Six Sigma

The point

Look at many people doing the same function, measuring productivity, efficiency, yield, and see if one's doing more than another: what he or she does, how they're doing it differently, how you can train others.

Jay Abraham · Two Day Rapid-Result "Instant Immersion" — Day 1 Cued to 3:20:33 3 min 18 sec loading…

Why it matters

Subir Chowdhury gathered what he called the forty-billion-dollar lessons of quality, and the sentence underneath all of them is the one owners find hardest to accept: good enough compounds. Quietly, stage by stage, it grows into failure.

Here is the arithmetic that makes it inescapable. Ninety-nine percent right, handed off ten times over, is not ninety-nine percent. Every stage multiplies the one before it, so a small flaw at each step does not sit still — it grows down the line until it becomes exactly what your client feels. Six Sigma simply puts a number on the distance between your work and a defect, and the further out you push it, the fewer flaws ever reach anybody.

It rests on three moving parts and missing any one of them lets the defect back in. The process is how the work gets done. The system is what holds that process in place when nobody is watching. The project is where you go and fix what the numbers have just exposed.

Now the part almost nobody does. This discipline was born on the factory floor and everyone still points it at the machine. Point it instead at the sale, the onboarding, the follow-up — everything that happens before, during and after the purchase. That is where the leverage is hiding, because the experience is what your client is actually buying.

The mistake almost everyone makes

Measuring what is easy to instrument and eyeballing the rest. The factory measured every part to a tolerance no human eye could catch; the client-facing operation gets judged on impressions, which is precisely where the defects survive.

The test: Count the handoffs between a first enquiry and a delivered result. If you cannot say the number, you are not measuring the part your client actually experiences.

Where it shows up — 8 worked examples

Who	What happened
Toyota	Any worker could pull the cord and stop the line, because a defect caught at the station costs a fraction of one caught by a customer.
Walmart and Delta	Both layered new capability onto an operating machine that already worked — forecasting and inventory, and predictive maintenance.
Zara	Turned daily store signals into new product in weeks, which is a tolerance on feedback rather than on a part.
Best Buy	Measured store usefulness rather than foot traffic, matched online prices and invested in expertise until the shop became indispensable.
Expedia	One confusing optional field was making payments fail. Deleting it reportedly produced roughly twelve million dollars in additional annual profit.
UPS	Built ORION to optimize every driver's route, shaving miles and fuel from each, so tiny per-route savings multiplied across millions of deliveries.
Jay Abraham's Chinese Motorcycle Maker	Without capital, it partnered with a non-competitor's idle factory shift, sales force and dealers; each side earned over twenty-five million dollars in the first year.
Renault-Nissan-Mitsubishi Alliance	Shared platforms, purchasing, research and development without a full merger, reporting annualized synergies of 5.7 billion euros while each kept its own identity.

What it solves

The words a business owner uses for this before anybody has told them the name of it. If one of these is a sentence you have said out loud, this is your strategy.

Good enough is our standard“Most things here are about ninety percent right and everybody has quietly decided that is fine.”The arithmetic that makes good enough fail — ninety-nine percent right, handed off ten times, is not ninety-nine percent.
The client feels what we cannot see“Our internal numbers look healthy and the complaints we get are about something else entirely.”The same tolerance the factory applies to a part, applied instead to the sale, the onboarding and the follow-up, which is what the client is buying.
Quality means the product only“We are careful about what we make and nobody has ever looked at the buying experience the same way.”The discipline pointed where almost nobody points it — everything that happens before, during and after the purchase.
It works when the right person is on“Everything runs properly when a particular person is here, and differently when they are not.”The system, which is what holds a process in place when nobody is watching, separated from the process itself and built deliberately.
Nobody counts the dropped balls“Things fall between people fairly regularly and there is no record anywhere of how often.”Handoffs counted to a tolerance, because what you measure to a tolerance stops leaking and what you eyeball does not.
The advisor's version“I improve a client's marketing and their delivery leaks most of the gain back out before anybody notices.”A way to stop the gain you produce leaking back out through a delivery operation nobody has ever measured.

The challenge

Count the handoffs and the dropped balls.

Draw the path a client actually travels from first enquiry to delivered result, and mark every point where the work passes from one person, system or step to another. Those are your handoffs. Now go back through the last twenty clients and count how many were dropped, delayed or repeated at each one. You are looking for the stage with the highest count, and it is rarely the stage anybody suspected.

How you will know it is done Every handoff on the client path drawn out, twenty recent clients counted against it, and the worst stage named with its number.

The AI layer · second pass

Once this is working, here is what to multiply

Improve the system first. Then multiply it. This panel is the second run at the strategy above, and it is deliberately useless until the first run is done.

Before you point anything at this You have separated the denominators inside one process and improved one of them.

What it multiplies Holding a thousand small operating variables at once and telling you which is drifting, continuously.

The trap Walmart and Delta layered AI on operating machines that already worked. Without the discipline underneath, you get a very precise measurement of a process nobody follows.