How to read a productivity number
Output per hour is a ratio, and ratios rise for reasons that are not improvement. A practical reading guide for people who have to repeat a figure in a meeting and defend it afterwards.
What the statistic is asking
Labour productivity is a division problem. Take the real value a business sector adds, divide it by the hours worked to add it, and you have output per hour. The official version for Canada's business sector is published quarterly by Statistics Canada, alongside hourly compensation and unit labour cost.[1]
Three words in that definition do most of the work. Real means the value is adjusted for prices, so a price increase does not masquerade as extra output. Value added means the contribution of the business, not the value of its inputs. Hours worked means hours, not people — a shorter shift, an extra overtime block or an unfilled vacancy all move the denominator.
So the number answers one question: how much real value did an hour of work produce, on average, in this sector and this period? It does not answer whether the work was good, whether it was needed, whether anyone was better off, or whether a particular machine paid for itself.

Staging, not capacity. Units waiting on a pallet are inventory, not output, until they are labelled and released.
Illustration for InSignal Works, 2026
Four shifts the headline hides
1. Composition
An average is a weighted average, and weights move. If a sector's mix tilts toward higher-value work, the average output per hour can rise without any process improving. The reverse is just as easy: a sector can do everything better and still post a weaker average because the low-value work grew faster.
This is why sector-level figures sit awkwardly next to firm-level experience. A manufacturer that invested well can be told the sector went backwards, and both statements can be true at once.
2. Hours are not effort
Hours measure time at work, not intensity or quality. Unpaid overtime is invisible. Vacancies shrink the denominator by leaving work undone. A plant running short-staffed can post better productivity while accumulating delayed maintenance and exhausted people — improvements that are borrowed rather than earned.
3. Prices and volume
Real output is a volume concept, but the world arrives in dollars. In a year when input prices move sharply, the deflators that convert dollars into volume do a lot of quiet work. Two analysts using different deflators can describe the same year differently without either being careless.
4. Capital and technology
A firm can raise output per hour by giving each worker more capital to work with — a faster machine, better software, less waiting. That is a real gain in output per hour, and it is not necessarily a gain in how cleverly the work is organised. Economists separate the two with multifactor productivity, which estimates how much growth is left once measured capital and labour inputs are accounted for. The residual is informative precisely because it is a leftover, and it is a mistake to read it as a clean measure of innovation.
A ratio tells you that something moved. It never tells you which thing moved, or whether the movement was worth having.
Why one quarter is not a trend
On 3 September 2026, Statistics Canada reported that labour productivity in the business sector rose 1.0 per cent in the second quarter of 2026. The release described it as the largest quarterly increase since the second quarter of 2020, when productivity rose 7.7 per cent.[1]
Both facts belong in the same sentence, and the second is the more useful one. The comparison quarter is the first pandemic quarter, when hours collapsed faster than output and the ratio jumped for reasons nobody would call a productivity strategy. A “largest since” comparison is a description of the distribution of past quarters, not a quality judgement about this one.
Quarterly figures are also revised. A first estimate is a first estimate. If a number is going to be used for a decision with a multi-year horizon, it should be treated as a signal about direction — and only after several quarters point the same way.
From a national ratio to a shift
The translation from a national percentage to a shop-floor consequence is where most misunderstandings start. Here is the arithmetic, with invented numbers, clearly labelled as an illustration.
Whether that $24,000 becomes cash depends entirely on what happens to the hours and to the demand for the output. The ratio does not say.
A one per cent gain is small enough to be dismissed in a headline and large enough to fund a hire. That is the scale at which productivity actually operates: not as a quarterly score, but as a slow accumulation of removed friction.
Five questions before you quote it
- Which measure? Labour productivity, multifactor productivity and unit labour cost are different ratios. They can move in opposite directions in the same quarter.
- Which population? Business sector, manufacturing, a province and a two-digit industry are not interchangeable. A national figure cannot settle a plant-level argument.
- What moved — the top or the bottom? Falling hours and rising output look identical in the ratio. They are different stories with different policy consequences.
- What changed in the mix? If the composition of output shifted, say so before comparing periods. A mix-adjusted comparison is a better conversation.
- How old is the number, and is it revised? State the period and the publication date. A dated figure is honest; an undated one is a rumour with a decimal point.
What the ratio refuses to say
Productivity says nothing about defects, lead time, safety, retention or resilience. A process can be more productive and more fragile in the same redesign. It says nothing about distribution: the value can be reinvested, returned, or absorbed by prices. And it says nothing about whether the output was wanted — a firm can produce more of something nobody is buying and record an excellent hour.
Those are not reasons to ignore the number. They are reasons to hold it next to a second number. At the level of a firm, that second number is usually cash. Working out how saved hours become cash — and what happens when they do not — is the subject of the second explainer, and of the worksheet that goes with it.
Sources
Method and citationLabour productivity, hourly compensation and unit labour cost, second quarter 2026
Published 3 September 2026. Read the source. Source of the quarterly business-sector labour productivity figure discussed in this issue.
Canadian ICT Sector Profile 2024
Data year 2024. Read the source. Source of the sector size, growth, employment, salary and research spending figures cited here.
Statistics are reproduced from the publications listed above, with their publication dates. They are fixed reference points, not live indicators. Nothing on this page is a forecast or investment advice.