William Beach, BLS chief from 2019 to 2023

Beach helmed the BLS during the first Trump administration and into the Biden administration. Without the monthly Employment Situation Summary — the official label for the jobs report — he’s relied on models from Moody’s and S&P Global. 

“If I was chief economist for a large company, I would say let's just use the forecasting models for the next few weeks,” Beach told me. 

Another popular, private-sector alternative is the payroll processor ADP. The firm draws from the data of 26 million workers at its client firms to measure employment levels. It still relies, though, on BLS figures for sizing up different industries across the U.S. economy.

ADP data does have its limits, Beach says. Its information is likelier to draw from medium and larger companies that can afford to outsource payroll and human resources functions to an external firm. That risks overlooking anywhere from six to eight million small businesses that technically don’t have employees, Beach says. For example, a restaurant could count family members as staff who are compensated with a portion of the sales.

In addition, ADP clients are clustered on the coasts rather than the interior of the country, with less representation around metropolitan areas of the South.

“They are one of the most successful companies, but they don't have a customer base that represents the businesses in this country,” Beach told me. “They represent a slice of it.”

Erica Groshen, BLS chief from 2013 to 2017

Groshen, who is now a Senior Economics Advisor at the Cornell University School of Industrial and Labor Relations, has been checking ADP along with a newer gauge.

“ADP is one that, of course, I'll look at,” Groshen told me. “Revelio Labs has a series of new kinds of estimates obtained from various kinds of web scraping that they do.”

Revelio is a newcomer to the employment data environment. It scrapes statistics and on employment levels and wages from more than 100 million U.S. profiles from social networking platforms such as LinkedIn.

For September, Revelio calculated that 60,000 jobs were added, with two-thirds of employment growth stemming from the health, education, and retail sectors. It estimated that BLS would have reported 38,000 jobs gained in September.

Like ADP, there are drawbacks to excessive reliance on Revelio. Groshen described it as a “convenience sample” that’s easy to collect but might include unintended biases in the data. For example, it could oversample for an industry like tech or manufacturing.

“They can't fully substitute,” Groshen said, “because they just don't have the breadth, the transparency, the historical track record data that really allows you to compare current conditions to the past.”

—Joseph Zeballos-Roig

Joseph Zeballos-Roig is Quartz’s Washington Correspondent. Email him at jzeballos-roig@qz.com and follow him on X at @josephzeballos.