The most common sequence is also the one that works worst: you hire somebody with a good technical profile, ask them to “innovate”, and eighteen months later they are firefighting for production. What has to exist before the first hire costs nothing, and the order you hire in afterwards matters more than the headcount.
The most common sequence is also the one that works worst: you hire somebody with a good technical profile, sit them in an office and ask them to “start innovating”. Eighteen months later that person has become production support, because production did have urgent problems and the department had nothing to defend.
What has to exist before the first hire is three things, and none of them costs money:
If production has a serious emergency tomorrow, does the R&D person go and put it out? If the answer is yes, there is no department yet: there is a person with a job title. And that is a legitimate decision in a small company — what does not work is having it without having decided it.
The most expensive mistake is copying the structure of a company four times your size. Pick the size closest to yours and you will see what is sustainable in-house and what is better kept outside.
—
—
You bring in-house what has to be repeated many times and what cannot be written down: product knowledge, the customer relationship, design judgement. You outsource what is used in bursts and requires equipment or accreditation: testing, occasional simulation, patent searches, regulatory writing. Hiring somebody for something done six times a year is expensive and it also bores the person doing it.
The order matters more than the headcount. These four hires, in this sequence, hold a department together; in any other order they take themselves apart.
Before anybody with a specialist profile. Somebody who can keep a portfolio alive, say no, and get things finished. Without this role, any specialist who joins afterwards becomes production support within months.
Whichever matches what actually defines your product: mechanical design, electronics, materials, process. One, not three. And the one that solves the real bottleneck, which is rarely the most exciting one.
The bridge to production and to suppliers. It is the hire that pays for itself fastest, and the one usually made latest — normally after a piece of tooling has gone wrong.
From here it depends on the sector. If the product is regulated, the fourth is regulatory and there is nothing to discuss. If it is not, it is the second technical discipline that shows up in the portfolio twice running.
It is the most counter-intuitive decision on the list and the one most often made backwards. A department made of one technical person has nobody to defend it when urgency calls, and urgency always calls. The first hire is not the one who knows the most: it is the one who holds the portfolio together.
Measuring an R&D department by the number of open projects rewards exactly the opposite of what is needed. These signs can be read without a dashboard.
A stalled project and an explicitly cancelled one are different things. Cancelling with a written reason means there is judgement and the portfolio is real.
When the plant starts taking problems to the department instead of solving them on its own, it means it recognises the department as useful. It is the most reliable sign and it appears in no indicator.
After a couple of years there has to be product sold that did not exist before. Not prototypes or proofs of concept: product with an invoice.
If looking back at last week almost all of it was firefighting, the department exists on the org chart and not in practice.
Twelve open projects and none closed in a year is not capacity: it is a lack of judgement about choosing. Fewer projects at a time finish sooner.
If when somebody leaves the reasons behind the decisions go with them, there was no department: there was one person doing their job very well.
This piece is ordered experience, not a standard: there is no standard that says how to build a department. These references do give a frame to the parts that have one.
A certifiable Spanish standard defining an innovation management system: processes, responsibilities, portfolio and follow-up. Useful as an index of what has to be decided, even if you never certify.
International guidance on the innovation management system. More general than the Spanish standard and with no certification attached; useful for shared vocabulary with international customers.
The international reference for defining what counts as innovation and how it is measured. It is what sits behind official statistics and a good share of public funding forms.
The phase-gate model. Its practical contribution is forcing an explicit decision at each gate about whether a project continues, which is why it fits the sign “projects get closed by decision”.
It proposes splitting the portfolio explicitly between core, adjacent and transformational. It is the conversation that gives the separate budget its meaning.
The document that says best whether there is a department or not. If in six months no decision to close or change course appears, the portfolio is not being managed.