Growth Is a System, Not an Effort

Every founder I meet has a version of the same story. They spent three weeks on outreach in March and had a good April. They got busy delivering in May and had a terrible June. They conclude that they need to be more consistent.

That conclusion is wrong, or at least incomplete. Consistency is a property of people, and people are the least consistent component you own. What they actually need is a system whose output does not depend on their attention this week.

The difference between a flow and a stock

Systems dynamics gives us precise vocabulary here. A stock is a quantity that accumulates — cash, customers, published pages that rank, reviews, an email list, a body of proof. A flow is a rate that changes a stock — new inquiries per week, churned customers per month.

The reason this feels like a treadmill is that it is a treadmill. You are supplying the entire flow yourself, in real time, and the stock you are building — a bank of demand assets that keep producing after you stop touching them — is empty.

The bottom system has a property the top one does not: it has a closed path from output back to input. Delivered work produces evidence — a review, a case study, a page that answers the question the next buyer will type. That evidence raises demand without further founder input. The loop turns.

Why you cannot intuit your way through this

You might reasonably object that this is obvious. The uncomfortable research finding is that knowing it does not help much.

John Sterman ran a controlled experiment at MIT in which subjects managed a simulated inventory distribution chain containing multiple actors, feedback, nonlinearities, and time delays Sterman 1989. The task is not hard in the abstract. The results were not close to optimal. In Sterman's own summary of the Beer Distribution Game, optimal performance calculated using only the information actually available to the players is about $200 per team, while average costs come in around ten times that Sterman, MIT. Order amplitude grows steadily from customer to retailer to factory, and the peak order rate at the factory averages more than double the peak at retail.

The mechanism is misperception of feedback: people manage the variable in front of them and discount the delayed consequences of their own actions. That is precisely what a founder does when a slow month triggers a burst of outreach whose leads land two months later, arriving exactly when delivery capacity is saturated by the previous burst.

A loop is only as strong as its weakest stage

Here is where the nine-stage map at Ariadne Growth Systems↗ comes from. The path from a stranger's first search to a closed, delighted customer who generates the next stranger passes through nine distinct handoffs: demand, conversion, capture, routing, follow-up, sales, measurement, automation, compounding.

Each one is a place the loop can break. And because they are in series, their yields multiply. If a visitor converts to an inquiry at rate $y_1$, the inquiry reaches a human at rate $y_2$, and so on, then the end-to-end yield is

which has an unpleasant property: nine stages that are each 90% effective give you

Under 40%. Not because anything is broken, but because "pretty good" nine times in a row is not good. This is why buying nine tools from nine vendors, each optimizing its own stage, reliably produces a system that nobody is accountable for.

That third number deserves its own post, and it will get one. For now, note what it means: in the peer-reviewed study by Sabnis and colleagues, roughly 70% of leads generated by marketing departments were not pursued by sales representatives at all Sabnis et al. 2013. Their evidence base was 461 sales reps across four firms. The single largest leak in most systems is not at the top of the funnel. It is a handoff in the middle that nobody owns.

What "build the system" actually means

Three commitments follow from the above, and they are the ones I work under.

One map of the whole path. If you cannot draw the nine stages of your own business on one page with a number on each arrow, you are not managing a system. You are managing whichever stage most recently annoyed you.

One place the data lives. Yield per stage is only knowable if a single record follows the person from first touch to closed deal. Nine dashboards give you nine local truths and no global one.

One person accountable for the outcome. Not for impressions. Not for open rates. For whether more of the right customers arrived and stayed.

The promise on the front page of Ariadne is that your business should not stop growing the moment you stop pushing. That is not a motivational statement. It is a structural claim about closing a loop, and it is testable: instrument the nine arrows, and watch whether the stock keeps rising in a month you did nothing.

References

1. Sterman, J. D. (1989). Modeling managerial behavior: Misperceptions of feedback in a dynamic decision making experiment. Management Science, 35(3), 321–339. https://doi.org/10.1287/mnsc.35.3.321
2. Sterman, J. D. (n.d.). The Beer Distribution Game. MIT Sloan School of Management, System Dynamics Group. https://web.mit.edu/jsterman/www/SDG/beergame.htmlAuthor's own summary of the 1989 experiment; source of the widely repeated 'ten times optimal' figure.
3. Sterman, J. D. (2000). Business Dynamics: Systems Thinking and Modeling for a Complex World. Irwin/McGraw-Hill. https://www.mheducation.com/highered/product/business-dynamics-sterman.htmlTextbook, not peer-reviewed; the standard treatment of stocks, flows, delays and policy resistance.
4. Sabnis, G., Chatterjee, S. C., Grewal, R., & Lilien, G. L. (2013). The sales lead black hole: On sales reps' follow-up of marketing leads. Journal of Marketing, 77(1), 52–67. https://doi.org/10.1509/jm.10.0047