The assumption nobody examines
Most organisations do not have a data problem. They have a decision problem, and the data has quietly become a way of looking like it is solved. Dashboards multiply, reports pile up, the volume of available information climbs every quarter, and the quality of the decisions that information is meant to improve does not move with it.
The phrase “data-driven” has become a credential rather than a discipline. Underneath it sits an assumption that is rarely said out loud: that if you put enough good data in front of capable people, good decisions follow on their own. They do not. Data informs a decision. It does not make one. The thing that makes the decision is the system around the data - who owns the call, on what cadence, against what definition, with what record of whether it worked - and in most organisations that system was never designed. It accreted.
There is also a point at which more data stops helping and starts hiding. Past it, additional dashboards do not produce more clarity. They produce more places to look, more figures to reconcile, and more cover for putting the decision off another week. The organisation feels rigorous because it is surrounded by numbers, but rigour is not proximity to data. It is what you do with the data when the figures disagree and a choice still has to be made.
Where decision-making actually breaks
In live environments the breakdown is almost never the analytics. It is the connective tissue between the analysis and the action.
The most common version is the trusted number that is never interrogated. A figure lands on a slide and is treated as fact because it came from “the system.” Nobody asks how it was defined, what it excludes, or whether the question it answers is the question being decided. In our experience the most dangerous metric is rarely the one that looks wrong. It is the one that has sat on the dashboard so long that questioning it feels rude. More often than not, a number everyone trusts is measuring something subtly different from its label: an “active user” that counts a login rather than a use, a “resolved ticket” that counts a closure rather than a fix. The label travels and the definition quietly does not. This is partly a question of how the numbers underneath are defined and governed.
Close behind it is the decision with no owner. Analysis gets commissioned, circulated, and discussed, but accountability for acting on it is spread thinly enough that no single person carries it. When everyone has seen the data and no one is named as the decider, the default outcome is delay, and the same analysis returns to the same meeting a month later.
The quietest failure is the missing feedback loop. The organisation decides, acts, and never checks in any structured way whether the decision was right. Without that loop the analytics function cannot improve, because it never learns which of its outputs actually changed an outcome. Every decision becomes a one-way door with no note of what was on the other side.
Fixing the system, not the dashboard
The fix is not another platform. It is treating the way decisions get made as something you design on purpose. In practice that means naming an owner and a cadence for every material recurring decision, so analysis attaches to a real moment of action instead of floating free. It means letting each key metric carry its definition and its known limits with it, so interrogation is built in rather than heroic. And it means logging decisions with their reasoning and revisiting them against what happened, so the organisation learns instead of repeating itself. It also depends on people positioned to act on what the system surfaces, which is a question of how data teams are built and wired in.
This is the gap we built Satchel & Boot to close. Rather than hand over a dashboard and an invoice, we put our people inside the client environment to own the decision system day to day. We call them the Boots: embedded delivery, accountable for the outcome, not for the deck. The change we see most often is unglamorous. A recurring decision that used to drift between meetings acquires an owner, a date, and an agreed definition, and then it simply gets made, on time and on the same basis each cycle. The technology rarely changes. The habit does.
Key takeaways
- The constraint is rarely the volume of data. It is the design of the decision system around it.
- The failures are organisational: trusted-but-uninterrogated numbers, decisions with no owner, and no feedback loop.
- Owners, defined metrics, and decision logs change outcomes more than another dashboard does.
- A decision system has to be lived inside the business, which is why we embed rather than advise and leave.