In 2015, a fintech company launched in Lagos with a genuinely differentiated product. It solved a real pain. Early traction was fast. Word spread. Users invited other users. Revenue compounded. The press covered it. More users came from the press. More revenue came from the users. More press came from the revenue.
By 2017, it had ten times the users it had started with. By 2019, it was gone.
When people tell the story of what happened, they tend to identify a specific cause — a bad hire, a competitor move, a regulatory intervention. These are accurate as facts. But they are the wrong level of analysis. What actually destroyed the company was a balancing loop that was always there, building pressure beneath the surface, waiting for the reinforcing loop to push hard enough to wake it up.
That loop — the one that kills growth-at-all-costs companies — is one of the most important patterns in all of Senge's and Meadows' work. And understanding it is the difference between building a company and building a rocket that flies briefly before the structural failure becomes visible.
Reinforcing loops: the engine of growth and the engine of ruin.
A reinforcing loop is a feedback structure in which a change in one direction produces more change in the same direction. Meadows uses the phrase "virtuous cycle" for the positive version and "vicious cycle" for the negative. But the structural reality is the same either way: the loop amplifies whatever is already happening.
The classic growth flywheel is a reinforcing loop. More users create more word-of-mouth. More word-of-mouth creates more users. A marketplace with more sellers attracts more buyers, and more buyers attract more sellers. A developer platform with more third-party apps becomes more valuable, which attracts more developers, who build more apps. These are not strategies — they are structures. The question is whether you have designed your business to contain one.
Most founders in Africa understand this concept intuitively, even if they have never called it a reinforcing loop. What they do not understand as clearly — and this is where Senge's work becomes most valuable — is that every reinforcing loop generates a balancing loop in response. Every engine of growth contains a constraint. And the bigger the reinforcing loop, the bigger the balancing constraint it will eventually encounter.
In the Lagos fintech story: rapid user growth (reinforcing) was eventually met by a customer service quality collapse (balancing). The company's ability to serve users well degraded faster than it could hire and train support staff. Users who had a bad experience told other users. The word-of-mouth that had previously driven growth began driving churn. The same mechanism — social proof and peer recommendation — that had built the company was now dismantling it. Same loop. Reversed direction.
The balancing loops founders miss.
Senge describes a pattern he calls "limits to growth" — one of the most common and most underappreciated structures in organisational dynamics. The pattern works like this: a reinforcing loop drives growth. The growth pushes against a constraint that was always in the system. The constraint creates a balancing loop that slows and eventually stops the growth. The founder, seeing the slowdown, pushes harder on the reinforcing loop — more marketing, more sales, more product features — which accelerates the encounter with the constraint rather than resolving it.
The critical insight is this: pushing harder on the reinforcing loop when a balancing constraint is active makes things worse, not better. The leverage is not in the engine of growth. It is in identifying and addressing the constraining structure.
For an African edtech company I am thinking of — I am not naming it because the founder is still building and the story is still live — the reinforcing loop was clear. Schools that adopted the platform reported better learning outcomes. Better outcomes produced referrals to other schools. More schools produced more data. More data improved the product. The loop was real and working. Growth was fast.
The balancing constraint was teacher capacity. The platform required teachers to change their classroom methodology significantly. In underfunded Ghanaian schools, teachers are already overextended. The additional learning burden created quiet resistance. Adoption was nominal — schools bought the product, but teachers did not use it deeply. The learning outcomes that were supposed to drive referrals did not materialise at the expected rate. Growth slowed. The founder pushed harder on sales. More schools adopted nominally. The pattern worsened.
The leverage point was not in the growth loop. It was in the teacher capacity constraint, which required a different kind of intervention entirely. Not a sales investment. A teacher training investment, a curriculum simplification, a product redesign that asked less of already-overloaded educators. The constraint, not the engine, was where the work needed to happen.
Delays: the silent distorter.
One of Meadows' most important observations about feedback loops is that delays in the system produce oscillating behaviour that looks irrational but is perfectly logical once you understand the structure. When information about the state of the system arrives late, decision-makers overshoot their targets — correcting too hard in one direction before the feedback from the previous correction has arrived.
She uses the example of a shower with a slow hot water heater. You turn it on. The water is cold. You turn it hotter. Still cold. You turn it hotter. A minute later, scalding water arrives. You lurch toward cold. Cold water arrives. You lurch back. The oscillation is not because you are incompetent at temperature regulation. It is because the delay between your action and the feedback from your action is long enough that you cannot behave optimally.
African markets have long delays built into almost every feedback loop. Customer behaviour changes slowly because formal switching costs are low, but informal social switching costs (telling your community about a bad experience) are high and move on social timelines. Regulatory feedback is slow because institutions move slowly. Investor sentiment feedback is slow because the African investor community is small and information travels slowly. Talent market feedback is slow because hiring and onboarding cycles are long.
The result: African founders routinely overcorrect. They make a pricing decision and do not see the customer response for three months, by which point they have already changed the price twice more. They launch a product feature, and the market feedback arrives six months later, long after they have built on top of the feature. They enter a partnership and discover its strategic implications only after the operational costs have become embedded.
The discipline that long-delay feedback loops require is patience with information. The temptation — which Senge describes as almost universal — is to mistake the absence of feedback for the absence of response. The system heard you. It is just responding on its own timeline, not yours. Your job is to act, wait for the signal, and resist the urge to act again before the signal arrives.
Drawing the loops in your own business.
I want to give you a practical exercise that I use with founders and that I learned from working through both Meadows and Senge's frameworks.
Take a single important outcome in your business — customer growth, revenue, churn, team performance, anything that matters. Draw it at the centre of a blank page. Now ask: what directly increases it? Draw an arrow from that thing to your outcome. Now ask: does your outcome, when it increases, affect that thing in return? If yes, you have a feedback loop. If the return effect reinforces the original direction, it is a reinforcing loop. If it resists or counteracts the original direction, it is a balancing loop.
Now keep going. For every element you identified, ask what affects it. Draw the arrows. Keep asking. Keep drawing. You will very quickly produce a map that looks complicated — because the system is complicated. But within that complication, you will start to see patterns. You will see where reinforcing loops are driving your growth. And you will start to see where balancing loops are waiting, underappreciated, for the growth to push against them.
When you have the map, the question becomes: where are the delays? Where does action in one part of the system take a long time to produce a signal in another part? These delays are the places where founders most commonly make their worst decisions — acting before the feedback from the last action has arrived, stacking interventions on top of each other until the system is so full of uncompleted corrections that reading it becomes impossible.
The companies that survive the long run in Africa are not the ones that push their reinforcing loops hardest. They are the ones who understand their constraining loops well enough to address them before they become limits. And they are the ones that read delays correctly — distinguishing between a system that has not yet responded and a system that is not going to respond.
Those are very different situations. They require very different responses. And telling them apart is the discipline that this series is ultimately about.
This is Post 2 of 7 in the Systems Founder series.
