If nobody can trace the decision, accountability gets much easier to evade.
A lot of bad decisions do not “disappear.”
Someone loses time and misses an important deadline. Someone else gets blocked/locked out of a system because they’ve forgotten a password. An executive gets exposed for something unethical. A frontline employee gets told to work around a problem that (unbeknownst to them) could have been prevented in the first place. Other times, they are told nothing and are expected to just fix it. Of course, there are opposing views on this (here’s one titled “Relining the garbage can of organizational decision-making: modeling the arrival of problems and solutions as queues”).
In many of these scenarios, the problem just gets “absorbed.” Usually, that means someone else is left to carry the consequences. The structure stays in place, but the reasoning behind the problem is somehow harder to find than it should be. Note: Research studies dispute similar claims, like this one that focuses on cognitive microfoundations and organizational decision-making (public management focus).
In my work, that is part of what I keep coming back to.
If a decision shapes risk, and governance helps determine who gets to make decisions, then perhaps better tracking of decisions matters, too. That is also why accountability needs a clearer way to connect harm, responsibility, and remedy. Even more so for us operations people when we realize that once a decision disappears into a process, platform, queue, or a chain of approvals, accountability gets much easier to deflect and fragment.
This post is about what happens when decision-making is poorly tracked, weakly documented, or treated as “too routine” to question.
What gets tracked?
What to track is not simply an administrative question. It is a question used by marketing, the C-suite, and everything in between. That is also part of what gets enforced once safeguards start failing in practice. If a decision affects access, privacy, staffing, services, communication, or public trust (notice the hierarchy?), then somebody should be able to explain how it (a specific decision) was made, who shaped it, and why it moved forward in the first place. That question starts even earlier, with who actually gets included while those decisions are still taking shape.
Clear decision-making is rare for any one person, but for corporations? They seem to always know what is best for everyone, right? Yet the reasoning behind “the most important decisions” companies make remains scattered across emails, side conversations, meetings, assumptions, and invisible power dynamics that never make it into the record.
Poor tracking is not a neutral administrative problem.
Many organizations treat data tracking as a back-end issue. Of course, something could be more operational than technical, but we’ll save that for another post. Usually, when we build something from scratch, the data doesn’t necessarily start aggregating until people begin using it. So, “data cleanup” can be pushed back until there is something more tangible to work with, and it can easily slip off our radar.
Why does that matter, you ask?
Well, bits of data don’t automatically translate into KPIs. That is also where data-driven decision-making starts to get messier than the phrase suggests. They go from Bits > Bytes > Field (Data Item) > Record > File/Database > Metric/Measure > OKRs and KPIs. And since data flows aren’t inherent (I’m literally talking biologically; KPI hierarchies aren’t an innate cognitive ability), the main components of KPIs aren’t either. To begin tracking data more effectively, we need to understand the above “data hierarchy,” how that fits into developing a KPI flow, and after that, if you’re hands-on, you could build a simple dashboard to test this process out.
What gets tracked, and what disappears
Thankfully, since we weren’t born with it (inherent knowledge of data and KPI hierarchies, not Maybelline), we have computers! However, not understanding what I mentioned above will put you at a disadvantage. Because if you don’t understand what you have to work with and (not or) how a system is layered, you cannot write effective AI prompts. Nor can you accurately troubleshoot. The catch: AI will still be your friend regardless, but AI is developing rapidly, so who knows?
When a harmful decision becomes permanent simply because nobody can identify where the real decision point was, what can we do about it?
Well, that setup is not just inefficiency, but a structure. It is the kind of underlying structure people rely on without always seeing it until something breaks. To make things more complicated, it can be intentionally structured that way. Either way, once it becomes normalized, it becomes easier to confuse “administrative disorder” with inevitability. In other words, these situations make it easier to say, “That is just how the system works.”
My answer is: No. That is how the system is allowed to work.
If nobody can explain why, that is already a problem.
I am not talking about having team members raise their hands to give a hyperbolic explanation. They need to be able to demonstrate and prove it. Especially nowadays, when it seems that decisions can become (magically) legitimate because someone is good at reductionism/summarizing, or they’re an expert without question or protocol. When nobody asks questions, a process can become “good” just because it exists.
To get to my point, if nobody can explain why a certain path was taken, what alternatives were considered, who was consulted, and what risks were known at the time, then the organization is already in weaker shape than it probably realizes. For non-project/program managers, this technique is called a Decision Log (or Decision Tracker or Decision Register).
Decision Logs thankfully help prevent the need for a twenty-page memo for every decision. However, decision tracking is tricky and, from what I have seen, is usually reserved for those who execute rather than strategize and/or operationalize. Keep in mind that there are many challenges in this area. Here is a specific one: large-scale decision-making. This leaves a clear gap, as tracking should not simply vanish depending on your role.
Science says: Trust is harder to sustain when the reasoning always seems to disappear at the exact moment accountability becomes relevant. This research on transparency and trust helps explain why.
The paper trail is part of people-power.
Who documents, what gets documented, and knowing what never makes it into the record matters, too. After all, a weak paper trail does not just create confusion; it creates an uneven power imbalance.
Some people know what happened, while others are left guessing. And if the people affected by a decision are always downstream from the explanation, then the explanation itself becomes a form of control. Because now they are not only bearing the consequences, but bearing the burden of interpretation.
Weak tracking protects the system more than the people in it.
This is a part, I believe, that many organizations don’t want to admit or even think about. Not because corporations are evil, but because it is increasingly impossible to catch up on this type of organizational understanding. There are a lot of moving pieces and layered history. However, some say that since most of us have AI at our fingertips, the tables could turn (for the better) if we approach it as diligently and professionally as business entities have.
In other words, instead of poorly tracked data creating distance, we can collectively uncover and adopt truly ethical approaches to build bridges. We just have to take the time to figure it out, and that is what some of us are doing, and probably why you found this page.
We no longer want poorly tracked decisions (built systemically into systems) that become easier to defend than the people they affect. Build trust, do not break it.
And let’s face it, systems are better at processing issues than explaining them, and we need to shift away from that. That is also why accountability has to go beyond metric-driven evaluations of risk and harm.
Good tracking is not about bureaucracy for its own sake.
Disclaimer: This is where people sometimes get defensive!
Good tracking is not about making every decision more complicated than it needs to be. Instead, we can shift our thinking and approach to ensure meaningful decisions do not “disappear” (into that same system) without ample context to understand, challenge, or improve them later. In a lot of cases, that is the difference between learning from a mistake and quietly repeating it.
Instead, it is time to start being intentional about naming bad decisions what they are. To do that, we must put in the work to understand weak, systemic processes. Not making the effort to understand something and just berating it in front of your coworkers only creates noise and distracts from the real issue (figuring out an all-inclusive solution and approach that goes beyond just roundtable decision-makers).
What should be tracked?
Not everything, but certainly more than just the final action. If we don’t know where to start, perhaps we can mirror what corporations are tracking about us? For example, decision points should be traceable and not hidden in the code. Reasoning should be clear enough to revisit, and ownership should not disappear. Although ownership is quickly disappearing as companies replace people with AI.
We also must be careful not to imply something as a risk when it should be named. You can be an early adopter (with questions) in a free world, but you cannot be once it turns into an autocracy.
Lastly, if the decision affects public-facing access, privacy, services, rights, staffing, safety, or a significant workflow change, then the standard should be even higher. That is especially true when privacy and protection decisions get buried until the damage is already done. For example, can we build tools to establish accountability defined by decision-making rules to track operations within a company, even if they use proprietary tools, or is all transparency lost once something becomes proprietary? These are not minor decisions, but structural (and strategic) ones. And structural decisions deserve more than a shrug, a timestamp, and a completed status box.
At the end of the day
If a decision cannot be traced, it becomes much easier to protect the system than the people inside it. That is why tracking matters. Once a decision “disappears” into routine, accountability gets much easier to evade.
And when that keeps happening, people stop trusting not only the outcome, but the structure behind it. That is also part of what gets trusted once public doubt starts spreading faster than evidence can travel.
So when I ask what gets tracked, I am not asking for perfect recordkeeping. I am asking whether the decisions shaping people’s lives, work, access, and risk are visible enough to be understood before they become one more thing everyone is expected to absorb and never have an opportunity to comprehend, until it is too late.
These aren’t technical details, folks. These details are ethical, and they are part of the question of whether true accountability is real at all or even still possible.