How do small firms experience worker churn? Early data from Small Firm Diaries USA

In our last blog post, we explored how employment numbers are recorded among U.S. firms at a single point each year—March 12—and how this approach fails to capture the volatility of employment among small firms with 1–20 workers. In today’s post, we use new data from the SFD USA sample to further illustrate this point.

Our researchers gather the number of workers at the firms every two to four weeks. This higher frequency approach allows us to observe changes as they happen: when a worker is hired, when staffing levels are reduced, when someone leaves voluntarily, or when a position disappears altogether. Over time, these repeated observations reveal patterns that would be hard to see in data that is collected only once every March 12th. 

What does the SFD USA data show us?

Data collected from the first 6 months of the study show a wide variety of employment experiences. More than half of the firms (56%) reported no change in worker count since entering the study. In contrast, nearly one in five firms (19%) reported a net reduction in workers, with declines ranging from the loss of a single worker to reductions of up to 6 workers. At the same time, 25% of firms added workers, ranging from 1 to more than 10 new workers in the same period. 

Beyond net outcomes, we see that firms with identical numbers of workers at their first and most recent data collection points can conceal important fluctuations over time. Ten firms in our sample demonstrate this point—at months 1 and 6 they have the same number of workers but have experienced fluctuations in the intervening months. In several cases, staffing levels doubled or were cut in half before ultimately returning to their original size. 

How static headcount measures can hide volatility: an example from Chicago

A music school in Chicago demonstrates how employment measures that record only net change can mask the change within that firm. When it entered the study in early September, the school had 4 workers. Just a few weeks later, it jumped to 6 but that growth was short-lived, dropping back to 4 in the beginning of October. The school added a new worker in late October, but in November, it had returned once again to 4 workers. 

Why similar headcounts among firms do not mean similar employment experiences: an example from New York

Two firms in New York City illustrate how similar workforce sizes at a single point in time can mask very different employment trajectories. For example, a dance and fitness studio entered the study with 18 workers in August 2025. By early September, they had lost a worker, but gained a new worker a few weeks later, bringing them back to a total of 18. At the end of September, they were down to 14 workers and by mid-October, the studio employed 13 workers, just over 70% of its workforce two months earlier.

By comparison, a childcare provider in the same city ended October with a similar headcount as the dance studio—14 workers—but rather than ups and downs, this firm saw steady growth, adding three new workers between August and October in response to an increase in demand for its services.

Implications

Counting workers at a single point in time masks the true number of hires and departures and obscures the volatility inherent in employment decisions, particularly among small firms. In doing so, it can overstate firm stability while understating the burden on owners navigating the perils of hiring and retaining skilled workers.

Without accurate data representing the true state of small firm employment, programs, policies, and products may not be as effective at supporting small firms and their workers as may be possible.

As we continue to collect data on the firms in our sample, we’ll dig deeper into what drives these changes such as how cash flow, access to credit, owner aspirations, and external shocks shape employment decisions over time.