When evidence cannot change the plan
Collecting feedback helps only if it reaches the people who can revise assumptions, decisions, and priorities. Evidence can sit in reports and reviews without changing the work.
A traditional control model often reads feedback as variance from plan and responds by correcting execution. A complexity-informed model also examines whether the evidence calls the plan's assumptions into question.
An epistemic bet specifies what the organisation expects to know early, what it can defer, and what must stay open to evidence. Feedback debt accumulates when spending continues on assumptions the evidence no longer supports.
How learning breaks down
Double-loop learning
Single-loop learning adjusts action within the current frame. Double-loop learning revisits its governing variables. Teams accumulate feedback debt when they repeatedly correct execution while leaving the assumption that caused the correction untouched. Reviews need to allow that assumption to change.
Absorptive capacity
People need the vocabulary, attention, access, and decision rights to recognise and use feedback. Evidence can be collected and discussed without anyone being able to act on it. Connect the people observing a signal with those able to interpret it and authorise a change.
Learning myopia
Organisations tend to favour nearby, recent, successful, or comfortable evidence. Delayed, distant, failed, or politically inconvenient signals are easier to discount. Local improvements can then hide damage elsewhere. Review what is missing from the evidence as well as what is available.
Debt Taxonomy
These labels describe how feedback fails to change organisational decisions. Even well-intentioned teams accumulate debt when their learning process cannot carry evidence through to action.
| Debt Type | Typical Signature | Learning Failure | Intervention Bias |
|---|---|---|---|
| Muted signal | Weak evidence is smoothed into status language before it reaches decision forums. | Absorption fails at recognition. | Preserve raw signal and context long enough for interpretation. |
| Local correction loop | Teams fix symptoms repeatedly while the planning premise remains untouched. | Single-loop learning crowds out double-loop learning. | Add explicit assumption review to retrospectives and portfolio decisions. |
| Delayed consequence | The cost of a decision appears after the review window has closed. | Learning myopia favours recent feedback. | Track lagging effects and revisit decisions after use, not only after delivery. |
| Distant edge signal | Customer, operations, or frontline evidence exists far from the planning centre. | Absorption fails at access and translation. | Shorten the path between edge evidence and governing assumptions. |
| Success trap | Past success makes the current model harder to question. | Learning myopia favours confirming evidence. | Separate evidence of success from evidence that the model still fits. |
| Decision orphan | Feedback is acknowledged, but no forum owns the authority to change the bet. | Absorption fails at use. | Give feedback a decision path, not only a discussion path. |
Use It
Choose one recurring surprise and trace the evidence. Identify what arrived, where it stopped, whether it changed the governing assumption, and which signals were missing from the discussion. Repeated local corrections can indicate that the organisation is protecting an assumption from revision.
References
Argyris, C. (1976). Single-loop and double-loop models in research on decision making. Administrative Science Quarterly, 21(3), 363–375. doi.org/10.2307/2391848
Cohen, W. M., & Levinthal, D. A. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly, 35(1), 128–152. doi.org/10.2307/2393553
Levinthal, D. A., & March, J. G. (1993). The myopia of learning. Strategic Management Journal, 14(S2), 95–112. doi.org/10.1002/smj.4250141009