Innovation Management

Once a company finds its footing, the way it makes decisions gets more rigorous. A process appears that asks for evidence before money moves, and that process tends to ask the same questions every time. How much goes in, when does it come back, and what are those numbers based on. These are good questions, invented to keep companies out of trouble. The catch is that some proposals can answer them and some cannot.

A proposal to extend the current business has plenty of material. Last quarter's conversion rate exists. The number of customers who asked for the thing has been counted. There is a record of what happened to revenue the last time a comparable feature shipped. A proposal to try something nobody has tried has none of that. There are no customers yet, so there are no requests to count, and nothing has been built, so there is no precedent to compare against.

Quarterly planning therefore ended the same way over and over. Someone would finish presenting a new idea, someone else would observe that there was no revenue estimate attached, and the conversation would close. What made it hard to argue with is that the observation was correct. There genuinely was no estimate, and any estimate offered would have been invented. Put a feature that reliably brings in thirty thousand dollars this quarter next to an attempt that may or may not work in two years, on the same scale, and the near-term one wins. Not once — every time.

The way things get dropped is consistent too. Outright rejection is rare. Items get moved to next quarter, because rejecting something requires a reason and deferring it does not. The deferred item meets the same logic next quarter and moves again, and after the third move it quietly disappears from the list. Nobody ever decided to kill it, and it is dead all the same.

Three years of that leaves a company with one product that has gotten slightly better each quarter. If the market stayed still, nothing is wrong. The trouble starts when something else has appeared in the meantime that solves the same problem a different way. At that point the company has not failed to notice the shift. It noticed and could not act. The meeting notes contain someone naming the shift, and the line underneath records that something more urgent was due that quarter.

The economist Clayton Christensen was the first to describe this shape clearly. What he pointed at was not laziness. Well-run companies fall into it more readily than sloppy ones. Listening carefully to existing customers and directing resources toward the highest-margin work are correct decisions, and making those correct decisions repeatedly is what causes the miss. Being well run is both the compliment and the diagnosis here. The tighter a company's process, the less room an unevidenced proposal has to slip through, and the fewer chances there are to generate the evidence that would let it through.

Innovation management is the response that came out of that diagnosis. Instead of blaming individuals or urging everyone to be bolder, it separates the scale. Budget and headcount get set aside, the yardstick applied to them changes, and new attempts stop competing for priority in the same line as certain revenue.

The terms under innovation management fall into three groups: the concepts that explain why this is hard, the tool that partitions the problem, and the activities that fill those partitions.

Concepts holds three entries, and they read as one sequence.

1. Innovator's dilemma — the trap where an existing business is the reason a company misses what comes next. Building precisely what current customers ask for and investing where margins are highest accumulates into an inability to respond while a cheaper, simpler substitute climbs from below.
2. Disruptive innovation — the move from the other side that sets the trap. It starts out worse on the dimensions incumbents care about, which makes it easy to find reasons to ignore. Because it is cheap and simple, it wins over people who were previously buying nothing at all, and once that base funds the climb in capability, responding is already late.
3. Moonshot — an attempt with a high chance of failure and a chance of changing the board. The dividing line is the size of the goal. Making document editing ten percent faster is doing the current job well. Making it unnecessary to open a document at all attacks the premise of the current job. The second kind usually fails, which is exactly why nobody touches it when success rate is the measure.

Tools holds one entry. Three horizons splits a portfolio into current, expanding, and exploratory work and manages the three separately. The first bucket is what earns revenue now, the second is what should reach revenue in two or three years, and the third is exploration that may not pan out. Its real function is not sorting projects but separating the yardsticks. The first bucket is judged on profit, the second on growth rate, the third on what was learned this quarter. Attach a revenue target to a third-bucket project and it is immediately being graded by first-bucket rules, which returns the company to the original problem. The seventy-twenty-ten split often quoted in resource allocation comes from this frame; the value is less in the specific numbers than in writing down a non-zero share for the third bucket in advance. Where the partition usually collapses is people rather than money. Budget is an accounting line, so moving it somewhere else is visible. People get reassigned quietly whenever something urgent appears, and a name that sits on the exploratory roster while spending four days a week on incident response looks perfectly fine on paper.

Activities holds two entries, separated by where the new business is placed.

1. New business development — building a business in a different area from the existing one. If Pagedeck, which sells a document tool, wants to sell workflow automation as well, that work belongs here. The market gets sized again, whether anyone will buy is tested with an MVP, and pricing is decided from scratch rather than inherited.
2. Intrapreneurship — building it the way a startup would, inside the company. The term names a working arrangement rather than a subject, which puts it on a different layer from new business development. A small team gets its own budget and decision rights and an exemption from part of the approval process, so it can change direction on its own instead of waiting for a committee that meets every two weeks.

The three groups complete each other. The concepts explain why leaving this alone fails, the tool redraws the scale to match that diagnosis, and the activities fill the partitions with actual work. With only one of the three, things usually fade. Share only the diagnosis and people nod in the meeting while the budget stays as it was. Draw only the partitions and the third bucket exists as a label with nothing in it. Run only the activities and the people get pulled back to urgent work before the first quarter ends.

"This is a third-bucket project, so judge it on what we learned this quarter, not on a revenue estimate."

Someone is asking for a different yardstick to be applied inside the same meeting.

"If we only keep pushing the plan that sells well now, in three years this is the only thing we sell."

The concern is not this quarter's result but an empty product list three years out.

"Take that team out of the central approval process. If decisions slip two weeks at a time, that's how fast they learn."

Budget alone is not enough; decision rights have to be separated too.

At Pagedeck the subject got its first real hearing while planning the budget for the year after the Series A. Grace set aside ten percent of the total for exploratory work. A quarter later, almost none of it had been spent, and the people assigned to it were handling incidents and enterprise customer requests. The money had been separated and the people had not.

On the second attempt Mia pulled two people out full time and removed every revenue-linked item from their quarterly review. That produced grumbling from other teams about one group having it easy. With nothing visible to show, the reaction was predictable.

What Grace settled on was making the learning public in place of results. Each quarter the team presented what it had tried and what turned out to be wrong, including the projects it was shutting down and why. Two quarters in, the slot had become less a progress report than a shared record of where the company had been looking, and the complaints stopped. The part about announcing what got shut down did most of that work. A team that kills nothing looks idle; a team that says it killed two of three ideas reads as a team making calls.

The most common mistake is reading innovation management as idea collection. Hackathons and internal idea contests carry this name and are the visible part, so the association forms early. Ideas are rarely the scarce input, though. Most companies already have more proposals on record than they can execute. What is short is the basis for choosing which ones to test, the people and weeks to test them with, and a procedure for shutting down the ones that come back negative. Leave those empty and collect anyway, and the board fills with a few hundred submissions against a handful of things actually attempted. The cost of that state is easy to miss. The people who submitted stop showing up to the next round, and then the collecting stops too. This is why work in this area usually starts with filtering rather than gathering: count how many things can genuinely be tested in a quarter, pick that many and staff them, and mark the rest as not this time rather than pending. Keeping the list short is harder than lengthening it.

The second mistake is narrowing innovation to new technology. The word gets used mostly at technology companies and the cited examples are usually technical, so the link forms naturally. In practice, the changes that move the board often come from how money is charged or how something is sold. Software moving from a one-time purchase and installation to a recurring subscription was not caused by a new technology; the pricing structure changed. Read narrowly, this work gets assigned to engineering alone, and attempts to change pricing or the sales motion never reach the innovation list — they scatter into each team's ordinary improvement backlog. Then choices that should be examined together at the corporate strategy level get split across departments and pull in different directions. In one quarter sales pushes for longer annual contracts while product prepares usage-based pricing. Each is reasonable on its own, and together they make it impossible to read from the outside where the company is headed.