Reorganizations tend to assume that authority is what the chart says it is. Redraw the boxes, reassign the reporting lines, and control moves accordingly. Anyone who has watched a decision get made knows this is half the story. The person with the title and the person who actually shapes the outcome are frequently not the same, and pretending otherwise produces reorganizations that change the diagram and nothing else.
Two authorities, not one
The cleanest way to name the gap comes from strategy+business (2022), summarizing the argument that managers matter. It distinguishes two kinds of managerial authority. The first is the power to hire, fire, and instruct, the formal authority that fits neatly into a box and a reporting line. The second is the power to lead, to structure a problem, and to build consensus around a direction. Only the first is easy to draw. The second is where organizations are actually steered, and it does not respect the chart.
This distinction dissolves a lot of confused debate. When a company declares itself bossless, it is usually retiring the first kind of authority while quietly relying on the second. Foss and Klein (2022) find precisely this pattern across firms that claim to be flat: they retain real bosses and hierarchies, just informal or renamed. Formal authority is removed from the chart and real authority reappears off it. The power did not leave. It changed address.
Real authority is earned and mobile
The healthiest organizations do not try to force the two authorities to coincide. They make the second one legitimate and visible. W. L. Gore is the clearest case: Grønning (2016) documents a lattice where leadership attaches to contribution rather than title, and even the CEO is chosen through peer recognition rather than appointed. Real authority there is conferred by the people who will follow it, which is what makes it real. It can also be withdrawn, which is what keeps it honest.
Semco reached the same place through transparency. Semler (1989) describes workers voting on major decisions and reading the monthly books, so influence rests on shared information rather than on position. A manager cannot lean on privileged access to facts everyone can see. That is a deliberate way of aligning real authority with something other than the title, and it is why the firm could eliminate its org chart, as the teaching case records, without descending into chaos (Siehl, Killian and Perez, 1999).
What these firms grasped is that the second kind of authority behaves differently from the first. Formal authority is a possession: you are granted it, you keep it until it is revoked, and its holder does not need the ongoing consent of the led. Real authority is a relationship, renewed or withdrawn continuously by the people who choose to follow. That is why it cannot be assigned by a memo and why a promotion does not confer it. A newly appointed manager holds the full first kind of authority on day one and none of the second, which is the awkward gap every new leader has to close by earning it. The chart records the grant. It cannot record the relationship, because the relationship is decided by everyone else.
Manage the gap instead of denying it
The failure mode is treating the gap as an error to be engineered away. It cannot be. Lee and Edmondson (2017) note that even radically decentralized organizations require structure and discipline to function, which means informal leadership will always exist and the only question is whether it is acknowledged. A firm that denies the gap gets the worst version of it: authority that operates without a name, accrues to the well-connected, and cannot be contested because officially it is not there.
The practical work is to make real authority legible. Ask, for any important recurring decision, who actually shapes it, and whether that influence rests on something defensible, expertise, contribution, the consent of the people affected, or merely on tenure and proximity. Where the answer is defensible, formalize it lightly so others can see and appeal to it. Where it is not, that is a problem the chart will never reveal, because the chart only ever showed the first kind of authority.
There is a corollary for the fashion of deleting managers to cut cost. If the design authority, the encouragement, and the enforcement described in strategy+business (2022) are real functions rather than decorations, removing the role does not remove the work, and the saving is partly an illusion. The functions reappear as unpaid, unaccountable labor performed by whoever cannot stand to let them lapse. A firm can genuinely shed the first kind of authority, the raw power to command, and often should. It cannot shed the second kind by decree, because the second kind was never the manager's to give away in the first place.
The lesson for anyone about to reorganize is modest and freeing. You are not going to make the org chart tell the truth about power, because half of power was never chart-shaped. The achievable goal is to make the uncharted half earned, visible, and reversible. Foss and Klein (2022) are right that the boss rarely disappears. The choice is whether the real boss is someone the organization chose on purpose or someone it acquired by accident.
Sources
- strategy+business (2022). Can bossless management work? Distinguishes the authority to hire, fire, and instruct from the authority to lead, structure, and build consensus, the framework of the piece. https://www.strategy-business.com/article/Can-bossless-management-work
- Foss, N. & Klein, P. (2022). Who's the boss? The persistence of entrepreneurial hierarchy in flat organizations. Finds flat firms retain real bosses, informal or renamed, showing formal authority leaving the chart while real authority persists. https://link.springer.com/article/10.1007/s41469-022-00132-8
- Grønning, T. (2016). Working Without a Boss: Lattice Organization at W. L. Gore & Associates. Documents leadership tied to contribution and a peer-recognized CEO, an example of real authority made legitimate and mobile. https://sk.sagepub.com/cases/download/person-to-person-communication-at-wl-gore-associates-inc
- Semler, R. (1989). Managing Without Managers. Shows influence grounded in votes and open books rather than title, a way of aligning real authority with shared information. https://hbr.org/1989/09/managing-without-managers
- Siehl, C., Killian, D. & Perez, F. (1999). Ricardo Semler and Semco S.A. Records Semco eliminating the org chart without collapse, supporting the idea that the chart never held all the power. http://www.saludygestion.com/archives/Ricardo%20Semler%20&%20Semco,%20Thunderbird.pdf
- Lee, M. Y. & Edmondson, A. C. (2017). Self-managing organizations. Notes decentralized organizations still require structure and discipline, meaning informal leadership always exists. https://www.sciencedirect.com/science/article/abs/pii/S0191308517300059