Your Capital Plan Needs a Glucose Monitor

You spent two hundred million dollars rehabilitating your sewer system.

Did it work?

I do not mean did the projects finish, or did the money get spent. Those always happen. I mean: did the system get healthier, and can you show me the curve?

I have watched rooms full of smart people fail to answer that question, and the reason is not laziness. The reason is that we measure infrastructure the way medicine measured patients in 1985: one blood pressure reading a year, filed away. A snapshot. A static KPI.

Now look at what a continuous glucose monitor gives a diabetic. Not a reading: a curve. The spike after the meal, the response to the insulin, the drift overnight. The trend, and the trend's answer to the intervention. Same body, same chemistry, completely different kind of knowing.

A continuous glucose monitor does what your SCADA dashboard mostly does not: it shows you the curve, not just the number.

This is what managing by curve looks like: not a number frozen in time, but a system talking back to you in real time.

A static KPI tells you a number. A cybernetic KPI tells you whether what you are doing is working.
Hardeep Anand

Wiener's insight from 1948 is still the cleanest definition of what separates a reporting system from a steering system.

That difference has a name, and it is older than you think. In 1948, Norbert Wiener at MIT founded cybernetics, the science of systems that steer themselves through feedback. He took the name from the Greek kybernetes: the steersman of a ship. The founding word of the science behind automation and AI is a man steering a vessel through water. This science was always ours.

This is what managing by curve looks like: not a number frozen in time, but a system talking back to you in real time.

The framework: the Curve Test

One framework, four steps, drawable on a napkin. Run it against any program your utility funds.

Name the vital sign. Which KPI is this money supposed to move? Not "improve the system." Name the number.
Draw the baseline. What has that curve done for the last three years? No baseline, no test.
Spend, then watch the curve. Not the budget curve. The physiology: did the trend bend after the intervention?
Feed the answer forward. The bend, or its absence, reshapes the next capital decision. That is the loop. A loop without step 4 is just reporting.

Walk the four steps in order and the Curve Test turns a vague program goal into a falsifiable claim about system health.

Four steps on a napkin: name the vital sign, draw the baseline, watch for the bend, and ask what the loop is telling you.

Three vital signs, read like a doctor

Non-revenue water is internal bleeding. Treated water, already pumped, cleaned, and paid for, vanishing before a bill. The US loses roughly six billion gallons of it a day. The static report says "our NRW is 28 percent" the way a chart note says "patient is bleeding." The Curve Test: district metering finds the bleed, repairs close it, and the next four quarters of trend are the proof of healing. The IWA water balance and AWWA M36 are the instruments; the loop is what most utilities are missing.

Infiltration and inflow is fluid overload. Groundwater leaking into sewers overloads the plant the way edema overloads the kidneys. The honest test of a $200M rehab: when the next big rain comes, does the flow spike at the plant bend lower than last year's? Flow against rainfall, year over year. If nobody can produce that chart, the program was measured by its spending, not its physiology.

Pollutant loads are cholesterol. An MS4 load target is an LDL number. You do not cure it in one procedure; you manage it down with BMPs, diet and exercise for a watershed, and you watch the labs trend across seasons. The permit report is the lab panel. The trend is the health.

Side by side, a distribution system schematic and a cardiovascular diagram reveal the same pressure, flow, and loss logic.

Your distribution system and your cardiovascular system are running the same operating logic, and the vital signs that matter look almost identical.

Boards do not need more numbers. They need to see the curve bend after the money lands.
Hardeep Anand

"But my meters are twenty years old"

This is the honest objection, and it is usually right. Most systems cannot pull a clean continuous reading off the infrastructure they have. Here is the part that matters: you almost never need new sensors to start. The first curve comes from data you already own, sitting in pieces. Flow at the plant against rainfall is in your SCADA history. Consumption against production is in your billing. The baseline for non-revenue water is in records you already file. The first project is not instrumentation. It is connecting what is already there into a line you can read. New meters make the curve sharper later. They are not the price of admission.

Even imperfect data can show you a trend, and a trend is infinitely more useful than a single number reported once a year.

The bend in the curve is the moment your investment proves itself, and if you cannot see it, you cannot defend the next dollar you spend.

The money: prevention versus the emergency room

Healthcare knows the arithmetic: the colonoscopy costs a thousand dollars, the surgery it prevents costs a hundred thousand. Infrastructure runs the same arithmetic with more zeros. A capital plan is preventive medicine; every project buys down a specific, nameable risk before the event. The alternative is the emergency room, which in this sector is called a consent decree. Atlanta's ran on the order of four billion dollars. The program worked, but it was forced, late, and enormous: the price of deferred prevention, set by a court instead of a board.

The cost gap between a diagnostic program and an emergency repair is the arithmetic that every board presentation should open with.

The bend in the curve is the moment your investment proves itself, and if you cannot see it, you cannot defend the next dollar you spend.

And the cascade does not stop at the fence line. Energy is the system's metabolism. Resilience is its immune system. The end of the cascade is property values, public health, and the rates of the people the system serves.

Trace the feedback loop from pump energy to effluent quality to public health and you see why no single metric can tell the whole story.

A utility's vital signs end up in its citizens' bodies and wallets. That is not a metaphor. That is what the loop does when you follow it far enough.
Hardeep Anand

Follow the loop all the way out and you will find that infrastructure health and community health are not two stories, they are one.

Takeaways

A static KPI is a snapshot; a cybernetic KPI is a trend plus its response to intervention. Demand the second kind.
Run the Curve Test on every funded program: name the vital sign, draw the baseline, watch the bend, feed it forward.
NRW is internal bleeding, I&I is fluid overload, pollutant loads are cholesterol. The body twins are not decoration; they are how boards finally see.
Buying down risk early is the colonoscopy. The consent decree is the emergency room.
None of it works without connected, trustworthy data underneath: be data-ready before you are AI-ready.

Next steps

If you sit on a board or run a utility, take one question into your next meeting: "For the program we just funded, which KPI curve will prove it worked, and when will we see it bend?" Then ask whether your data can even draw the baseline. If it cannot, that is the real first project.

Go deeper: the OneWater Lexicon pages on cybernetics and PDCA explain the loop science in three layers each, and the APAS Labs course on data readiness turns the Curve Test into a working program.

Next in this series: The Steersman, on why three systems that look nothing alike, nature's water loop, the one we built, and your own body, all run the same machinery. Subscribe so it lands in your inbox.

Reply with one KPI curve that bent after a capital program in your system. That chart is the most underrated document in public infrastructure, and I collect them.