54% of a Room Full of Product Leaders Said Their Default Is to Make Every Call

At the PDMA summit in Atlanta this October, Anshul Garg, a principal product manager at Amazon, opened his workshop with a quick poll. He described four leadership modes and asked everyone in the room to pick the one closest to their default. The Explorer provides direction and makes fast decisions. The Builder coaches, asks questions and develops the team's capability. The Scaler removes blockers, designs systems and builds collaboration. And the Steward sets the vision, delegates outcomes and builds the next generation of leaders.

The results: 54% Explorer, 33% Scaler, 8% Builder and 4% Steward.

Garg wasn't surprised. Most rooms of leaders gravitate to Explorer, he said, because "we feel ourselves more in control when we are providing the direction, we are providing fast decisions." It's one room's self-assessment, not a study, but Garg's own story explains why it matters. And the number at the other end is just as telling: only 4% of the room picked Steward, the mode whose whole job is building the next generation of leaders.

The mode that launches a product

Back around 2015 and 2016, Garg was at Capital One, leading a cross-functional team building the next generation of credit card payment alerts – around 1.8 billion alerts a year across 19 dimensions, with real consequences for customers who might miss a payment or get hit with a late fee. The stakes were high and the timeline was tight, so he did what most of that room said they'd do. "I tried to make all the decisions," he said. "I was involved in every engineering discussion."

At first, it worked very well. The team got clarity and speed, and they loved it. Then, as the problem got defined and the work picked up momentum, he started to feel friction. The same team that had been happy for him to make every call was hesitating to accept his decisions, and "somehow I was becoming a bottleneck."

That's the moment a lot of leaders misread. When a team starts pushing back on decisions it used to welcome, it's tempting to conclude the team has gotten less committed. Garg's read was that the team hadn't changed at all. The work had, and it needed something different from him.

Around the same time, he took a course on situational leadership, the idea that people need different kinds of leadership depending on where they are in their experience. That's when it clicked for him that products work the same way. Leading a two-person team searching for product-market fit calls for someone who defines a clear north star and resolves ambiguity fast. Leading a 200-person team scaling a mature product calls for someone who improves processes, builds mechanisms between teams and thinks long-term. The leadership mode that launches a product will hold it back at scale.

Founders get stuck in the first mode

This is a common trap for founder-CEOs, and it isn't a character flaw. Making every call is the right behavior when the company is five people and the founder has all the context. The problem is that the habit outlasts the stage it was built for.

The research on company growth has said this for a long time. Larry Greiner's classic model of organizational growth, summarized well here, rests on the idea that "management practices that work well in one phase may bring on a crisis in another." And investors act on it. Harvard Business Review reported that venture capitalists replace 20% to 40% of their founders "with more seasoned and 'professional' managers at critical transition points." In most of those cases, the job changed underneath the founder.

From the team's side, it looks like waiting. Decisions sit in a queue for the founder's calendar, and the small ones wait longest because the founder rightly spends time on the big ones. People stop building judgment of their own, because they learn that what's wanted is a well-prepared question for the founder rather than an answer. As the Rework piece puts it, that's how a company ends up with managers who can't act alone, "which then justifies the founder staying involved." And if the metric the company cares about isn't moving, one of the reasons may simply be that every path to moving it runs through one person.

Give away the outcome, not the steps

The way out of Explorer mode is to give people ownership of outcomes instead of handing them steps to follow, and then to change what you spend your own attention on. The Rework piece describes the end state well: the founder "moves from approving individual decisions to reviewing patterns," and the question changes from "should we do this?" to "is the system producing good decisions?"

I try to make the handoff explicit, because people can't own something they aren't sure they've been given. When I hand something off, I say it plainly: you're in charge of this, you're the owner, and you have to figure it out. I say that out loud because people often ask someone else to decide out of fear – of doing it wrong, or of doing it differently from how you wanted it done. So I also tell them the part that's hardest for a founder to say: you won't do it the way I would, and that's okay. The goal is you doing it. Afterward we look back together at what went well, what didn't and what we'd change.

That only works if the leader means it, and it rests on something more basic. I can't really control anyone on a team. All I can do is uphold my end of the deal between us – take care of them, give them what they need – and give them good reasons to do what the work needs. Leading by influence instead of control is slower in the moment and much faster over a year.

My own failure runs the other way

Where I sit on this isn't in the 54%. On the spectrum between doing everything yourself and being completely absent, I sit closer to the absent end. I give people autonomy on the assumption that they're capable or will learn to be.

That has its own characteristic error. Sometimes I've handed someone an outcome they weren't yet equipped to own, and when that happens I treat it as my failure, not theirs. It's the mirror image of the Explorer's problem, and it's a useful reminder that Garg's point isn't that one mode is good and the others are bad. Each one fits a stage, and the skill is noticing when the stage has changed.

Which mode does your product need now?

If you're a founder, it's worth asking which stage each of your products is in, and whether your leadership has moved with it. A new product still searching for fit may need you deciding fast and being in every room. The product that pays the bills and has 50 people working on it probably needs you designing how decisions get made, then getting out of the way.

The 54% aren't wrong to be Explorers. Plenty of products need one. The trouble starts when the product grows up and the leader doesn't.