(Originally published by C4SS)
During a recent business meeting, I noticed some ways in which corporate hierarchy leads to negative sum competition and distortions of reality. At my current job in data analytics, I build live business “dashboards” to provide actionable insights to upper management; this includes working with different departments to create new data pipelines and ensure the databases are up-to-date. As a recent entrant to the workforce, I did not expect that such a large proportion of my time would be spent navigating bitter office politics, something that long-term employees seem to take as a given. To use Graeber’s lexicon, I’ve become something of a duct taper.
Without getting into too many details, the scenario was as follows: the company was considering marketing a new product to existing customers. In our case, an important metric in making the decision was to look at the difference in profitability between the existing close substitute on offer and the new product– cannibalization was expected. To support the new product, a new manager with some expertise was to be hired and the employees who specialize in selling and developing the old product would be retrained and shifted over to work on the new one. The company had no plans to immediately stop selling and supporting the old product, but to eventually phase it out over time.
To these ends, an outside consulting firm was brought in to assess the profitability of the new product. My job was to calculate the profitability of the existing product, a relatively trivial task. Management also wanted a visualization of the sales performance of the existing team to assist in their decision-making.
There were multiple interests to consider in the context of this scenario. Firstly, the sales manager wanted to keep his job in the long run, maintain his commission numbers in the short to medium run, and look good in the eyes of management. Secondly, upper management wanted to maximize profits in order to line their own pockets, a goal our analysis found to conflict with the interests of the sales manager. Finally, my boss wanted to maintain his reputation for providing accurate data to upper management, but also ensure the cooperation of the sales team, who he knew could make our lives difficult by refusing to cooperate with us in the future.
After I finished building the visualization, I sent it out to “low level” stakeholders to ensure that the data was correct. Almost immediately, I received a call from the panicked sales head telling me that my numbers were wrong. After some tedious back and forth with the folks who handle data for the sales side and escalation to higher levels of corporate governance, it quickly became apparent that different ways of defining productivity showed the sales team in a different light. My numbers weren’t wrong, they just didn’t serve the sales manager’s interests. Depending on which exceptions one made on various cost and revenue categories, there were dozens of ways to define profits.
Naturally, the sales manager favored definitions that highlighted his team’s performance because this would deter management from introducing the new product, which would cut into his commission and potentially cost him his job. However, to make an accurate comparison and maximize profits, we’d need to use the same methodology to assess the profitability of both products, otherwise, we’d be comparing apples to oranges. Even headcount came under dispute, with the sales team arguing that certain non-performing people ought to be excluded since they’d be fired soon anyway.
After much heated back and forth, and to my surprise, my boss and the head of the sales team decided on a “compromise” between two competing versions of reality, one backed by statistics and the other by the need for self-preservation. Insightful information, but not insightful enough to cause any trouble. In other words, they agreed to fudge the numbers (just a little). Of course, things weren’t phrased in those terms, but some numbers were generously rounded, they decided to adopt a slightly different definition of total revenue to the one used by the consultants, and they excluded a few people from the final headcount. This brought up the team’s productivity just enough to ensure that they wouldn’t be completely phased out, but not so much as to invite suspicion from the untrained eyes of upper management. The final number was marginally less than the figure calculated by the consultants. The primary beneficiary of this arrangement was the sales manager, but not necessarily his team, all of whom were going to be transitioned into selling the new product. After the week-long bureaucratic nightmare, my boss gave me a single piece of career advice, “it’s always important to keep the people you work with happy.”
This is both banal and sagacious advice in the context of capitalism. As far as employees are concerned, maintaining good relations with their boss and colleagues ensures job stability and opens up the possibility for promotions and a higher income in the future. However, this mindset can lead to externalities that harm the organization as a whole.
In our case, misaligned and perverse incentives turned out to be highly costly in terms of both time and money. The corporation spent over $200,000 on consultants, all to end up with an uncertain trade-off. We spent over a week butting heads over definitions and reworking the numbers. Given the final difference in productivity, it’s unlikely that upper management will decide to invest more in developing the new product and make the necessary personnel changes. The cost of making the wrong decision or slowing down the transition will add up, costing millions. The irony here is that the data analytics team is supposed to notify management of excessive waste. Instead, we were recruited into helping to create it.
Beneath the neatly presented, audited, numerical reality found in annual reports and marketing materials there are competing illusions vying for supremacy. Imperfect information is given in practically all markets, but this condition is exaggerated under hierarchical capitalist institutions where managers compete with workers, opaque concentrations of private power are protected by the state, and politicians have a vested interest in misleading constituents. Seemingly simple statistical data is far more complex than it would seem on the surface. For example, in a previous position as an outside consultant to a state-funded infrastructure project (I burned out badly), my main project involved collating and manipulating extremely messy and incomplete location data to make it appear as if most of the spending was allocated to rural areas and small businesses. The end result, an interactive map demonstrating the economic impact of the project, was highly distorted relative to what data actually showed— most of the spending was going to big, politically connected consulting firms in major cities.
The less accountability and more top-down control capitalists have over the population, the more distorted the official reality becomes, something we’re witnessing in today’s post-truth political discourse. The narratives spun by data at the corporate level are not necessarily intended to inform rational business decision making but to justify predetermined imperatives that primarily serve to cement established power dynamics. The common thread I have found in my experience of corporate life is that we do not fit conclusions to data but data to conclusions. What we’re dealing with isn’t just noisy data and oversimplified projection models, but a starkly postmodern landscape where power distorts empirical reality. Of course, this is a common theme in anarchist literature, but it’s jarring find yourself in the reality lab, manipulating metrics people implicitly trust and make decisions with. Others have had similar experiences:
What I could not get my head around was having to force-fit analysis to a conclusion. In one case, the question I was tasked with solving had a clear and unambiguous answer… But the client did not want analysis that contradicted their own, and my manager told me plainly that it was not our place to question what the client wanted.
Naturally, informational distortions have consequences, namely waste. Estimates say that excess management costs the U.S. economy $3 trillion per year. Large corporations can easily absorb these heavy costs, but there is a massive opportunity cost to consider. In an anarchic economy, dominated by a robust commons and lean worker coops, the cost savings of more efficient production would benefit workers and lead to increasingly positive-sum interactions. Better decision making across all production units would be incredibly beneficial on a global scale.
At this point, one might argue that a possible solution is a stronger central body to render chains of signification transparent — in other words — to create top-down system of conceptual clarity to prevent the power struggles and ensure information is accurate, with managers monitoring employees, executives monitoring managers, the board monitoring executives, and the state monitoring the board. However, this proposal runs into information problems and it’s exactly what we have now but with more stringent oversight. In a huge corporation with thousands of employees, it is impossible for management to keep track of everything. This is precisely why the sales manager and my boss felt comfortable about adopting a generous methodology to calculate productivity — management would never find out. More-stringent, top-down imperatives wouldn’t erase underlying conflicts of interests, but would only push them into the shadows while also further alienating workers who already lack control over their day-to-day lives. Even at my level, my boss doesn’t fully know or understand everything I do. A common ploy I know a lot of my friends’ use is to pretend to have unfinished work when they don’t, allowing them to spend time reading articles or listening to podcasts before they’re given another task. More-stringent control to prevent such behavior would be costly to implement and lead to increased stress and burnout.
In a hierarchical organization, there are misalignments of incentives at every level as a result of the hierarchy. These misalignments produce the need for people to bridge gaps in communication and self-interest in order to facilitate action. Graeber notes these inefficiencies at length in Bullshit Jobs, where flunkies, goons, box tickers, duct tapers, and taskmasters all respectively exist to validate, protect, justify, navigate, and reproduce hierarchies. Bullshit jobs are entirely unnecessary. While my position doesn’t necessarily fall into the bullshit category, there are bullshit components to it that only exist because of the corporate hierarchy.
While some disconnect in incentives and information is inevitable, the same levels of waste and inefficiency would not occur in a worker-run coop where all employees have a stake in the final accounting. For example, when introducing a new product, salespeople wouldn’t have a boss facing replacement in the first place, they’d have a say in transition plans and could invest in their own retraining if necessary. On the other hand, in a corporation, workers interests don’t align with and are necessarily subordinate to those of upper management, creating perverse incentives that encourage practices such as fraud and deceit. Self-management and organization around consensus and delegation of roles turn out to be more efficient, less stressful, and less alienating.
Thanks to this and other effects, networks of smaller, leaner producers are more efficient than large bureaucratic corporations. This is in part due to the increased transparency, heavier competition, and lower transportation and warehousing costs experienced by smaller firms.
One argument that comes up as a defense for large scale production is economies of scale, which occur when fixed costs are distributed across more units of output, so marginal cost decreases as production increases. However, this is a topic already addressed by Kevin Carson who points to the knowledge problems I’ve discussed here, expensive machinery which needs to be amortized by mass distribution, warehousing and transportation costs, and promotion costs to deal with the sheer quantity of goods that would need to be sold. Moreover, if fixed capital becomes part of a shared commons, it would be possible to harness both the cost savings from distributed economies of scale and competition between producers. As Carson has argued at length, decentralized production would quickly outcompete centralized, bureaucratic organizations, if not for the fact that our current, heavily subsidized, capitalist markets dedicate much of their capacity to reproducing concentrations of power.
Conflicting incentives created by hierarchy lead to informational loss and waste that could be harnessed to significantly increase efficiency and standards of living as a consequence. Today, “data-driven insights” are more often just a marketing gimmick rather than a real source of value to the economy. Actually being able to make full use of the data without these distorting forces would lead to efficiency gains on a far wider scale than we’ve already seen.