Why almost fixed is still broken

A Product PM simplifies onboarding to boost signups, cutting a verification step that was slowing new users down. The change works. Activation numbers climb within weeks, and the team celebrates a clear win.

A month later, Customer Success is quietly absorbing a spike in cancellations. Users who signed up without understanding what they were signing up for are now churning as fast as they arrived. Yet nobody connects the two. The dip looks like a Customer Success problem, because that's where it surfaced.

Optimize the part or optimize the system

This is not a rare failure. It is one of the most common patterns in organizational life, and it rarely gets named correctly. The instinct is to treat it as a Customer Success problem: a retention issue, an onboarding experience that needs polish, a team that should have caught the churn sooner. Each explanation is plausible on its own. None of them explain why the decision that caused it looked, at the time, like an unambiguous win.

The same pattern shows up at the strategic level, not just the operational one. A plan gets built with clear rationale, strong logic, and genuine ambition. It still fails. Not because the reasoning was sloppy, but because the reasoning stopped at the effect the plan was designed to produce. It never extended to the effects the plan would produce elsewhere in the system. The strategy was locally sound and organizationally blind — the same way the onboarding fix was locally sound and organizationally blind.

What connects the two is not incompetence, and it is not lack of effort. It is what the decision-maker was able to see. Some leaders, faced with a problem, see the problem. Others see the problem as one node in a wide set of relationships. It's this second kind of seeing that determines whether a fix holds, or simply relocates the damage to wherever it's least visible from where the decision was made. This second kind of seeing is what's usually meant by "systemic thinking": not analyzing more variables, but registering, by default, that a decision belongs to a wider set of relationships rather than to one part alone.

Not seeing the bigger picture isn't the only factor blocking the implementation of systemic solutions to complex organizational problems. Other blockers have already been described in other editions of this newsletter. Here, we examine how they must be addressed together as a set to understand what it actually takes for organizations to embed systems thinking into their decision-making. To see how they interact, we'll trace the lifecycle of that single onboarding decision through the six blockers that stand in its way.

The anatomy of a systemic break

1. The Frame of Attention

The first blocker is the most basic, and the easiest to misdiagnose. Before an organization can fix information flows, collaboration structures, or accountability, someone has to notice that a decision reaches beyond their own part of the organization.

The Product decision wasn't blocked by missing information; the PM could likely have found out how onboarding friction affected retention if the question had occurred to them. It didn't occur to them because the decision was framed, and experienced, as a Product decision. Nothing in that frame included Customer Success, so nothing prompted a check.

What was missing was a habit of attention. Paying attention is what determines whether a decision registers, by default, as touching a wider set of relationships before any deliberate analysis begins. Habits of attention are shaped by exposure and by what a role has historically been asked to notice. An organization can cultivate this habit or suppress it, but it cannot install it in an afternoon. This habit is what separates tactical competence from strategic judgment: a leader missing it isn't failing at strategy through lack of vision, but missing the raw material strategic reasoning is built from.

2. The weight of the metric

Even where a decision-maker clearly sees that a choice reaches beyond their own function, seeing isn't the same as weighing. If a PM's targets, review, and compensation are built entirely around activation, Customer Success's outcome has no place in that calculation — however clearly it is seen. This isn't reluctance; it's that nothing in how the role is measured gives the wider consequence any weight in the decision.

This blocker can survive every other one being cleared. A PM might close the attention gap completely, know exactly what Customer Success is measured on, and still make the same call, because the decision is still evaluated, formally and informally, against activation alone. Seeing the wider system and being measured against only part of it can coexist indefinitely. Nothing forces them into alignment.

3. The blind spot

Even where the attention gap is closed, noticing is not enough on its own. Closing it means knowing there is something you don't know — that a decision reaches beyond your own function. It doesn't mean knowing what it is you don't know. A decision-maker can sense a gap in their picture without having any sense of its shape. What actually fills that gap is knowing what the organization is trying to achieve, why it matters, what the decision-maker is empowered to do, and how escalation works when the answer isn't clear.

Suppose the PM notices that Customer Success might be affected. Without knowing what Customer Success is measured on, what tradeoff the company actually wants between activation and retention, or who to ask before shipping, that awareness has nowhere to go. The PM correctly senses that Customer Success is affected, but has no reliable way to reason about it. It becomes awareness without the material required to act.

4. The distortion downstream

A fourth blocker shows up even after the first three are cleared. A leader can have the attention, the weight, and adequate information, and still watch a sound decision distort as it moves through the organization. Each layer beneath the original decision reinterprets it locally. Each local interpretation is reasonable on its own, but the compounding effect by the time it reaches implementation can bear little resemblance to what was intended.

The original decision might have been narrow and well-reasoned: simplify one specific verification step for one segment of new users, while leaving the rest of onboarding untouched. By the time engineering builds it, "simplify" has become "remove," and "one segment" has become "everyone," because each team along the way made a locally sensible call without the full picture the original decision was built on. Systemic thinking fails here not at the point of decision, but at the point of transmission. The person who made the decision may have seen the whole picture correctly, but the chain that carried it downward did not preserve it.

5. The missing channel

A fifth blocker is structural in a literal sense. A decision-maker can notice that a choice touches another function, understand the tradeoffs, and still have nowhere to take that insight. No channel exists for negotiating a shared decision across the two functions involved.

Even a PM who does everything right up to this point — notices the risk, weighs it properly, understands the tradeoff, and wants Customer Success's input before shipping — may simply have no standing meeting, no shared decision forum, and no process where that conversation is expected to happen. Raising it becomes a favor to ask rather than a step in the process. Peer agreement between functions has a ceiling, and goodwill is not a system. Systemic attention identifies that a conversation needs to happen, but it doesn't create the forum for having it.

6. The altitude dilemma

The last blocker is the most decisive, because it defeats systemic decision-making even when every other blocker has been cleared. A role can be correctly identified as accountable for a decision that spans two mandates — the crossover point role — and still be unable to act on that accountability.

The higher up that role sits above where the decision gets implemented, the less likely it is to have real visibility into downstream conditions. Authority over both mandates and distance from the ground tend to be related. This isn't a manager declining to get involved; it is a role with authority over a system it often cannot see clearly, because the same altitude that gives it authority also puts distance between it and where the system actually operates.

The matrix trap: trading one blocker for another

Organizations are not blind to these blockers. Matrix structures, for example, attempt to solve this problem by splitting the formal accountability otherwise held by the crossover point manager between two functional heads, who jointly oversee one shared report. This is meant to build the missing channel for negotiation, forcing the two heads to work out competing priorities directly rather than escalating upward.

In practice, that negotiation often doesn't happen in time. The tension between the two mandates lands instead on the person reporting to both heads, who has to act when their two bosses want different things but without having the authority to resolve that conflict themselves.

The organization has made a trade, not a fix. Having a single decision-maker was good, but being several levels removed from implementation was bad. That was the altitude problem. The matrix trades the single decision-maker for two, which reopens the missing channel problem in a new form. But it also closes the distance to the implementation line, which resolves the altitude problem. One problem improves, while another reappears in its place.

This is the clearest evidence that these blockers aren't independent obstacles to clear one at a time. They interact. A structure built to solve one can quietly reintroduce another, because the underlying forces are deeply connected.

What systemic decisions actually require

Systemic decision-making requires addressing all six blockers together, because none of them alone is enough.

An organization can train people to notice wider relationships, yet watch decisions ignored because nothing in how the decision-maker is measured gives them weight. It can align incentives, yet watch decisions stall for lack of information. It can supply better information, yet watch a sound decision distort as it cascades downward. It can build a forum for two functions to negotiate, yet leave the role accountable for the outcome unable to see the ground it is accountable for.

Each fix is real, but none of them, alone, closes the gap. Each blocker is independently capable of breaking a systemic decision, regardless of how well the others are handled. Fixing one doesn't fix the rest. It simply moves the point of failure to whichever blocker remains unaddressed.

Real progress requires stepping back to treat the six blockers as a single, interdependent system. That means auditing the entire chain before intervening: asking for any critical choice where the signal is most likely to break, recognizing how a structural fix at one point might recreate friction at another, and accepting that an organization's decisions are ultimately only as systemic as its weakest link.

Alexander Spradling, PhD, is an organizational consultant and executive coach based in Austin, Texas. He writes Why Organizations Misbehave.