Choosing Among Good Ideas Is the Hard Part

Category: Focus, Direction, and Completion Capable environments fail when idea generation outpaces selection.

The whiteboard at the end of the offsite was the most encouraging thing the founder had seen in months, and it was also the problem, though it took her a while to understand that those were the same sentence.

Her team had spent two days generating ideas, and they had been brilliant at it. The board was covered — three columns deep, written small to fit — and not one of the ideas was filler. There was a partnership that could open a whole new channel. A pricing change that several people had wanted for a year and could clearly defend. A product line that customers kept asking for unprompted. A hire that would unblock a team that had been stuck for two quarters. A piece of automation that would pay for itself by spring. She walked the board slowly and found that she agreed with all of it. Every idea on that wall was something a competent person had thought hard about and could argue for convincingly. There was nothing to cross out, because nothing was wrong.

And that was the trap she didn’t see closing. She left the offsite elated — look how much we could do — and the team left believing the offsite had been a success, because generating thirty good ideas feels like winning. Over the following weeks she watched the board turn into the year, and the year went the way the board had: a little progress on everything, the channel half-opened, the pricing change drafted but not shipped, the product line scoped but not started, the hire discussed in three separate meetings and made in none of them. The team was working hard on a generous spread of genuinely good things, and at the end of it she could not point to a single one that had landed with its full weight.

What she had been short of was never ideas. Her team manufactured those by the dozen, effortlessly, and would do so again at the next offsite. What she had been short of — what she had never built, never staffed, never treated as real work — was the thing that comes after the ideas: the deliberate, difficult act of choosing which few of thirty good things would get funded properly and which twenty-seven good things would be killed so the few could live. She had a machine for making ideas and no machine for choosing among them, and in a capable team that second machine is the one that’s actually scarce.

The Principle

In a capable environment the bottleneck is never the supply of good ideas — it is the capacity to select among them. Generating worthy options is the easy, abundant half; deciding which few to fund fully and which many to kill is a distinct, difficult, under-valued skill, and the environments that get somewhere are the ones that build selection as a real function rather than assuming it will happen by itself.

The flattering story a talented team tells itself is that its edge is creativity — that the hard part is having the ideas, and that once the ideas exist the rest is just execution. For a weak team that story might even be true; if you can barely generate one decent option, generation is your constraint. But the moment a team becomes genuinely capable, the constraint moves, and most teams don’t notice it move. Ideas stop being scarce. They arrive faster than anyone can act on them, each one defensible, each one championed by someone who isn’t wrong. At that point the limiting resource is no longer invention but judgement under abundance — the ability to look at a wall of things all worth doing and deliberately choose the two that get everything and the twenty-eight that get nothing. That choosing is not a formality that follows the real work. In a capable team it is the real work, and it is hard precisely because every option you kill is a genuine loss.

Why It Is Inevitable

This isn’t a quirk of badly run teams; it’s a structural consequence of getting good, and it lands on anyone who crosses a certain threshold of capability.

The first reason is that competence inflates the supply of ideas far faster than it inflates the capacity to execute them. A team gets better and the rate at which it generates worthy options climbs steeply — more skilled people, more context, more customers, more pattern-recognition all pour more good ideas onto the pile. But the capacity to actually carry an idea to completion grows slowly, if at all; it’s bounded by hours and attention and the stubborn fact that finishing things takes concentrated push. So a widening gap opens between what the team can think of and what it can do, and that gap is filled by exactly one activity, whether the team names it or not: selection. The better you get, the more selecting you are forced to do, because the more good ideas you are forced to leave on the floor.

The second reason is that good ideas don’t sort themselves, and they actively resist being sorted. A bad idea is easy to drop; it has a flaw you can point at. A genuinely good idea has no flaw — that’s what makes it good — so there is no natural reason it falls off the list, and worse, it has a champion who can defend it sincerely and an upside that is real. When everything on the board has a credible case, the board does not narrow on its own. Left alone, all the good ideas stay, each one too defensible to cut, and the team’s finite capacity gets divided among them until every one is starved. The only thing that narrows a list of good ideas is a deliberate act of selection — somebody choosing to kill worthy things on purpose — and nothing in the ideas themselves will ever supply that act.

So any environment that becomes capable is pushed to the same hard place: it must build a way to choose among more good ideas than it can fund, or it must accept that its abundance of good ideas will quietly translate into a shortage of finished ones. The environments that go on to do something are simply the ones that treated selection as a function to be designed and staffed, rather than a thing they assumed would take care of itself once the clever ideas were in the room.

How It Shows Up

• The number of good ideas in circulation visibly exceeds what the team can act on, and everyone knows it — the constraint is plainly selection, not invention.
• There is an explicit, owned step where ideas are chosen among, not just collected; the funnel narrows on purpose, at a known point, by a known method.
• The criteria for what gets funded are written down and shared, so a selection can be explained — “this beat that, for these reasons” — rather than feeling like a mood or a favour.
• Good ideas get killed, openly and on the record, and the killing is treated as the system working rather than as a failure of nerve or imagination.
• The few things that survive selection get the whole weight of the team, not a slice, because the point of killing the rest was to make that concentration possible.
• Nobody mistakes a productive brainstorm for progress; the team values the act of choosing as highly as the act of generating, and resources it accordingly.

Why It Causes Benefit

When selection is built as a real function, a capable team finally converts its abundance into output instead of letting it dissolve, and the gain is larger than it first appears.

The immediate benefit is concentration. Because the many good ideas have been deliberately killed, the few survivors receive enough force to actually cross the line into done — and a capable team aimed at two things will land both with a completeness that the same team, spread across thirty, could never achieve on any of them. The talent that was being diluted across a generous spread is now pushing a narrow front, and a narrow front moved with full weight breaks through. The very capability that made the idea-glut inevitable becomes formidable the moment it’s pointed, and selection is what points it.

The deeper benefit is that selection makes the team’s quality legible and therefore improvable. When choices are made against explicit criteria and recorded, you can look back and see why the surviving ideas were chosen and ask whether the choosing was sound — which means the selection skill itself gets better over time, the way any skill does once its outputs are visible and examinable. A team that selects by gut and never writes down why learns nothing from its choices; a team that selects against stated criteria builds, over years, an actual taste — a sharpening sense of which of many good things is the one worth everything. That compounding judgement is the rarest asset a capable organisation can have, and it exists only where selection was treated as a discipline rather than an afterthought. The abundance of ideas was always going to be there. What separates the teams that do something with it is that they got good at the second, harder thing — choosing — and kept getting better at it on purpose.

How To Cultivate It

• Name selection as its own step, owned by someone, happening at a known point — distinct from idea generation. Generation and selection are different activities with different temperaments, and collapsing them into one meeting means the loud generation drowns the quiet, unglamorous choosing. Give the choosing its own time, its own owner, and its own respect.
• Write the criteria down before you choose, not after. A selection made against stated criteria can be explained, defended, and later examined for whether it was right; a selection made on instinct and rationalised afterward teaches nobody anything and feels, to the people whose ideas were cut, like a whim. The criteria are what turn killing good ideas from politics into method.
• Practise killing good ideas explicitly, on the record, with the reason being “we chose something else, not that this was bad.” The hardest cut is the cut to a genuinely worthy thing with a sincere champion, and that is exactly the cut selection is for. Doing it cleanly — honouring the idea while still ending it — is the core skill, and it has to be made normal or it won’t happen.
• Cap how many ideas can be live at once and treat the cap as binding, so that admitting a new good idea requires choosing it over an existing one rather than simply adding it. A funnel with no cap isn’t selecting; it’s collecting. The forced trade-off — this instead of that — is where real choosing lives.
• Separate “this is a good idea” from “this is one of the few we will fund,” out loud, every time. Almost everything that reaches the selection step will be a good idea; that is no longer the question. The question is whether it beats the others for the scarce capacity, and keeping that distinction sharp is what stops the list re-filling with everything that merely passes the low bar.
• Review your past selections deliberately, to sharpen the judgement rather than just to make this round’s calls. The point of recording why you chose is so that, later, you can see which choices paid off and refine the criteria — building the taste that lets a capable team pick well. Selection that’s never reviewed never improves; selection that is, compounds.

What Good Looks Like

A capable team that generates good ideas easily and is entirely unflattered by that fact — because it understands the ideas were always going to come, and that its real edge is what it does next. There is a clear, owned moment where the many good ideas are chosen among, against written criteria, and the funnel narrows on purpose; good ideas are killed openly and without rancour, the killing understood as the system working rather than failing. The few that survive get the full, undivided weight of a talented team and so they land — completely, on a rhythm — and the team can always point at what it actually finished rather than the impressive spread of what it considered. Over time the choosing itself gets better, because the choices were recorded and examined, and the team develops a taste — a sharpening sense of which of many worthy things is the one worth everything. The place is not short of ideas and never was; what it built, and most capable teams never do, is the harder machine downstream of the ideas — the one that chooses among them. And that machine, unglamorous and quiet next to the brilliant brainstorm, turns out to be the thing that decides whether all that capability becomes anything at all.

A Reflective Question

When your environment last failed to deliver, was it really short of good ideas — or was it drowning in them, with no deliberate, owned way to choose which few would get everything and which many would be killed? And if it was the second, have you ever treated selecting as a skill worth building and resourcing — or have you kept investing in generating more good ideas you already had no capacity to choose between?