Founder Data ReportMore Small Firms Report Revenue Down Than Up. In All 74 Fortnights on Record.
Get clarity — download the free Entrepreneur Survival Guide Summary.
Census asks every size of business the same question in the same fortnight. Firms with 1-4 employees have answered ‘down’ more often than ‘up’ every time since 2023. Firms with 250 or more have answered the opposite.
Quick answer: In the latest Census fortnight, 30.6% of businesses with 1-4 employees said revenue had fallen and only 13.6% said it had risen. At businesses with 250 or more employees the same fortnight reads 15.2% down against 29.6% up – the mirror image. Across all 74 published cycles since 2023, more small businesses have reported revenue falling than rising in every single one. Large businesses have reported the opposite in 69 of them.
- Latest fortnight, revenue: 1-4 employees runs 30.6% down against 13.6% up. 250-plus runs 15.2% down against 29.6% up. Same fortnight, same economy, opposite sign.
- The net reading – up minus down – falls in a straight ladder across the seven size bands: -17, -15, -9, -7, +2, +6, +14. It crosses from negative to positive between 20-49 employees and 50-99 employees.
- This is not one bad quarter. Over all 74 cycles, the smallest band’s net revenue reading was negative 74 times out of 74. The largest band’s was negative 5 times.
- It is not simply that small firms are lumpier. The 1-4 band reports NO change in revenue 55.3% of the time, more often than the 10-19 employees band at 50.5%. They report change less often and still skew down.
- Demand tells the same story a step earlier: 23.7% of the smallest businesses reported demand falling against 9.5% of the largest.
One fortnight, seven different economies
Every two weeks the Census Bureau asks US businesses whether their operating revenue rose, fell or held steady over the previous fortnight, and publishes the answers cut by how many people the business employs. It is the same question, asked in the same window, of businesses living in the same economy.
For the cycle covering 10 August to 23 August 2026, businesses with 1-4 employees answered 30.6% down and 13.6% up. That is 2.2 businesses reporting a fall for every one reporting a rise. Businesses with 250 or more employees answered 15.2% down and 29.6% up – about 1.9 rises for every fall.
Between those two ends the reading does not jump. It slides. Net of increases and decreases, the seven bands run -17.0, -14.6, -9.0, -7.4, +2.4, +5.9, +14.4 points, and the sign changes between 20-49 employees and 50-99 employees. Somewhere around fifty people on the payroll, the economy stops being something that happens to you and starts being something you are part of.
Source: U.S. Census Bureau, Business Trends and Outlook Survey (BTOS), response estimates by employment size class, 74 published fortnightly collection cycles from 11 September 2023 to 23 August 2026, downloaded from census.gov on 9 September 2026 and last updated at source on 26 August 2026. Q4 asks how operating revenues changed in the last two weeks and Q10 asks the same of demand, so every figure is the SHARE OF BUSINESSES reporting a direction of change over a fortnight – not the size of the change, not a level, and not weighted by how large each business is. A net figure is the share reporting an increase minus the share reporting a decrease. Employment size bands are the Census codes A to G. Three cycles (202521, 202522 and 202523, covering 6 October to 16 November 2025) were cancelled by the federal funding lapse and are absent rather than zero, and small cells in this file are disclosure-suppressed. BTOS refreshes its sample once a year; both series were tested for that break at each end of the size range before any multi-year claim was made, and the refresh steps sit inside ordinary fortnightly variation.
The ladder, band by band
Put both sides of the answer on one track and the pattern is hard to unsee. At every size below 50-99 people, the decrease dot sits to the right of the increase dot. At every size above it, they have swapped.
Demand runs the same ladder one step ahead of revenue, which is what you would expect if demand is the cause and revenue the consequence. 23.7% of the smallest businesses reported demand down against 11.1% reporting it up. At 250-plus employees it is 9.5% down against 19.1% up.
Hiring follows the same shape. In this fortnight 3.3% of the smallest businesses added anyone and 7.9% lost someone; at 250-plus it is 22.6% adding against 10.1% losing. The employment gradient is not a separate finding – it is the revenue gradient turning up in the payroll a few weeks later.
The track runs from 0 to 36%, so the length of each bar is the gap between the two answers in that band. Read the printed figures rather than the bar lengths when comparing two bands. The remainder in each band reported no change, which is why the two dots do not sum to the width of the track.
Source: U.S. Census Bureau, Business Trends and Outlook Survey (BTOS), response estimates by employment size class, 74 published fortnightly collection cycles from 11 September 2023 to 23 August 2026, downloaded from census.gov on 9 September 2026 and last updated at source on 26 August 2026. Q4 asks how operating revenues changed in the last two weeks and Q10 asks the same of demand, so every figure is the SHARE OF BUSINESSES reporting a direction of change over a fortnight – not the size of the change, not a level, and not weighted by how large each business is. A net figure is the share reporting an increase minus the share reporting a decrease. Employment size bands are the Census codes A to G. Three cycles (202521, 202522 and 202523, covering 6 October to 16 November 2025) were cancelled by the federal funding lapse and are absent rather than zero, and small cells in this file are disclosure-suppressed. BTOS refreshes its sample once a year; both series were tested for that break at each end of the size range before any multi-year claim was made, and the refresh steps sit inside ordinary fortnightly variation.
The economy is not one weather system. It is seven, and yours is decided less by your industry than by your size.
Three years, not one bad fortnight
A single cycle proves nothing. So take the net reading – the share reporting an increase minus the share reporting a decrease – and average it over all 74 cycles Census has published since 2023. The ladder survives intact, and it is steeper than any one fortnight suggests.
Businesses with 1-4 employees average -20.4 points on revenue across three years, and the reading has been negative in 74 of 74 cycles. Businesses with 250 or more average +5.6 and have been negative in 5 of 74. In 74 of the 74 cycles the smallest band’s net reading was below the largest band’s, which is to say: in all of them.
That covers three years, an inflation cycle, a rate cycle, a federal funding lapse and a national election. There is no fortnight in the record where the smallest businesses had the better of it. The gradient is not a phase of the cycle – it appears to be the shape of the thing.
Bars run left of the centre line for a negative net reading and right for a positive one, scaled to the largest bar on the chart (20.4 points). A net of -20.4 means that in an average fortnight, 20.4 more businesses in every hundred reported revenue falling than reported it rising. It does not mean revenue fell by 20.4%.
Source: U.S. Census Bureau, Business Trends and Outlook Survey (BTOS), response estimates by employment size class, 74 published fortnightly collection cycles from 11 September 2023 to 23 August 2026, downloaded from census.gov on 9 September 2026 and last updated at source on 26 August 2026. Q4 asks how operating revenues changed in the last two weeks and Q10 asks the same of demand, so every figure is the SHARE OF BUSINESSES reporting a direction of change over a fortnight – not the size of the change, not a level, and not weighted by how large each business is. A net figure is the share reporting an increase minus the share reporting a decrease. Employment size bands are the Census codes A to G. Three cycles (202521, 202522 and 202523, covering 6 October to 16 November 2025) were cancelled by the federal funding lapse and are absent rather than zero, and small cells in this file are disclosure-suppressed. BTOS refreshes its sample once a year; both series were tested for that break at each end of the size range before any multi-year claim was made, and the refresh steps sit inside ordinary fortnightly variation.
It is not just that small firms are lumpier
The obvious objection is volatility. A four-person business loses one client and its revenue moves 20%; a thousand-person business barely notices the same client. If small firms simply report more change in both directions, a gradient in the down answers proves nothing.
The third answer settles it. Across the full record, businesses with 1-4 employees report no change in revenue 55.3% of the time – more often than the 10-19 employees band, which reports it 50.5% of the time. The smallest businesses report a change less often than mid-sized ones, and when they do report one it is far more likely to be a fall. Volatility does not explain that.
One honest caveat, because the check is not perfectly clean at both ends: the largest businesses report no change most often of anyone, 58.7%, so some of their calmer reading is genuinely less movement rather than better movement. The volatility story is ruled out at the bottom of the range. What it does not do is tell you why the small-business answer skews down, and this table cannot settle that.
If small businesses were simply noisier, this line would fall as size falls. It does not: it dips in the middle of the range and rises at both ends. Bars are drawn neutrally because the finding is the shape, not any single band. Note the scale – every band sits between about half and three-fifths, so the bar lengths exaggerate a spread of roughly eight points.
Source: U.S. Census Bureau, Business Trends and Outlook Survey (BTOS), response estimates by employment size class, 74 published fortnightly collection cycles from 11 September 2023 to 23 August 2026, downloaded from census.gov on 9 September 2026 and last updated at source on 26 August 2026. Q4 asks how operating revenues changed in the last two weeks and Q10 asks the same of demand, so every figure is the SHARE OF BUSINESSES reporting a direction of change over a fortnight – not the size of the change, not a level, and not weighted by how large each business is. A net figure is the share reporting an increase minus the share reporting a decrease. Employment size bands are the Census codes A to G. Three cycles (202521, 202522 and 202523, covering 6 October to 16 November 2025) were cancelled by the federal funding lapse and are absent rather than zero, and small cells in this file are disclosure-suppressed. BTOS refreshes its sample once a year; both series were tested for that break at each end of the size range before any multi-year claim was made, and the refresh steps sit inside ordinary fortnightly variation.
What to do with this if you are running one
Stop reading the national number as your number. When a headline says demand is holding up, it is describing an economy whose weight sits with businesses of 50-99 employees and more. In the same fortnight, 23.7% of four-person businesses said demand fell. Both are true. Only one is about you.
Benchmark against your band, not against the economy. A quarter where more weeks went down than up is, for a business your size, the ordinary condition – it has been the ordinary condition in 74 of the last 74 fortnights. That is not a reason to relax. It is a reason to stop treating it as evidence that something is uniquely wrong with you.
Assume the recovery reaches you last. The gradient runs the same way in good fortnights as bad ones. When conditions improve, the 250 or more employees band’s net reading goes positive first and by more. Plan the cash for a turn that arrives at your size later than it arrives in the news.
Watch demand, not revenue. Demand moves first and moves less noisily – the smallest band reports demand down 23.7% of the time against revenue down 30.6%. Whatever you use as your own demand signal – enquiries, quotes sent, pipeline added – it will tell you weeks before the bank balance does.
Frequently Asked Questions
Are small businesses doing worse than large ones?
On the Census Bureau's fortnightly measure, consistently yes. In the latest cycle 30.6% of businesses with 1-4 employees reported revenue falling against 13.6% reporting it rising, while businesses with 250 or more reported 15.2% falling against 29.6% rising. Across all 74 published cycles the smallest band's net reading was negative 74 times and the largest band's 5 times. Note this counts businesses, not dollars.
Does this mean small-business revenue is falling?
No, and the distinction matters. BTOS counts how many businesses report a direction of change, not how much money changed hands, and it does not weight a business by its size. A net of -20.4 points means 20.4 more businesses per hundred reported a fall than a rise. Most firms shrinking a little while a few grow a lot would produce exactly this pattern alongside a growing total.
Isn't this just because small businesses are more volatile?
That is the first thing to check, and the data does not support it at the small end. Businesses with 1-4 employees report NO change in revenue 55.3% of the time, more often than the 10-19 employees band at 50.5%. They report a change less often than mid-sized businesses and still skew heavily toward decreases. The largest band does report no change most often of all, at 58.7%, so part of its calm is genuinely less movement.
At what size does the pattern flip?
In the latest cycle the net revenue reading crosses from negative to positive between 20-49 employees (-7.4 points) and 50-99 employees (+2.4). Averaged over all 74 cycles the crossing sits higher again, between 100-249 employees (-0.7) and 250 or more employees (+5.6). The boundary moves with conditions; the ordering does not.
How current is this data?
The most recent published collection cycle ran 10 August to 23 August 2026. BTOS publishes fortnightly, so the lag is short. Three cycles between 6 October and 16 November 2025 were cancelled by the federal funding lapse and do not exist rather than reading zero, which is why the record covers 74 cycles rather than 77.
Published by The Lonely Entrepreneur — the community and coaching platform for entrepreneurs who are building alone. lonelyentrepreneur.com
This article is for educational purposes and is not a substitute for professional financial or legal advice.