BOLDGROUPTHAILAND Back to overview
All insights
FIELD NOTE 04 / INNOVATION STRATEGY

Your innovation strategy only knows one way to win

Most approval frameworks recognise a single route to advantage. There are two, they need opposite kinds of evidence, and the companies that last use each in service of the other.

01IN BRIEF
  • There are two routes to advantage, and most approval frameworks recognise only one of them.
  • The Shaper builds a market that does not exist yet. The Dominator wins the market that does, on price and performance. Both are legitimate; neither is sufficient alone.
  • The companies that last use each code in service of the other — and the ones that mastered only one, from Kodak to Intel, were at their most confident just before it stopped working.
  • The practical fix is one question asked before a project enters your pipeline: is this a Shaper project or a Dominator project? The answer decides which framework should govern it.
02THE ARGUMENT

There is a question that tends to surface in the quiet after a difficult quarter. You look at what the organization has put into innovation — the capital, the time, the attention of your best people — and ask why none of it produced the advantage the textbooks promised.

The problem is the framework, not the ambition

Most companies run every proposal through a single approval process: define the stages, require a forecast return, release funding against milestones. That process was designed for a world where the shape of the market was known and the question was execution.

It encodes one theory of how a company wins. And there are two.

The Shaper sets out to create a market that does not exist yet, and to write its rules. The work is the systematic de-risking of a conviction about the future.

The Dominator sets out to win the market that already exists, so decisively that competing becomes uneconomic. The work is the systematic removal of cost from delivering a thing people already want.

Neither is the superior strategy. They are different jobs, they need different evidence, and a framework built for one will quietly starve the other.

The Shaper's code: build the market that is not there

A Shaper's defining move is usually one that looks irrational from outside, because it is priced against a future the organization believes in and the market has not yet seen.

Netflix is the cleanest example. It held the DVD-by-mail market outright, and its leadership concluded the future was streaming. Acting on that, it built a service that would take apart its own profitable business — choosing to make itself obsolete on its own timetable rather than someone else's. Airbnb assembled a home-sharing market that had no prior category. SpaceX rebuilt the economics of private launch.

What these have in common is not boldness for its own sake. It is that the conviction came first and the evidence was then built deliberately, in stages, to test it. That is the part most organizations skip.

The Dominator's code: win the market that is

A Dominator's engine is tuned to a single measure: the ratio of price to performance. The question is not what new thing to make, but how to deliver the existing thing at a cost nobody can match.

Walmart held retail for decades on the strength of a supply chain, not a product. McDonald's reached global scale through process standardisation rather than cooking — a predictable product, at a predictable price, in any country.

This is innovation, and it is frequently misfiled as operations. The work of taking cost out of a system at that scale is as demanding as inventing a category, and considerably better understood by most finance functions.

The duality most leaders miss

Presented as above, the two codes look like a choice. Examined closely in the companies that have sustained an advantage, they turn out to be a pair — each deployed in service of the other.

Tesla is read as a Shaper, and its most consequential move was a Dominator's. The electric car was not Tesla's invention; it had existed for years as an expensive niche. What changed the market was solving the price-performance problem at volume. Gigacasting is that move: a Shaper-scale bet on unproven manufacturing technology, made in service of a cost position. A single press casts an underbody section that was previously assembled from many separate stampings.

Apple is read as a Shaper, and it protected the iPhone with a Dominator's discipline. Two projects ran in parallel. One was the safe evolutionary path — a phone built on the proven iPod platform and its click-wheel. The other was the revolutionary one: a desktop operating system shrunk onto a phone, driven by a multi-touch interface built from nothing. Funding both was portfolio risk management of the most conventional kind, used to shelter a fragile idea long enough for it to survive.

The pattern is consistent. A Shaper-scale bet is made safe by Dominator-grade discipline, and a Dominator's cost position is reached by a Shaper-scale technical wager. The codes are not alternatives to pick between. They are tools that work on each other.

Why purebreds go extinct

Mastering only one code is not a stylistic preference. It has a consistent failure mode, and the failure arrives at the moment the company looks strongest on its own measures. That exposure is widely felt: in McKinsey’s Global Innovation Survey, 80% of executives said business models are at risk. The survey states no fieldwork date, so the figure is used here for the scale of the felt exposure, not as a current measurement.

The pure Shaper invents the future and does not get to own it. The companies that created the portable digital music market in the late 1990s built the category and then lost it, because they had no comparable discipline in supply chain, retail or software. What displaced them was not a better device but a better system — a player and a store designed as one thing.

The pure Dominator becomes excellent at making something nobody wants. Kodak's mastery of chemical engineering and global-scale manufacturing was genuine, and it was aimed entirely at film. The company held an unmatched price-performance position in a product category that was ending — and it had helped invent the digital technology that ended it.

The pattern is not historical. Intel's long dominance rested on perfecting the price-performance of one processor architecture. That focus became a blind spot when the highest-value computing work shifted to massively parallel processing, and Nvidia — a Shaper in that new market — took it.

The Intel case is worth understanding mechanically, because it explains why a competent incumbent misses a shift rather than losing a fight. A traditional processor is built like a small number of extremely fast engines, executing complex instructions one after another. A graphics processor is built like a very large fleet of modest ones, each handling a simple task at the same moment. Training an AI model is millions of simple calculations that can happen simultaneously — perfectly suited to the fleet, and a poor use of the fast engines however fast they are. Intel had perfected the wrong machine for the work that turned out to matter.

The test that sorts them

The answer is not to discard your approval framework. It is to stop applying it before you know what kind of thing you are approving.

Add one question ahead of the pipeline, for every initiative: is this a Shaper project or a Dominator project?

If it is a Dominator project, your existing framework is well suited. Forecast return is the right test, because the thing being improved already exists and can be extrapolated from. The leadership challenge is to push the team toward a genuine technical leap rather than settling for safe increments.

If it is a Shaper project, your existing framework is actively dangerous, because it will ask for evidence the project cannot produce yet and will stop it for failing to. This work belongs in a protected space, measured on validated learning and the systematic reduction of uncertainty — and governed with Dominator-grade discipline in staged funding and explicit validation milestones.

That single question converts the framework from a uniform hurdle into an instrument that can tell two different kinds of work apart. Most of the value is in having asked it before the money is committed.

Where this does not apply

This is a strategy diagnostic, not an excuse. A Shaper label does not exempt a project from evidence; it changes what evidence is appropriate and when. A project that cannot say what would have to be true for it to work, or what it would cost to find out, is not a Shaper project — it is an unexamined one.

Nor is the balance meant to be even. For most established companies the Dominator portfolio should be considerably larger. The argument is that the split should be a decision somebody made, rather than the residue of a process that could only see one kind of proposal.

One boundary worth stating plainly: every company named above is a public example, analysed from published history. None is a Bold Group client, and nothing here reports work we did.

03SOURCES
  1. McKinsey Global Innovation Survey, Exhibit 2

    McKinsey & Company · Undated; read from the 15 November 2017 capture

    First-party exhibit, and the source HBR's own reporting links to. States no fieldwork date and no sample size. Source of the 80% business-models-at-risk figure; the same exhibit carries the 84% and 6% cited in The innovation mirage. Its text is outlined to vector paths, so it is readable only as a rendered image.

  2. Exploration and Exploitation in Organizational Learning

    James G. March, Organization Science · 1991

    The academic ancestor of the distinction this article draws in strategic terms. Attribution rather than a cited statistic: March's argument concerns organizational learning, and the framing here concerns approval criteria.

  3. Gigacasting: the hottest trend in car manufacturing

    S&P Global Mobility · November 2023

    Background for the Tesla example. Reported figures vary with what is being counted: the rear underbody casting alone replaced roughly 70 parts, while the front and rear castings together are reported to have replaced about 171 parts and eliminated some 1,600 welds. The article states the mechanism rather than a headline number for that reason.

04COMMON QUESTIONS
Is this the same as exploration and exploitation?
It shares an ancestor — James March set out that distinction in 1991 — but the framing here is deliberately about strategic intent and approval criteria rather than organizational learning. The Shaper and Dominator codes describe what a company is trying to win, which is what determines the evidence a project should be held to.
How do we tell which type a project is?
Ask what has to be true for it to pay. If the answer can be extrapolated from something that already exists in your market, it is a Dominator project. If it depends on a market, behaviour or price point that does not yet exist, it is a Shaper project — regardless of how the proposal is written.
Does a Shaper project need a separate team?
It needs separate criteria and a protected funding route. A separate team is one way to achieve that and carries its own cost, because work built outside the business often fails to return to it. The essential thing is that the project is not judged on forecast return before it can produce one.
Are Netflix, Tesla, Apple and the others Bold Group clients?
No. They are public examples analysed from published history, used because their strategic decisions are documented and verifiable. Nothing in this article reports work Bold Group performed.
What proportion should go to Shaper work?
There is no general answer, and any figure offered without knowledge of your market cycle and balance sheet should be treated with suspicion. The useful discipline is that the proportion is chosen, stated, and reviewed rather than left to whatever the approval process happens to permit.
05WHERE THIS LEADS
WHAT SHOULD CHANGE?

Bring the outcome, the constraint, and the accountable decision into the same conversation.

Discuss the outcome that matters