AI can't help you with Discipline

I was on a call with a Director of Product at a company that sells smart meters to the utility industry.

She expressed her problem and wish for her org very succintly:

"Leadership changes our priorities mid-launch and that scrambles everyone. I thought we had a good process in place. But I think a good process would prevent changing priorities."

She wasn't describing a broken team. She was describing a good one — with most of the right pieces in place.

The gap was discipline.

Here's what's inside:

• What happens to the business when leadership changes priorities mid-quarter
• What it looks like when a team builds the evidence first — and why the scramble almost disappears
• It's not better prioritization, its something else

Be the PM who Ships a great product. That's the only thing that gets you promoted or poached. Join Cohort 5 next week. I'm in.

​

The cost of switching priorities

Someone catches what looks like a "prioritization error".

So the roadmap decided 6 weeks ago is now thrown out and a new one handed to everyone, with 6 weeks left in the quarter.

Now everyone is context switching. Handoffs get messy, the launch date is approaching and teams start to scramble.

When this happens its inevitable that great product work won't happen. Without a great product, you don't get measurable outcomes and without those, you don't get a Business Result.

There goes the win you were going to bring to your management for your promotion.

Or the story you were going to sell in your next product interview.

Wasn't the whole "prioritization process" and "quarterly roadmap review" meant to prevent this situation?

Teams who Kill ideas don't scramble

Not because they're better at prioritization.

Because the evidence was built before the pressure arrived.

When the criteria for Build or Kill an idea are clear — and everyone agrees on them before the meeting starts — the decision stops being political.

It becomes obvious. The idea either has evidence or it doesn't. If it doesn't, it gets killed. Quickly. Without drama.

The roadmap stays lean. The priorities stay stable. And when something does shift, as things inevitably do, there's room to absorb.

Killing product ideas is not a personality trait and it doesn't come from having a tough VP or a rigorous PM.

It's a set of constraints that make the right behavior the default behavior.

Here's what a Discovery culture looks like

I've seen it work. Here's what's in place when it does:

• A VP who won't approve a roadmap item without evidence. Not instinct. Not a strong opinion. Evidence.
• Trios — PM, UX, and Engineering — running discovery together and showing up to leadership as a unit.
• A sprint cadence built around Build or Kill decisions, not feature delivery.
• A team that presents kills as confidently as it presents builds. Killing is celebrated, not apologized for.
• Shared criteria that make the decision obvious rather than political.

That kind of culture creates the conditions for great product work to happen.

Install the Process, maintain Discipline

A structured discovery process is something you install once.

The install always starts with a team working in a new way, showing a small win and then scaling it to the rest of the product org.

Now nothing gets on the roadmap without evidence. That's the rule.

But after the install, its the discipline to stick to the process that matters. The temptation to switch priorities will always be there.

When that happens, you look at the evidence your team uncovered to keep you from losing focus.

That's what discipline does and AI can't.

From Insight to Action

• In my product career, wish I had a place to ask dumb questions. Join my free Office Hours and ask me and other PMs anything: ​April 16, or April 23​​
• The only thing that matters in your product career is what you ship. Check out my free webinar: "Stop Guessing. Ship the right product with Claude Skills."​
• ​Here's my Claude Skill you can use to de-risk your launch. Copy/paste it into Claude and press enter. Let Vera guide you.