“The System Sucks”: Computer Programs and Technical Control in Entry-Level White-Collar Work

2.1 Regimes of White-Collar Technical Control

Two salient forms of managerial control Richard Edwards’ (1979) theorized are technical and bureaucratic. In technical control, control resides in the work process. An assembly line structures workers’ tasks and pace. Technical control is effective because workers are replaceable. Bureaucratic control, on the other hand, controls workers through the promise of a career. Bureaucratic control is the “rule of law,” where jobs are stratified in an organization and each position is governed by job descriptions and criteria for evaluation and promotion (Edwards 1979: 21). Control, in this form, resides in the social structure of work.

According to Edwards, technical control applies to blue-collar workers and bureaucratic control applies to white-collar workers. This distinction continues to drive research on managerial control. Though blue-collar factory work has declined in developed countries, technical control is still applied to jobs that are precarious or otherwise at the bottom of the labor market hierarchy (Cameron 2024; Vallas et al. 2022; Vallas & Schor 2020). The most prevalent contemporary example of technical control is in gig or platform work. Algorithms determine the steps, pace, and evaluation for these workers, are a source of stress, and create conflict between them and employers (Cameron & Rahman 2022; Kellogg et al. 2020; Schor 2020; Rahman 2021). Regardless of whether work is part of the new gig economy or structured by traditional employment conditions, technical control is rarely examined in an office context.

In rare instances, scholars have examined technical control in office work, however, these cases tend to be in clerical occupations or among established professionals (Callaghan & Thompson 2001; Lebovitz et al. 2022). For example, Callaghan and Thompson (2001) show how technical and bureaucratic control blend together in call-centers where workers are tethered to workstations. Entry-level workers reside between the two extremes (top of the occupational hierarchy and clerical work) allowing for research in more “typical” settings. Researchers also theorize computer monitoring as a form of white-collar technical control (Stanko & Beckman 2015). Yet, computer monitoring is distinct from software that dictates the steps and pace of work.

Research on managerial control over white-collar workers often examines bureaucratic procedures or cultural norms that lead to compliance (Chown 2021; Cardinal et al. 2017; Sitkin et al. 2020). Healthcare is a common site for studying control of professional workers. Instituting new monitoring meetings or checklists can ensure employee compliance (Chown 2021; Pronovost & Vohr 2010). Company culture can also effectively garner compliance among professionals because workers feel a sense of buy-in and commitment (Kunda 2006). Though both bureaucratic and cultural/normative control can apply to manufacturing (Burawoy 1979; McLoughlin et al. 2005), these forms of control are often the only kind theorized to apply to white-collar work. This is because researchers often assume workers in white-collar jobs/occupations are too autonomous for technical control to succeed. Even when considering automated diagnostic criteria in healthcare, high-status medical workers can exercise discretion in whether they follow AI’s recommendation (Lebovitz et al. 2022). However, it remains an empirical question how much autonomy entry-level workers have when experiencing technical control.

In this article, I show that many entry-level white-collar workers are controlled by computer software that dictates the steps and pace of work, either through logging their work extensively (social-oriented) or following computer prompts to complete their tasks (task-oriented). Responses to control vary depending on which form employees experience. Though bureaucratic control may require documentation, the lack of flexibility makes technical control qualitatively different (see Aneesh 2009 on how software that coordinates workflow is distinct from bureaucratic forms of coordination and see Wu 2020 on how app-based documentation can be more authoritative than paper-based documentation). As Subramanian and Suquet show, “constant reporting” through computer software is an example of technical control because it easily allows employers to control the work process and evaluate employees (Subramanian & Suquet 2018: 64). In sum, using software to document work tasks is fundamentally more intense than paper reporting because it (1) allows constant monitoring and (2) tells staff how to document in ways that are more inflexible than traditional paperwork (Subramanian & Suquet 2018; Wu 2020).

2.2 White-Collar Technical Control and Gender Inequality

Technical control in entry-level white-collar work has implications for understanding gender inequality at work. Recent studies examine the role of work tasks in creating and reproducing workplace gender inequality (Cech 2013; Martin-Caughey 2021). Examining gender difference can uncover within-occupation inequality, which is growing at comparable rates to between-occupation inequality (Xie et al. 2016). And though worker variation in responses to control is oft studied (Gill 2019), we know little about how people respond differently by type of control and demographic variables. Gendered differences in responses to technical control likely reflect the gendered nature of occupations. Both within technical occupations and in the broader economy, women may end up in lower prestige or less technical roles where social interactions may be subject to technical control (Cardador 2017; Cech 2013; Reskin and Roos 1990). In general, as jobs become feminized, they often lose autonomy and become more routinized (Reskin & Roos 1990). Feminized roles come with lower status and fewer rewards as feminized skills are culturally devalued (Quadlin 2018; Ridgeway 2011). Therefore, even if technical control degrades all occupants of an occupation, women may be disproportionately affected if they are funneled into more social roles and/or roles traditionally feminized seeing the most deleterious effects of this type of control.

Expatiating the social processes behind technical control in an office setting may help elucidate how gender inequality manifests in early careers among white-collar workers. Though men and women might be subject to technical control at similar rates within organizations, forcing employees to log all their social interactions, essentially replacing social roles with computer-mediated interactions, both disproportionately affects women (see Cech 2013) and is a more intense form of control than task-oriented technical control. Routinizing or simplifying work that is otherwise social creates a sense of alienation and dissatisfaction (Ikeler 2016; Leidner 1993). Uncovering how men and women respond differently to technical control both in similar roles and differentially based on their task assignments contributes to our understanding of within-occupation inequality and responses to task inequality.

Edwards foresaw the impact of “smart machines” and “program machinery” on managerial control. He argued, “rather than producing qualitative differences, this new technology is best understood as simply expanding the potential contained in the concept of technical control,” (Edwards 1979:117–119). I build off that insight to examine how recent college graduates working in entry-level work experience technical control. This takes one of two forms. First, task-oriented technical control occurs when employees perform all their duties in a computer program that dictates their tasks. This can come with resistance, but also with consent by re-framing their role as “data experts.” Second, social-oriented technical control occurs when employees must document every task in a computer program allowing the software to control the work process. Employees experiencing the latter, who are disproportionately women in the sample and likely in the broader economy, were those most frustrated by technical control. Though previous studies highlight how technology can coordinate work globally through standardization or the diffusion of software across organizations (Aneesh 2009; Pollock et al. 2007), white-collar office workers were previously thought as largely immune from technical control because limited studies show professionals effectively working around software (Lebovitz et al. 2022; Rahman 2021). This article demonstrates that work can be degraded by technical control even at the level of entry-level office work with high skill requirements that is often well paid. I also uncover a mechanism through which the introduction of technology turns gender disparities in work tasks into inequality in employment experiences.

3.1 Research Setting

This study is based on in-depth, semi-structured interviews with graduates of a U.S. Midwestern public research university. Studying recent graduates’ entry-level jobs has advantages. Graduates can be compared across industries and occupations at the same career stage. Second, entry-level job quality has long term social and economic consequences for individuals (Bills et al. 2017). Finally, the kind of tasks subject to technical control may be most evident in entry-level jobs. With firms cutting clerical and front-line workers (Gordon 1996), entry-level employees may have more reporting or clerical tasks that make technical control observable.

3.2 Research Design and Sample

I interviewed 85 individuals six months to one year after graduation. Most interviews took place over Skype, though some were conducted in-person on campus. The aim was to understand how recent graduates entered the labor market, what their jobs were like, and their ambitions. For the purposes of understanding the contours of technical control and workers’ reactions, only the 46 respondents who were employed and experienced white-collar technical control are included in this article. I left out the other respondents for three reasons. First, many did a range of jobs outside of an office setting (or were underemployed/in school). For example, some Engineering graduates work in manufacturing facilities. Second, many respondents conducted tasks not captured by technical control. For example, some English graduates created content for organizations. Third, to keep this study focused on technical control in a new context.

I recruited students from four majors: Industrial and Systems Engineering (ISE), Business with a specialization in Operations and/or Logistics, Communications, and English. These disciplines include an STEM field, a professional field, a social science, and a humanities discipline. As a result, graduates work in a range of occupations and industries, rather than one setting as is typical of studies on control (e.g. Chown 2021). Further, two majors have low-occupational specificity (English and Communications) and two are more specialized (ISE and Operations/Logistics). This allows for occupational clustering that is useful for determining the parameters of, and likely settings for, technical control. Finally, two of the disciplines included (ISE/Business) are men-dominated and two (English/Communications) are women-dominated (Shauman 2016).

Prior to data collection, the study was approved by the IRB at the author’s institution. Recruitment occurred by emailing classes, professors, and major listservs. Respondents were asked to refer others or send an email to the classes they were taking in their major. Two forms of possible bias from this sampling method are homophily and overrepresentation of certain kinds of volunteers (Heckathorn & Jeffri 2001). Homophily was dealt with by limiting the number of referrals for each respondent, but also by asking respondents to send an email to class lists rather than personal contacts. Overrepresentation was dealt with by providing a $20 gift card. The incentive mitigates overrepresentation and ensures some socio-economic diversity. This strategy succeeded, respondents came from a range of backgrounds, as is common at large public universities.

The choice of school and majors create scope conditions for this study. The school primarily serves middle-class students rather than those that are the most elite or most under-represented in higher education. Therefore, findings may not be transferable to graduates from significantly more or less prestigious universities. Midwestern public research universities tend to be disproportionately White (Armstrong & Hamilton 2013). I therefore cannot make racial comparisons. Finally, while I suspect similar kinds of technical control occurs across individuals’ careers, results are confined to recent graduates holding entry-level jobs.

While the subsample of 46 is large enough to identify social processes, there are a relatively small number of graduates from each major. Roughly half of the English and Engineering students in the broader study experienced technical control, but only a third of Communications students did. Meanwhile, approximately two-thirds of the Business majors are included in the subsample. The gender, class, and racial makeups of the subsample closely mirror the original sample. Almost 60 % of participants are women, this is largely due to the gender balances of the majors, although women engineering graduates are overrepresented in those who experience technical control. In general, the sample is disproportionately White and middle-class, which aligns with the school under study. Most employees experiencing technical control work in large bureaucracies. The modal industry is automotive, but many respondents work in manufacturing, technology, finance, or consulting. Despite most graduates working in large firms, median salaries differed by major, with Engineering and Business majors earning significantly more than Communications and English graduates. These numbers align with previous research on inequality between majors (Moss-Pech 2021, 2025). Full sample demographics appear in Table 1.

3.3 Interview Procedures

In-depth interviews were appropriate because I was interested in studying workers across several jobs/industries and interviews allow multiple accounts to be pieced together to create a broader picture of social processes (Weiss 1994). I follow the extended case method approach (Burawoy 1998) in my orientation toward the relationship between theory and data. The goal of the interviews was to extend the concept of technical control by understanding individuals’ experiences in one context (entry-level white-collar work). Given the research design of comparing people across multiple occupations and industries at the entry-level job stage, ethnography would not have been feasible. And it is unlikely a survey would uncover the necessary information needed to inductively uncover new contours of technical control. Ultimately, following others who use interviewing methodology to study control (e.g. Vallas et al. 2022), I expose, and theorize, how technical control operates in the setting under study.

The interview guide covered respondents’ transition into the labor market, their current job, how their job is related to their education/schooling, and a section on their ambitions/plans. Though respondents were asked the same interview guide questions to ensure consistency, I asked probes and follow-ups as appropriate. I asked detailed questions about how employees did their current jobs and probed them to uncover the step-by-step process they took to complete their tasks. Decades of labor sociology shows that only through this kind of sustained micro-evaluation of work can we fully understand how individuals do their jobs (e.g. Burawoy 1979; Vallas et al. 2022). For example, respondents often answered in vague euphemisms when I asked what they did at work. They might describe projects’ intended outcomes and how they contributed to the organization’s goals or profits. But upon probing I would frequently uncover they exclusively followed prompts in computer software that dictated their work process.

3.4 Coding and Analysis

For data analysis, I treated each respondent as a case. Research assistants and I wrote a one-page summary for each interview that I reviewed in conjunction with transcripts. Coding proceeded using NVivo qualitative analysis software. Following the extended case method principle of theoretical reconstruction, I coded interviews based on the interview guide and relevant literature to examine managerial control. This included codes for work steps, pace of work, and evaluation procedures. Ater review of coded transcript sections it became clear many respondents were experiencing technical control in a way not yet fully theorized.

I found many employees were required to log or enter all their work into an online system in a way consistent with technical control. Given this pattern, I tasked two research assistants with flagging instances of technical control in the sample following Edward’s definition of machinery dictating the steps, pace, and evaluation of employees’ performance. After initial coding, researcher assistants and I further coded each case as experiencing technical control if respondents were unable to exercise autonomy over their work vis-vis the computer programs. As such, each transcript was checked by three coders. We had a high degree of intercoder reliability, as only a handful of cases did not receive unanimous coding. We discussed anomalous cases as a team and came to consensus, arriving at the subsample of 46 respondents.

Next, I analyzed the cases inductively for emergent patterns among those experiencing technical control, and I found task and gender disparities. I coded for the two types of technical control (logging work and completing work in the software) and examined differences between and within each type. Though men and women experienced technical control at rates proportional to their representation in my study, women were more likely than men to respond negatively. Analysis revealed people were most upset with control when it replaced or routinized social interactions with colleagues. Since women were more likely than men to be assigned these kinds of tasks, even in technical occupations, it became clear this at least partially explained gender differences in experiences with, and reactions to, technical control in my sample.

4.1 Task-Oriented Technical Control

Thirty-two respondents completed their work tasks in a computer program following steps laid out by the software. For example, many organizations that hire Business and Engineering graduates use an enterprise resource planning system. These systems lay out exactly how workers must complete their tasks, leaving no discretion in completing assignments, and as such fit within the definition of technical control (Wu 2020). Tonya, a Business graduate working in the automotive industry, has a job requiring a college degree and pays $69,000 per year. As a Buyer, Tonya purchases parts from suppliers. A computer program, SAP, issues purchase orders and pays suppliers. She spends much of her day reviewing orders. She explained:

I review the field order, make sure it’s accurate and make sure that it gets signed. On the back end when the work gets complete, they’ll issue another thing that comes through the SAP system. I review those quotes, saying what we’ve agreed upon from the initial purchase orders matches what they’re quoting for this additional scope. And then I put those through a pencil-pushing approval process.

Tonya ensures purchase orders are accurate and signed by her boss. When discrepancies arise, she sends them to her manager. Each step is mediated through SAP. She said of this assignment, “I’d say probably about 25 % of my day is that, which is my least favorite part of the day.”

When asked to tell me about her least favorite part of her job, Tonya came back to SAP. She said, “the paper pushing aspect it,” was her least favorite. Tonya disliked, “reviewing documentation to make sure it matches, the system sucks.” When there is a mistake in the system Tonya referred to fixing it as “non-creative problem solving.” This is because everything needs to be done in an exact way, which limits Tonya’s discretion in completing her work tasks.

Though Tonya and Thomas (mentioned in the opening vignette) studied Business, engineering graduates had similar experiences. Stephen earns $62,000 as an entry-level Quality Engineer for a manufacturing firm. He ensures products meet customers’ standards. When Stephen initially described his job, it sounded complex. He said he was responsible for, “translating customer need into actual product functionality, and meeting all of the requirements.” When asked specifically how he does that, Stephen said they use a tool called quality functional deployment (QFD). QFD’s job is to “translate customer requirements into technical requirements.” In practice, the QFD controls Stephen’s work. Stephen told me, “It’s basically an Excel spreadsheet on steroids, where you fill in all the requirements and it basically tells you what to prioritize.” The QFD dictates the steps of Stephen’s tasks. It lists the information to provide, and Stephen fills it in. The program converts those requirements into technical specifications engineers in other departments need to initiate manufacturing processes.

Technical control makes evaluating workers simple. Samantha, a Business graduate, is graded on metrics produced by a system that is largely out of her control. She works for a fashion retailer earning $60,000 per year. Samantha forecasts how much product they send to stores. Like many others in this study with technical occupations, a computer system dictated the steps of her tasks. She said, “It’s a ridiculously complicated Oracle retail system that does a lot of the forecasting.” Samantha can manually override some of the forecasting, but most of the work is done by the Oracle system. Forecasts are compared to actual sales and Samantha is evaluated based on accuracy. She said, “It [the system] learns from its history so you need to keep it as accurate as possible. And it gets more accurate every week because you have more history.” She went on, “If they only sell 100 [out of 300] of those shirts, then I missed my forecast and then I get in trouble because it’s going to flag in my Monday selling meeting.” In many ways, the algorithm, rather than Samantha’s discretion, determines her evaluation.

The most common complaints reported by employees experiencing task-oriented technical control was their job was tedious, monotonous, and repetitive. Engineering graduate Artie works as a Core Product Engineer earning $67,000 at a manufacturing firm. In practice, senior engineers provide him with formulas and a list of injection molds, and he runs those formulas in Excel. He said of this work, it’s “monotonous because you’re doing the same crap every day.” Alan, a Communications graduate working at a bank uses software to run reports and called his job tedious. When asked if his job was engaging or challenging, he said, “No, not really. I don’t mind [though], I can just put on headphones and it’s fine.” He also said, “it’s a little tedious and repetitive honestly. Especially because I pull them [reports] every week.” Artie and Alan are two of the many respondents who experienced task-oriented technical control and complained their jobs were tedious, monotonous, or repetitive. And though responses were similar across fields of study, most jobs subject to task-oriented technical control were held by ISE and Business majors, due to the technical nature of the roles. And though respondents responded similarly by gender, due to gender segregation of majors roughly 80 % of the men in the subsample experienced task-oriented technical control, compared to approximately 60 % of women.

Reactions to task-oriented technical control were not universally negative. Some graduates discussed their work positively despite task-oriented technical control. They reframed their experience as “data experts.” Consider three examples – ISE major Mike and Business majors Luke and Burt. All three had tasks related to data management and reporting. Yet, they found pride in being data experts likely because of the utility of the data they provided to the firm or clients and the high cultural value of “data skills” in the economy (Moss-Pech 2025; Xie et al. 2015). Mike works for a digital services consulting firm and said of his job, “I have some opportunities to take raw data and figure out what to do with it.” Though Mike is the “lowest guy on the totem pole” and mostly conducts data entry into Excel, the resulting metrics are important to deliver to clients, which may explain his positive impression of working with data.

The other two respondents were more explicitly positive. Burt works for an automotive manufacturer and compiles data on spending forecasts. He claimed to have more knowledge of the data than anyone else, describing that as a “feather in my cap.” That Burt presents this information to his division Vice President is a source of pride that may partially explain his positive view toward his work even if he primarily looks up parts prices and project codes.

Finally, Luke, like Mike, works in consulting and similarly said he has “become the go-to guy for all the data in the system.” Though in practice, he ensures data has been accurately entered into a computer system, that this project is a deliverable for a major airline appeals to him. All three respondents mentioned above, despite having little discretion over their work, framed themselves as data experts. Technical and data skills are masculine-associated skills and highly valued in the economy (Quadlin 2018; Xie et al. 2015). That these men experienced task-oriented technical control and re-framed it through gendered cultural notion of data expertise, sheds light on how technical control may be re-framed when it aligns with tasks that are culturally and economically valued. All three said the data they worked with were part of big projects or reported to supervisors high-up in the organization. This stands in contrast to the narratives of those who experienced social-oriented technical control, logging all their work into a computer system. Logging work often replaced social interactions and therefore took place in social occupations, which are culturally associated with women (Quadlin 2018).

4.2 Social-Oriented Technical Control

For 14 graduates experiencing social-oriented technical control, they completed work outside a computer system and then recorded it in software. Since these were detailed descriptions of work individuals completed that needed to be entered a certain way at scheduled times, they controlled the work process (Subramanian & Suquet 2018: 64). English graduate Colleen works in the publishing as a Project Manager earning $38,000 a year. Colleen uses a time management program to shepherd book projects through production. The system provides “a couple of templates” for her to use. Colleen constantly puts status updates in the system. She said, “From when we get the manuscript all the way to when it gets sent out. So, you’re constantly having to go in and sort of update where it’s at and as things get finished, mark them complete. So that way it tracks, accurately, how long we’re spending on projects.” For example, if a manuscript goes through multiple rounds of editing, after each round Colleen goes into the computer program and enters that update. The software therefore dictates Colleen’s behavior because her job is to receive these updates and make sure they are logged correctly.

Brett also studied English. He worked as a logistics coordinator for a trucking company. Brett uses a computer program to go down a queue of deliveries. He explained the process:

I would come in and there’d be a program where it lists all of them and then you just go down the list. It would have either the company’s information or the driver’s information and you just call them up. “Hey, I’m Brett with [company name] just trying to see if your driver’s going to be able to make it on time for his delivery,” and if they yes, you type that into the computer. If they say no then you get the issue, “Okay, so your driver got a flat tire,” then you let the company know.

Brett told me “We’re basically just the middleman there.” He had no troubleshooting or problem-solving responsibilities. He said, “I was only responsible for taking the information and relaying the information.” He said this could get frustrating because sometimes drivers wanted him to help solve whatever problem they were having. Brett’s example highlights an important phenomenon. The trucking industry has long been on the forefront of technical control by surveilling drivers (Hubbard 2000; Levy 2015). Now technical control infiltrates other jobs in the industry. In this case, logistics coordinators like Brett, who go down a list of drivers in a computer program, checks on these drivers, then enters the relevant information. This pattern was apparent in other industries as well, such as manufacturing, where office workers like Stephen, in addition to assembly line operators, are subject to technical control.

Jade, a Communications graduate, works in Healthcare as a Pharmacy Business Consultant. She goes to pharmacies in her assigned region and discusses programs or discounts her employer offers. Every interaction must be documented in an online system. She explained:

That’s a really big push now is for people to put in there, they call them touches and they’re basically just notes of your interactions on a daily basis. You’ll have different categories. You can say it was an on-site visit, a phone call, an email. You just go in and you say what you talked about, the gist of your conversation, you can put a follow-up date, you can tag other reps in it, so like after I met with [client] today, you come home or whenever, you can do it on your phone too, and just said, “Met with [client] to introduce myself as his new rep. He said that he’s having issues with understanding the new reconciliation portal. I reached out to [colleague]. He said he’s up for a demo. We’ll follow up, blah blah.” They’re supposed to be as detailed as possible.

Jade estimated she spent 25–30 % of her day logging these “touches.” This system allows her employer to closely monitor Jade. That Jade needed to input all her work interactions into a computer program allowed technical control to take place. Jade’s experience was representative of several others who likewise had to log or upload all their client interactions into a database while supplying detailed notes. Though many individuals need to document their work, the lack of discretion in executing their tasks and intensity of surveillance make this form of logging or documenting work different from professional workers who need to provide accountability reports (e.g. a lawyer calculating billable hours). The level of detail and continual logging shaped the way Jade did her job, the language she used to interact with customers, and the pace at which she was able to work. Not only were there specific ways she was supposed to document each meeting following prompts in the software, but the online nature of the task allows her employer to continuously and quickly monitor and evaluate her.

Despite conducting work offline, programs often dictated the pace of workers’ tasks. This was the case for Emily and Roger, a Communications and Business graduate respectively, who work in sales for the same technology company. Both employees needed to make 50 phone calls per day to prospective clients. Roger said, “it’s a hard job man [being] responsible for X amount of calls a week, X amount of meetings.” Making 50 calls is Emily and Roger’s pace of work. But like many individuals experiencing technical control, Emily resisted by making “fake dials.” She explained, “Sometimes I make fake dials just because they monitor that stuff…so realistically maybe I make 20–25 dials a day and then I’ll email the rest.” When I asked her to explain fake dials, she said calls get counted if they last longer than 10 seconds. She said, “So, you can dial any number, let it sit there until it reaches 10 and then you hang up and dial again and then you hang up and dial again until 10.” She said this worked best when customers have automated messages that she can let talk for 10 seconds before hanging up. Emily and Roger register their calls in the online system and include whether it led to a follow-up sales meeting. However, the next step was outside of the scope of their work; Roger explained that after the initial call, the work is passed off to another employee. In addition to having their calls monitored, they are evaluated based on the percentage of calls that turn into sales leads. Emily told me she needs to get 10 % interested. Unfortunately, Emily is not meeting her numbers. She said, “I’m not doing well right now and it’s not because I’m not trying… I know at the end of the day it’s about the number, so I feel like I need to find somewhere else [to work] quickly before they kick me off.”

Eleven of the fourteen workers who experienced social-oriented technical control expressed dissatisfaction, especially since it routinized their social interactions. Sophie, a Communications graduate, works for a marketing firm. She schedules client’s social media posts using an online platform and interacts with clients. Similar to Jade’s “touches,” Sophie uploads and describes all client interactions in a database. She said of this work, “it’s a lot when you’re trying to do everything else, making sure everything is uploaded properly and we got a new system we’re trying to learn and that [makes it] take longer.” Jade also complained about the “touches” she had to put in the system saying, “it gets to be a lot,” if she’s out all day visiting clients, she must do them when she returns home. Brett, the English major working at a trucking company, felt constrained by his ability to help drivers. He said:

There were times where people would be like, “I need you to do this,” and it’s like, “Sir, I can’t book this appointment.” They’d be late and there’d be places like Procter and Gamble for example, if you don’t show up at your appointment time they won’t let you in until you make a new appointment. We get people that call us, they’re four hours late, they know they’re going to be late…they want us to change the appointment. I couldn’t change the appointment…You get a lot of guys yelling and screaming but that’s just the way the system is set up.

Recall Brett said he was just able to relay information. As a result, he was constrained in his ability to help truck drivers who might yell at him.

Though respondents from all majors and job types reported frustration with the constraints of technical control, the rate was higher for those with social-oriented technical control. This is because respondents expressed the most negative feelings when technical control affected their social interactions at work. Research shows workers often respond negatively to the monitoring or routinization of social interactions (Ikeler 2016; Leidner 1993). This was evident in Communications majors Jade and Sophie complaining about logging interactions. This phenomenon was also clear in the example of Brett who reported his interactions with truck drivers could become combative because his autonomy was constrained. Further, individuals experiencing social-oriented technical control were disproportionately women. This is because the types of roles routinized in this way were often held by Communications and English majors, two women-dominated college majors (Shauman 2016). Though the reverse logic is also possible, these roles are more likely to be dominated by women and therefore more Communications and English alumni enter them. In either event, the result is unequal gender sorting into jobs with different forms of technical control.

The case of Engineering major Jackie further shows how replacing peer interaction with online communications can frustrate workers. Jackie works as a Cost Engineer for an automotive firm. Design Engineers send her parts to include in cars and she sends the cost of those parts to Buyers who purchase them. She must log or document all these interactions. She explained:

Realizing that everything needs to be documented is another thing that I struggle with a lot because I’d rather just go and talk to the person and get information like that. All I’m doing is just managing this information that they’re telling me. And little things can mess it up, like if I wasn’t that organized about it or didn’t document something right, or I mistyped something or the system is just slow because the system is very slow. That doesn’t make me feel good.

Here Jackie complains about aspects of her job that are difficult because of computer software. These kinds of complaints were common in the sample and though men and women responded similarly to the type of technical control they were subject to (see for example, Brett), frustration was disproportionately expressed by women given the gendered division of work tasks within and between occupations. In other words, the two types of technical control map onto qualitatively different work experiences, with social-oriented disproportionately experienced in roles more likely to be held by women due to larger labor market conditions such as workplace gender segregation, the cultural typing of social skills as feminine, and the routinization of work in occupations that feminize (England et al. 2020; Puzio & Valshtein 2022; Quadlin 2018; Reskin & Roos 1990).

5.1 Contributions to Technical Control and Gender Inequality at Work

Technical control permeates white-collar work that is relatively highly paid, requires a college degree, and is often described as skilled (Kalleberg 2012). Most respondents work in large bureaucracies. Those not working in large bureaucracies often had technical tasks or project management-type responsibilities. This gives us insight into the contexts we may be most likely to observe technical control. However, it can also apply to other settings. If people conduct their work in computer software, they can be subject to technical control. Technical control may become more salient as a larger percentage of employees work remotely and as AI gets further integrated into the workplace. As a result, a large segment of the workforce previously thought to be autonomous may be subject to technical control.

I developed two sub-types of technical control, task-oriented and social-oriented, which varied among respondents based on occupational contexts. Firms hiring graduates of technical disciplines for data-centric roles often deployed task-oriented technical control. Graduates from less technical disciplines (Communications and English) often had social roles more subject to social-oriented technical control. This is largely due to college field of study and subsequent occupational gender segregation (England et al. 2020; Puzio & Valshtein 2022; Shauman 2016), though I suspect within technical disciplines, women were more likely funneled into social roles subjecting them to social-oriented technical control (Cardador 2017; Cech 2013). Social-oriented technical control was perceived more negatively, but part of the variation may be explained by the cultural and economic value of “data skills” and how important people thought their tasks were to their organization. The three men who framed their role as “data experts” re-framed technical control through the cultural lens of data expertise and reporting to important teams/clients. Though I cannot speak to contexts beyond my study, it is likely variation in responses to technical control depend upon how central the role seems to the organization and the nature of tasks being routinized. However, this must be interpreted with caution as studies show men are more likely than women to inflate their performance and importance at work (Kroska 2009); and I cannot rule out this explanation for why men portrayed themselves as data experts.

This paper extends our understanding of managerial control. Most research on control of white-collar work shows how bureaucratic rules and procedures control workers or how cultural norms create compliance (Chown 2021; Edwards 1979; Kunda 2006). Meanwhile technical control is often limited to clerical, gig, or blue-collar work (Callaghan and Thompson 2001; Cameron and Rahman 2022; Hubbard 2000; Kellogg et al. 2020; Levy 2015). Some researchers examine control of freelance or platform work (e.g. Rahman 2021), or those working in healthcare (Lebovitz et al. 2022), but these skilled professional workers operate at a different career stage and in a different employment context than those I studied. Research shows health care workers tend to maintain some autonomy over their work even with the introduction of new technologies, while blue collar work continues to become de-skilled through automation. In this paper, I show that technical control has extended into entry-level white-collar work previously thought to be immune from technical control. Results suggest these jobs may become more like fully automated blue-collar work than the more autonomous healthcare work, despite employees maintaining their status in the organization and the jobs requiring college degrees. This forces us to rethink a core tenet of how certain workers can be controlled and key features of technical control. Specifically, technical control ensures compliance due to the interchangeability of workers. Perhaps with the proliferation of college degrees and degradation of professional work, entry-level college educated employees have become more replaceable and therefore more precarious. As credential inflation continues to rise and more jobs require a college degree, it may no longer reflect the broader economic reality to describe these jobs as “high skilled” (Aneesh 2001; Moss-Pech et al. 2021). Though many respondents in this study were well paid, we may be underreporting how many college graduates are underemployed if we simply look at whether jobs require a college degree and exclude jobs that require a degree but do not ask graduates to draw on skills learned in college and/or subject them to technical control. Future research can further unpack this by investigating later career stages.

This paper also extends our understanding of how algorithms control workers. While algorithms often control workers through ratings (Cameron & Rahman 2022; Kellogg et al. 2020), they can also control office workers by creating both the metrics and output for evaluations. Recall Samantha was evaluated on the accuracy of sales forecast determined by an algorithm. I also move beyond algocratic modes of organization (Aneesh 2009) by showing software not only allows for more global coordination, but also the managerial control of workers with high skill requirements and pay.

The expansion of technical control into the white-collar workforce is qualitatively different, and more intense, than other forms of control. Essentially, technical control deskills work. Take Edwards’ example of a factory foreman, whereas prior to the assembly line, the foreman had supervisory power and technical expertise, the introduction of the assembly line transformed the foreman into an “enforcer of the requirements,” (Edwards 1979: 120). Similar transformations could occur in white-collar work, where previously technical employees are reduced to data entry and software monitoring. Since technical control resides in the work itself rather than the rule of law, it pervades deeper into employees’ day-to-day experience. This is even the case in jobs that require logging or monitoring social interaction. As Subramanian and Suquet (2018) show, constant reporting and immediate monitoring and feedback elevate logging work into technical control when it moves from a paper process into a computer system. Given the relationship between technical control and de-skilling, and the relationship between technical control and the ostensible status of those subject to it (usually workers at the bottom of the occupational hierarchy – Cameron 2024; Vallas et al. 2022) it may be experienced as more intense for white-collar workers than forms of bureaucratic control.

This article makes a novel contribution to gender inequality at work. Though scholars research workers’ reactions to managerial control (Gill 2019), studies rarely separate reactions by demographics. I show how managerial control can be gendered. Despite men and women reacting similarly to each type of technical control, in general task differences between men and women is an important way within-occupation inequalities manifest (Cech 2013; Martin-Caughey 2021). People responded most negatively to technical control when computer software replaced or routinized social interactions and women are more likely to be given these tasks at work. Structural processes such as technical skills being culturally associated with men and social skill being culturally associated with women (Quadlin 2018), gender segregation in the labor market (England et al. 2020), and the devaluation of work that is feminized (England et al. 2020; Quadlin 2018; Reskin & Roos 1990), combine to explain why men and women might be subject to different types of technical control, with social-oriented technical control experienced by women more likely to elicit negative responses. I therefore contribute to literature on gender inequality by showing how a single form of control, in this case technical control, can intersect with task discrepancies to create employment differences by gender. Though technical control did not create the task disparities by gender, the introduction of this kind of control, was more likely to make employment conditions worse for women than for men.

Technical control permeating entry-level white collar office work can also help explain the relationship between deskilling and the standardization of work on the one hand, and the status of white-collar workers on the other. The general degradation of work in the United States is well documented. For many workers, their jobs have become deskilled and standardized (Aneesh 2001; Braverman 1974; Ikeler 2016; Timmermans & Berg 2003). When this happens, workers usually lose pay or status (e.g. contract lawyers – see Brooks 2011). The workers who maintain their status in the face of standardization, often retain their autonomy over the work process even when attempts are made to automate or routinize their work (Lebovitz et al. 2022; Timmermans & Berg 2003). However, respondents in my study experiencing technical control neither had much autonomy nor lost any pay/status. What explains this outcome? Though full accounting is outside the scope of this paper, it is possible white-collar technical control captures work in flux; this work may be further automated or outsourced to lower status workers in the future, resulting in loss of pay or status. It is alternatively possible that managers who oversee a team of workers may not want their team dissolved even though technical control makes that possible. Another possibility is that these workers will advance into more complex work in the organization, making technical control a prominent feature of entry-level work. All possibilities reflect an economy where employers require ever-higher degrees for jobs that may not substantively require much skill-use (Moss-Pech 2025; Moss-Pech et al. 2021). It is possible work will both continually be degraded and the credentials needed for these jobs will continue to rise.

5.2 Limitations

I have used in-depth interview methodology to study control at work. This has advantages and limitations. Following others using this method (e.g. Vallas et al. 2022), I show how technical control in one context operates and workers’ reactions to it. A limitation of ethnography is it can be difficult to observe multiple workplaces. Most studies of control, surveillance, or quantification, focus on a single industry, like retail work or journalism (Christin 2018; Ikeler 2016). I examine workers across sectors gaining breadth that would be difficult to ascertain in a single case study or a comparative ethnography.

However, a limitation of my method is I cannot clarify the line between the affordances of the software itself and the organizational practices surrounding the software. Managerial oversight, formal rules, informal norms, and the software packages themselves all structure the level of intensity with which technology controls workers and to what extent employees find ways to exercise discretion or workarounds. It remains an open debate in STS scholarship whether the diffusion of software across organizations allows for each organization to adapt the software to its local context or if the software has a “generification” effect across firms (Pollock et al. 2007; Shestakofsky 2017). Though fully understanding the organizational practices and contexts in which technical control in white-collar work takes place is an essential endeavor for future research, I have taken the important step of uncovering it in entry-level work and documenting employees’ reactions.

5.3 Conclusions

Understanding managerial control and workers’ responses to it has been a touchstone of sociology for almost a century. As technology changes and algorithms run many social institutions, scholars are paying greater attention to technology, quantification, and inequality outside of work. However, workplaces continue to be on the forefront of how technology shapes our lives. Uncovering technical control in entry-level white-collar work can provide a jumping off point for inquiries into how computer programs constrain people in complementary ways, whether it be academic staff burdened with onerous and exacting reimbursement systems or electronic medical records specifying diagnostic criteria. By expatiating technical control in one context, I have given us new insights into what the future of work might look like, and how algorithms and technology might degrade work even further up the occupational hierarchy than previously observed.