Agile Meridian

I ran a leadership exercise for a year and blamed the wrong thing. I kept calling it a decision-making problem. It wasn't. It was an authority problem.

The exercise assumes something it never states out loud: that everyone running it already holds the authority the model requires. Give it to someone who has that authority and it looks clean. Give it to someone who doesn't and it looks broken. Same tool, same instructions, completely different result, because the missing variable was never the person's judgment. It was their standing.

I've seen the same pattern outside leadership training, too. It shows up anywhere a model, a simulation, or a framework gets handed down without anyone checking whether the person receiving it actually carries the authority the tool assumes.

So the fix isn't a better framework. It's a smaller question, asked first.

Before you hand someone a scorecard, a decision-rights chart, or a governance model, check whether they already have the authority it assumes. If they don't, the tool isn't the problem.

Where in your org are you handing someone a framework for a call they were never actually given the authority to make?

80% of the features your teams shipped last year are rarely or never used.

Your delivery didn't fail. It delivered. The features landed on schedule and nobody came.

Pendo's research has carried that number for years, and roadmap reviews still treat shipped as done. The demo happens, the box gets checked, the team rolls to the next epic. Nobody asks whether anyone touched the thing.

Gary Cohen and I dug into exactly this on the latest Meridian Point episode, Why 80% of Features Go Unused. The discipline underneath it is uncomfortable: building the right thing is different work from building the thing right, and most organizations only staffed the second one.

Usage is a reading almost nobody takes, because taking it puts the roadmap at risk. If shipped counts as done, the plan stays green. Start counting use and half of last year turns red retroactively.

I went deeper on the fix in Obtain a Yield: A Payoff You Defer Is a Payoff You Can't Steer. A yield is a result the customer would pay for, not output you can point at in a demo.

One move for your next roadmap review: for every feature shipped last quarter, ask who used it last week. Count the shrugs.

What did your team ship this year that nobody has checked for use even once?

🎙️ NEW EPISODE: His Teams Hit Every Deadline. 80% of What They Built Went Unused.

Gary Cohen spent twenty years in software product companies building a delivery machine that hit its dates almost every time. Customers liked the features. The engine hummed.

Then a private equity board member asked one question most teams never hear until it's too late. "What additional revenue do we expect from all these features?" The honest answer sent Gary chasing a harder problem than "are we on track," the problem of whether they were building the right thing at all.

Gary Cohen joins Kumar Dattatreyan, ICF PCC to explore:

✦ Why roughly 80% of features go unused, and the board question that exposed it
✦ What embracing uncertainty actually looks like, and why it's a competitive edge instead of a weakness
✦ Why red, yellow, green status meetings turn into theater, and the watermelon project problem
✦ The questions that turn a status interrogation into real discovery
✦ Where product management goes when AI makes shipping code cheap

"We can ship stuff people don't want a lot faster."

-Gary Cohen

Listen:
Apple:https://lnkd.in/gb92KnYQ
Spotify:https://lnkd.in/gbR5YzcW
YouTube:https://www.youtube.com/live/o2sH0PqZoSQ?si=tFk4ZWvvsCXhvHXy

The launch is where most initiatives stop getting attention.

The town hall happens. The rollout hits its dates. The deck gets filed. Everyone moves to the next thing.

That's output. It's not a yield.

A yield is a result someone would pay for, not the fact that something shipped on time. The initiative gets treated as done the moment it launches, and nobody goes back to find out if it actually changed anything.

The reason is simple. Going back means you might find out it didn't land. And if it didn't land, whatever you built on top of it is standing on nothing.

I went deeper on this in Obtain a Yield: A Payoff You Defer Is a Payoff You Can't Steer. A payoff you never check is one you can't steer either.

One move: pick the biggest initiative your team launched in the last six months. Ask three people it was supposed to help whether anything actually changed for them.

What's the last initiative your team launched and never went back to measure?

You set a target to cut cycle time 30% and you never measured what it is today.

The target is theater. You cannot improve a number you have never counted. It happens every planning cycle. Leadership picks an improvement number because a number looks like rigor, and the process it points at is a black box. Nobody has watched the work end to end. Nobody knows how long it actually takes. So the 30% is invented, and every status against it is invented too.

The honest first goal is not a target. It is a baseline. Go measure the thing once, cleanly, before you promise to move it. A baseline is boring and it is real, which is exactly why it gets skipped for a rounder, braver-sounding number. I wrote about this in Friction Is Data. Your org is already running a sensor network. It is called your people, and they can tell you where the work actually stalls if you ask before you set the target, not after you miss it.

This is the same gap one layer up. On the ground, a status goes green that should be red. In planning, a target gets set on a baseline that was never taken. Same habit. A number that feels like control, standing in for a number that would take real work to get.

So before the next quarter's targets, ask one question of every improvement goal. What is the baseline, and who measured it?

If the answer is a shrug, you do not have a target. You have a wish with a percent sign on it. What target are you carrying right now that was never baselined?

You trust the green status because red would mean a conversation you don't want to have.

I've read a dashboard I wanted to be true and called it a report. Most of us have. A status goes green and stays green. It feels like control. Underneath, the work is stuck, and the people doing it have known for weeks.

They watched what happened the last time someone said red out loud, so the report stays green and the truth stays in the hallway.
That's not a status problem. It's a trust problem wearing a status costume. The dashboard didn't lie to you. The people closest to the work decided it wasn't safe to correct it.

Reported readiness and real readiness pull apart quietly, and the gap only shows up when something ships broken. By then the report has been green for a month. The signal you trust the most is the one you built to protect yourself. Green is comfortable. Green means no hard meeting today. So you stop asking the people who actually know, because asking might turn it red.

So how do you find the real number? You go to the people doing the work and you make red cheap to say.

Not 'any concerns' at the end of a call when everyone's packing up. A real question, early, where telling you the truth costs them nothing.

Before you trust the next green status, ask who would have to feel safe for it to turn red in front of you.

What's the last status you called green because red was inconvenient?