CHOICES
A publication of AAEA
The dairy industry is a cornerstone of US agriculture, particularly in California and Wisconsin—the two largest dairy-producing states. Both states jointly produce a third of the nation’s milk and employ about 35,000 farmworkers (Jette-Nantel, 2018; Matthews and Sumner, 2019). However, dairies in both states face labor shortages, rising wages, and difficulties in retaining and managing workers—issues that worsened during the COVID-19 pandemic and have continued since (Peña-Lévano, Burney, and Adams, 2020; Liebrand, 2022). In response, dairy farmers are exploring automation in dairy activities to decrease labor reliance (Peña-Lévano, Burney, and Beaudry, 2023). In this article, we examine the findings of a large-scale survey conducted in 2024 aimed at assessing the perceptions of dairy farmers in California and Wisconsin regarding labor challenges and automation, with emphasis on robotic milking. The study reveals significant regional differences in labor sources as well as the key barriers and drivers for adopting automated milking systems (AMS), offering valuable insights into the future of the dairy industry.
Multiple processes within dairy operations can be automated. However, while AMS are gaining attention as a technology that can be used to decrease the reliance of manual labor, they require a large up-front investment and often modification of the barn infrastructure and changes in farm management (Peña-Lévano et al., 2025).
AMS are robots with a hydraulic arm, lasers, and cameras that allow the teat cups to autonomously attach to the cows’ udder for a hands-free milking operation. Overall, each robot box can milk 60–70 cows per day (Peña-Lévano et al., 2025). AMS systems are availablein two configurations: (1) linear, where cows approach the robot at their own pace, and (2) rotary systems, in which dairy cows are guided by workers into the rotating milking parlor (Siewert, 2017; Marques et al., 2023). The leading linear AMS technologies in the US are DeLaval’s VMS robots and Lely’s Astronaut, with emergent competition from GEA’s R9500, Boumatic’s Gemini and Galaxy USA’s Merlin (Marques et al., 2023; Peña-Lévano, Burney, and Beaudry, 2023). Linear AMS parlors may require new barn construction or retrofitting an existing barn. New barns are designed either in a free-flow, where cows move freely throughout the barn, or a guided-flow setup that uses gates and barriers to direct the cows to the robots (Marques et al., 2023).
Figure 1 explains the milking process in a free-flow linear robotic setup. Each animal has a unique identifier that allows the AMS to measure close to 100 different parameters and track a wide range of individual cow data (such as milk yield, somatic cell count, health metrics, milking frequency, activity levels, rumination patterns, and feed intake). Once a cow enters the robot, it is identified using its collar, and the system determines whether to accept or reject the cow based on milking permission settings. Once a cow is inside the robot, the first step is udder identification, where sensors and cameras locate the teats. After identifying their position, the prepping process begins, with automated brushes or cleaning cups disinfecting and cleaning the teats. To encourage visits to the robot, farms offer feed, though the type varies: some provide pellets, others a meal. Once the cow is prepped, the system attaches milking cups to each teat, and the milking process begins. After milking, the system applies a teat disinfectant to prevent infections. Finally, the gate opens, allowing the cow to exit and rejoin the herd (Peña-Lévano et al., 2025).
In early 2024, we conducted a comprehensive survey in California and Wisconsin, targeting 2,000 dairy farmers from both states. The survey covered labor availability, technology adoption (AMS and other automation), perceptions and utilization of dairy support programs, and sustainable practices. Sections of the questionnaire used in this study—which focuses on farmer and farms’ characteristics, labor, and automation—are described in Table 1.
The Survey Research Center (SRC) at the University of Wisconsin–River Falls implemented the survey on our behalf, which included survey distribution, data collection, cleaning, and compilation. The SRC utilized the Dillman (1978) Total Design Method to ensure systematic and robust data collection. The approach included sending a postcard reminder to nonrespondents 3 weeks after the initial mailing, followed by a second survey to nonrespondents 3 weeks after the reminder postcard. Based on survey research literature (Hoddinott and Bass, 1986) and prior experience of the SRC, this method is documented to boost the survey response rates in a cost-effective manner. Data collection was conducted over an 8-week period and concluded in April 2024.
To construct statistical estimates with a 5% margin of error and within a 95% confidence interval, a minimum sample size of 363 responses was needed. Stratified sampling was utilized to ensure that the overall sample is balanced relative to the population of dairy farmers in each state. The contact information of the 2,000 survey recipients of both states was obtained from multiple sources. For California, the SRC used the US Farm Service Agency (FSA) mailing list of all payment recipients from the Dairy Margin Coverage program—a low-cost insurance program in which most dairy farmersparticipate. For Wisconsin, the SRC utilized the list of licensed milk producers obtained from the Wisconsin Department of Agriculture, Trade, and Consumer Protection (DATCP).
The survey yielded 556 complete responses—100 from California and 456 from Wisconsin—resulting in a 28% response rate and exceeding the minimum sample size needed for statistical validity, which provides robustness to our analysis. The following sections discuss the major findings related to key challenges in labor retention and supervision, barriers and motivators for AMS adoption, and adoption of additional automated technologies in dairy operations.
Despite comparable contributions to the US milk supply, dairies in California and Wisconsin operate under distinct economic and operational structures. As the nation’s largest milk producer, California benefits from atopography and climate that is favorable for cattle andmilk production, supporting a flourished dairy industry (Sumner, 2020). In 2021, California dairies produced 41.9 billion pounds of milk—about 19% of the US milk supply—generating $57.7 billion in economic activity (California Dairy Press, 2022).
The most common nonrobotic milking facility reported by our California survey respondents was the herringbone pit parlor—where cows stand on an elevated platform at a 45-degree angle to the operator pit, allowing easy access to the cows’ udders. Most of the dairy farms (87%) were nonorganic (conventional), with an average of 918 acres and 2,143 dairy cows in their operations, largely comprised of Holstein cows (displayed in Table 2). This aligns with Sumner (2020), who states that approximately 70% of the dairy sector in California consists of midsized and large operations (>500 cows). The potential for economies of scale and the need for skilled dairy managers are among the reasons motivating the consolidation of the California dairy industry. In 2007, there were about 2,000 dairy farms in the state; by 2017, only 1,279 remained in operation (Sumner, 2020).
Wisconsin—commonly known as America’s Dairyland—has the largest number of dairy farms in the United States. However, this number has dropped dramatically, faster than in California, from 11,542 herds in March 2012 to only 5,321 license herds in March 2025 (USDA-NASS, 2025). This means that more than half (54%) of Wisconsin dairy farms closed operations in just 13 years. Despite this decline in the number of farms, overall milk production in Wisconsin has increased by about 18% during this period and the state of Wisconsin currently accounts for 14% of the US milk output, making it the second largest milk producer (USDA-NASS, 2025; Peña-Lévano et al., 2025).
Unlike California, Wisconsin dairy operations tend to be much smaller, an average of 608 acres and 247 dairy cows, according to our survey responses. Parabone pitparlors—which arrange cows at a 70-degree angle to theoperator pit—were reported to be the most common facility (Table 2), likely to allow more cows (in a smaller space) relative to herringbone pit parlors.
In California, senior farmers (56+ age group) comprised more than half of the respondents (58%), with 70% of owners earning more than $75,000 annually. In contrast, most of Wisconsin farmers (59%) are in the middle-age group (35–55 years old), with only half (50%) of the dairies earning more than $75,000. These results can be attributed to the difference in the size of operations and ownership in both states.
Decomposing the labor force into full-time and part-time employees (as depicted in Figure 2), the survey shows that about 91% of California farms hire external labor, with immigrants accounting for 43% of the total workforce—aligning with the findings of Charlton and Kostandini (2021). Family members support the dairy operation on a part-time basis, constituting almost half of the part-time employees. The sum of both payrolls constitutes about 12% of the cost of the California milk production (Sumner, 2020).
In Wisconsin, most of the dairies (about 95%) are family-owned (Peña-Lévano, Burney, and Beaudry, 2023), relying heavily on family members, which comprise almost half of the full-time employees. Most part-time workers (84%) were reported to be either domestic labor or family relatives, making the labor sources (statistically) different from California.
The adoption of AMS in California and Wisconsin is still in its infancy. Only 4.7% of surveyed respondents (hereafter referred to as AMS farmers) reported that they have already adopted robotic milking. The majority of farmers (approximately 80.5% from both states) reported that they operate with manual (or conventional) milkingsystems and would not adopt AMS. This estimate is consistent with the 4% AMS adoption rate in Wisconsin unveiled in the 2024 DATCP survey (DATCP, 2024). Similarly, a recent exploratory survey conducted by Darby (2022) found that only 8.7% of respondent farmers utilize AMS in the US Northeast.
A third group—categorized as potential adopters (14.8% of all our respondents)—expressed that they currently do not have AMS but are considering making the investment. Figure 3 shows how sociodemographic characteristics differ across the three groups (AMS farmers, potential adopters, and manual systems). Consistent with Peña-Lévano, Burney, and Beaudry (2023), experienced farmers—who usually own larger operations—are keener to adopt AMS due to their financial capacity and as a strategy to reduce their working hours and thus improve their quality of life.
Most farmers in all groups (30%–37%) fall into the middle-income category, earning between $75,000 and $150,000 per year. Almost a quarter of potential adopters and AMS farmers reported earnings over $150,000, compared to only 12% of farmers who prefer manual milking systems. This can be partially attributed to the size of their operations. While farmers with robotic milking or contemplating their adoption tend to farm the most land (AMS farmers: 965, potential adopters: 839acres), manual farmers run smaller operations (average = 621 acres). Likewise, most AMS adopters (88%) reported a higher milking frequency (more than twice a day) compared to 79% of manual milkers who milk twice daily. This is likely due to the labor required for manual milking, which is challenging amid labor shortages, especially in Wisconsin, as noted anecdotally.
The survey reveals similar reasons why AMS is adopted in California and Wisconsin. Among possible drivers, dairy farmers in both states indicated their primary motivation for automation is to reduce reliability on employees. This result was expected because labor shortages are of major concern for the dairy industry (Charlton and Kostandini, 2021). In California, the limited supply of domestic farmworkers cannot meet the need for full-time, year-round employees. In Wisconsin, where farms are largely family-run, the shortages come in the form of part-time workers that provide intermittent support in tasks such as milking, fetching cows, and managing herds. This situation is exacerbated by the incompatibility of the H-2A guest worker program, as the seasonal nature of H-2A employment does not align with the year-round demands of dairy farming (Escalante et al., 2019).
Farmers with aging milking systems (that will soon need to be replaced or will require a major repair) were motivated to automate their operations and thus remain competitive. Furthermore, the prospect of higher milk yield from cows voluntarily milking about three times daily incentivizes AMS adoption by offering potential revenue gains (Salfer et al, 2017). Wisconsin farmers also responded that their intentions of expanding herd sizes align with the automation of the milking process, consistent with the positive correlation between both practices (Martinsson et al., 2024).
The survey also shows that—independently from the state—the major barrier to modernize a parlor to a fully-robotic setting is the high upfront cost, estimated at approximately $250,000 per robot box. Considering each box milks between 60–70 cows daily, an average 250-cow Wisconsin farm would require an initial investment of about $1,000,000. In addition, the farm must incur expenses for either new infrastructure (free-flow or guided barn) or retrofitting an existing barn (Peña-Lévano et al., 2025). Maintenance expenses—ranging from $7,500 per year in the initial stages to as much as $16,500 annually as the system ages—further discourage investment (Bentley, Schulte, and Tranel, 2018). These barriers may particularly constrain middle- and low-income farmers from installing AMS in their operations, as it can represent a large financial burden.
Dairy farmers that perceived their current system as still functioning efficiently and those that were unclear about the benefits of AMS also expressed no interest in purchasing robotic milking systems. This can be attributed to the fact that despite the vast literature studying AMS in Europe and New Zealand, there are limited studies addressing the feasibility of AMS in the United States. Likewise, previous research also provides conflicting findings. These are summarized in a review by Jacobs and Siegford (2012), which attributes AMS benefits to improvements in management practices and facility design when the technology is adopted. In the same line, Steeneveld et al. (2012) report negligible differences in labor costs, net output, or technical efficiency in their Dutch study. Thus, it is still not clear whether this technology’s net benefits outweigh its costs (Peña-Lévano, Burney, and Beaudry, 2023). Last, Wisconsin dairy farms also report that without a clear successor to take over the operation, long-term capital investments such as AMS become less appealing to farmers.
In addition to labor shortages, the dairy sector also faces issues with labor retention, with a national turnover ratio averaging 38.8% in 2019 (Ribero, Adcock, and Anderson, 2020). Turnover imposes severe financial burden on dairy operations, estimated at 150% to 250% of a worker’s annual wage due to recruitment, onboarding, and productivity losses (Billikopf and González, 2012). Our survey (Figure 4) underscores the severity of these challenges. One-third of dairy farmers reported substantial difficulty in securing skilled labor, while more than half struggled with employee retention. Notably, farmers considering AMS adoption perceived greater challenges in hiring, supervising, training, and retaining workers than those who are largely uninterested in automation, suggesting that AMS serves as a strategic response to labor market constraints. However, farmers with a manual milking system had a similar satisfaction with and ease in finding, training, and retaining workers as AMS farmers, which explains why they chose not to adopt AMS.
Beyond economic implications, labor retention is also a social concern. Previous studies have emphasized the fact that wage increases alone do not fully resolve the issue. In a 2009 survey, Billikopf and González (2012) interviewed 209 dairy workers in selected California counties, finding that, in addition to compensation and benefits, workers also left their employer because of existing economic problems at the operation, working schedule, and family matters.
Beyond AMS, dairy farms are integrating a range of automated technologies to reduce labor reliance in specific tasks such as cleaning and feeding (Garcia-Covarrubias et al. 2023). Among the most common ancillary systems are automatic washers, cluster removers, scrapers, and feeders.Automated washers play a critical role in maintaining hygiene by using high-pressure water and chemical detergents to sanitize milking equipment and parlors, reducing bacterial contamination and the risk of mastitis. Automatic cluster removers prevent overmilking bydetaching milking clusters once milking is complete, thereby reducing teat stress and improving cow comfort. Automated scrapers help maintain barn cleanliness by systematically removing manure from alleys, while automated feeders deliver precise portions of feed tailored to each cow’s nutritional needs, optimizing herd management and efficiency (Palma-Molina et al., 2023)
Our survey results indicate that more than two-thirds of responding farmers (N = 403) have adopted some form of automation (illustrated in Figure 5), with automatic washers being the most prevalent (33.4%), followed by automated cluster removers (26.4%), automatic scrapers (9.4%), and automated feeders (4.0%). While automated washers are considered standard in modern dairy operations, responses suggest that some farmers may not perceive their systems as fully automated or only report recent investments in automation.
Dairy farms in California and Wisconsin face persistent challenges in recruiting, supervising, training, and retaining skilled labor. These issues—particularly labor shortages and high turnover—are key factors motivating some producers to consider automating the milking process. Although AMS adoption remains limited, approximately 15% of surveyed producers identified themselves as potential adopters, reflecting the growing interest in automation as a tool to improve labor efficiency and reduce dependency on increasingly hard-to-find workers.
However, this interest is far from uniform across all dairy operations. Our results suggest a more nuanced picture: One group of manual milking farmers, often operating at a smaller scale with more stable and family-based labor, reports fewer challenges in worker management and show limited interest in transitioning to AMS. In contrast, traditional farmers managing larger herds and experiencing labor challenges such as difficulty with recruitment, supervision, and retention express greater openness to adopting AMS technologies. This distinction highlights that labor pressures might be concentrated in certain segments of the dairy sector.
Among AMS adopters, average profits appear higher than nonadopters; however, these differences may broadly reflect variations in farm size rather than the technology itself. Previous research remains inconclusive on whether AMS-driven milk yield and operational efficiency improvements sufficiently offset the substantial investment and maintenance costs. Given these uncertainties and financial constraints, many producers are turning to ancillary technologies— such as automatic washers, cluster removers, feeders, and manure scrapers—as partial solutions to automate specific labor-intensive tasks without the full commitment to robotic milking systems.
The findings from this study offer important insights for dairy producers, policymakers, and technology developers. Our results highlight the value of weighing AMS adoption against farm-specific factors such as labor reliability, operational scale, and capital availability for medium to large dairy operations. In regions where farms depend more heavily on hired labor, policymakers might explore targeted incentives as a strategy to encourage automation. Moreover, AMS manufacturers and service providers can use these findings to adapt their technologies to better fit the needs of different farm sizes and management styles.
Bentley, J., K. Schulte, and L. Tranel. 2018. “Economics of Automatic Milking Systems.” Iowa State University Extension and Outreach. Available online: https://www.extension.iastate.edu/dairyteam/files/documents/economics_of_automatic_milking_systems_v2.1_2018.pdf
Billikopf, G., and G. González. 2012. “Turnover Rates Are Decreasing in California Dairies.” California Agriculture 66(4). https://doi.org/10.3733/ca.v066n04p153
California Dairy Press. 2022. “California: The Nation’s Dairy Leader” [fact sheet]. Available online: https://californiadairypressroom.com/sites/default/files/2022-08/2022_Fact Sheet_Nations Dairy Leader_CS2204D.pdf
Charlton, D., and G. Kostandini. 2021. “Can Technology Compensate for a Labor Shortage? Effects of 287 (g) Immigration Policies on the US Dairy Industry.” American Journal of Agricultural Economics 103(1):70–89. https://doi.org/10.1111/ajae.12125
Darby, H. 2022. Understanding Opportunities and Risks Associated with Alternative Milking Strategies. Final report for ONE20-360. Sustainable Agriculture Research and Education Projects. Available online: https://projects.sare.org/project-reports/one20-360/
Dillman, D.A. 1978. Mail and Telephone Surveys: The Total Design Method. Wiley.
Escalante, C.L., O. Williams, H. Rusiana, and L. Peña-Lévano. 2019. “Costly Foreign Farm Replacement Workers and the Need for H-2A Reforms.” Journal of ASFMRA:14–20. https://doi.org/10.22004/ag.econ.322659
Garcia-Covarrubias, L., D. Läpple, E. Dillon, and F. Thorne. 2023. “Automation and Efficiency: A Latent Class Analysis of Irish Dairy Farms.” Q Open 3(1):qoad015. https://doi.org/10.1093/qopen/qoad015
Hoddinott, S.N., and M.J. Bass. 1986. “The Dillman Total Design Survey Method.” Canadian Family Physician 32:2366.
Jacobs, J., and J. Siegford. 2012. “The Impact of Automatic Milking Systems on Dairy Cow Management, Behavior, Health, and Welfare.” Journal of Dairy Science 95(5):2227–2247. https://doi.org/10.3168/jds.2011-4943
Jette-Nantel, S. 2018. “Farm Employment in Wisconsin.” University of Wisconsin Center for Dairy Profitability. Available online: https://farms.extension.wisc.edu/articles/farm-employment-in-wisconsin/ [Accessed March 19, 2025]
Matthews, W., and D. Sumner. 2019. Contributions of the California Dairy Industry to the California Economy in 2018. A Report for the California Milk Advisory Board, University of California Agricultural Issues Center. Available online: https://cail.ucdavis.edu/wp-content/uploads/2019/07/CMAB-Economic-Impact-Report_final.pdf
Liebrand, C. 2022. “Dairy Outlook: 2022 Agricultural Outlook Forum.” US Department of Agriculture. Available online: https://www.usda.gov/sites/default/files/documents/2022AOF-dairy-outlook.pdf
Marques, T., C. Lage, C., D. Bruno, E. Fausak, M. Endres, , F. Ferreira, and F. Lima. 2023. “Geographical Trends for Automatic Milking Systems Research in Non-Pasture-Based Dairy Farms: A Scoping Review.” Journal of Dairy Science 106(11):7725–7736. https://doi.org/10.3168/jds.2023-23313
MacFarland, T.W., and J.M. Yates. 2016. “Mann–Whitney U Test.” In Introduction to Nonparametric Statistics for the Biological Sciences Using R. Springer, pp. 103–132. https://doi.org/10.1007/978-3-319-30634-6_4
Martinsson, E., H. Hansson, K. Mittenzwei, and H. Storm. 2024. “Evaluating Environmental Effects of Adopting Automatic Milking Systems on Norwegian Dairy Farms.” European Review of Agricultural Economics 51(1):128–156. https://doi.org/10.1093/erae/jbad041
Palma-Molina, P., T. Hennessy, A. O’Connor, S. Onakuse, N. O’Leary, B. Moran, and L. Shalloo. 2023. “Factors Associated with Intensity of Technology Adoption and with the Adoption of 4 Clusters of Precision Livestock Farming Technologies in Irish Pasture-Based Dairy Systems.” Journal of Dairy Science 106(4):2498–2509. https://doi.org/10.3168/jds.2021-21503
Peña-Lévano, L., S. Burney, and C. Adams. 2020. “Labor Disruptions Caused by COVID-19 in the U.S. Agriculture and Nonfarm Industries.” Choices 35(3):1–12. https://doi.org/10.22004/ag.econ.305276
Peña-Lévano, L., S. Burney, and J. Beaudry. 2023. “Automatic Milking Systems: An Exploratory Study of Wisconsin Dairy Farms.” Journal of ASFMRA:74–82. https://doi.org/10.22004/ag.econ.342887
Peña-Lévano, L., S. Burney, J. Salfer, J. Clark, L. Garcia-Covarrubias and C. Escalante. 2025. “Automated Milking Systems: A Case Study of a U.S. Midwest Dairy Farm Decision-Making Process.” Applied Economics Teaching Resources. https://doi.org/10.71162/aetr.652617
Ribero, L., F. Adcock, and D. Anderson, 2020. A National Survey of Hiring, Compensation and Employee Treatment Practices on U.S. Dairy Farms. CNAS Report 2020-1. Available online: https://nationaldairyfarm.com/dairy-farm-standards/workforce-development/
Salfer, J., K. Minegishi, W. Lazarus, E. Berning, and M. Endres. 2017. “Finances and Returns for Robotic Dairies.” Journal of Dairy Science 100(9):7739–7749. https://doi.org/10.3168/jds.2016-11976
Siewert, J.M. 2017. “Housing, Management, and Factors Associated with Efficiency of Automatic Milking Systems on Midwest US Dairy Farms.” MS thesis. University of Minnesota. Available online: https://core.ac.uk/download/pdf/211348981.pdf
Steeneveld, W., L. Tauer, H. Hogeveen, and A. Lansink. 2012. “Comparing Technical Efficiency of Farms with an Automatic Milking System and a Conventional Milking System.” Journal of Dairy Science 95:7391–7398. https://doi.org/10.3168/jds.2012-5482
Sumner, D. 2020. “California Dairy: Resilience in a Challenging Environment.” In P.L. Martin, R.E. Goodhue, and B.D. Wright, Eds. California Agriculture: Dimensions and Issues. Giannini Foundation of Agricultural Economics, University of California.
US Department of Agriculture National Agricultural Statistical Services (USDA-NASS). 2025. “Wisconsin Milk Cow Herds by Type of Milk Produced.” Available online https://www.nass.usda.gov/Statistics_by_State/Wisconsin/Publications/Dairy/2025/WI-DairyHerd-03-25.pdf
Wisconsin Department of Agriculture, Trade and Consumer Protection (DATCP). 2024. “Dairy Producer Survey Results, 2024.” Available online: https://datcp.wi.gov/Documents2/DairySurvey2024.pdf
