How to measure innovation

How to measure innovation: count the job, not the theater. Patents, hackathons, and “number of ideas” measure activity. Hours returned, scrap down, and people who did not go back to the old way measure innovation.

Leading vs lagging

Adoption this week is leading. Margin next quarter is lagging. You need both. A team that only reports activity will look busy and change nothing. A team that only reports lagging numbers will kill tests too early or too late — they cannot tell which.

Do not average bets

One dead project and one live one is not “50% success.” Kill the dead one in public. Fund the live one. Averages are how bad bets hide.

What a 20-person shop should measure this month

Before you buy another suite for “How to measure innovation,” write three numbers on a whiteboard: unplanned stops (hours), changeover (last chip to first good chip), scrap (pcs or $). Two weeks of honest numbers beat a demo.

If you cannot fill the board, the first job is a clock and a reason code — not a dashboard. Software that starts without a clock becomes a nicer argument.

• Unplanned hours by machine, not a plant average.
• One timed changeover per cell, written on the traveler.
• Scrap with a cause in one word: tool, program, stock, inspect, other.

When software is the lever — and when it is not

Software helps when the clock exists, the stop has a name, and a person still decides what ships. It does not help when the real problem is a missing fixture, a tribal setup, or a mill waiting on inspection.

SINLE Technologies LLC will scope a monitor, a report, or a lock on who stops a job. We will not sell you a MES to hide a queue. A Workflow Audit is paid discovery. Nothing is billed before a written scope.

A human lock on the floor

A model can flag a hole, a tool, a late job. A person still decides scrap or ship. Unsupervised scrap-or-ship is how you train a lawsuit. Name who clicks. Name the kill switch. Put both in the statement of work.

The steps — do these in order

1. 01

   Pick one outcome metric per bet

   Hours, scrap, cycle time, paid conversions. One.

2. 02

   Pick one adoption metric

   Weekly users of the new way. A tool with no users is a cost.

3. 03

   Time-box the read

   Two weeks or a month. Eternal pilots are unmeasured on purpose.

Keep reading

• Innovation strategy

  How to build an innovation strategy

  A strategy is what you will not fund. Pick a job family, a time box, and a kill rule — not a lab with a slogan.

• Failure

  Why some innovative ideas fail

  No user, no Tuesday test, no kill rule, and a team that needed the idea to be true. Failure is usually adoption, not IQ.

• OEE

  OEE software for small manufacturers — three numbers, not a suite

  Availability, performance, and quality. How a small plant uses OEE without a $100k MES or a dashboard nobody believes.