistics/understanding-statistics/statistical-terms-and-concepts/time-series-data . ABS, Time Series Analysis: The Basics: https://www.abs.gov.au/websitedbs/d3310114.nsf/home/time%2Bseries%2Banalysis%3A%2Bthe%2Bbasics . Reserve Bank of Australia, statistical glossary: https://www.rba.gov.au/statistics/glossary-statistical-data.html . Primary indexed passages consulted. Example and case are original.

Previous: Origin function Next: Orthogonal array