"We need an innovation department first." Why you don't

No, you do not need an innovation department before you start. You need a problem worth solving, a sponsor with authority, a small team with protected time and a simple way to decide what happens next. A department can come later, once you know what work it should hold. Built first, it usually becomes an office waiting for a mandate.

The belief, stated fairly

The case for a department first is not foolish. It goes like this: innovation is everyone's job, so in practice it is nobody's. Without a dedicated unit there is no budget line, no headcount, no one accountable and no signal from leadership that this matters. A department makes innovation visible, gives it a seat at the table and protects it from being crushed by daily operations.

Every part of that is partly true. Visibility, accountability and protection do matter. The mistake is assuming that a department is the only way, or the first way, to get them.

Why it fails in practice

It separates the idea from the owner. Most valuable innovation in an established organisation changes something that another unit runs: a service, a process, a product line. When a separate department generates the ideas, the people who must adopt them were not in the room. The handover becomes a negotiation, and the receiving team has every reason to say "not now". Vijay Govindarajan and Chris Trimble, in The Other Side of Innovation, argue that innovation initiatives need a partnership between a dedicated team and the existing organisation, which they call the "performance engine". A standalone department that skips the partnership is exactly the setup they warn against.

It lets everyone else off the hook. Once there is an innovation department, line managers can reasonably say innovation is "their job". The department ends up asking other units for time and access it has no authority to demand.

It front-loads cost before there is evidence. A department needs a head, staff, space and a budget before it has shown a single result. That makes it an easy target in the first tight budget cycle. Units that are built before they prove anything are often the first to be cut. The sibling article on why innovation labs get closed in year two covers this pattern in detail.

It defines the structure before the work. You cannot design a good unit until you know what it will do: run challenges, support teams with methods, manage a portfolio, scout partners, or something else. Each needs different people. Organisations that build first often hire for a generic "innovation" profile and then discover the work they actually need is facilitation, or portfolio governance, or procurement of pilots.

The standards don't require it. If the argument is compliance, check the text. ISO 56002, the guidance standard on innovation management systems, asks top management to assign responsibilities and authorities, and explicitly allows them to sit with existing leaders and functions as well as with dedicated roles. A department is one option, not a prerequisite.

What works instead

Start with a light structure that does the three things a department promises (visibility, accountability, protection) without the overhead.

1. A named sponsor. One senior person who owns the outcome, protects the team's time and has agreed to act on findings.
2. A small working team drawn from the units that own the problem. Three to six people, part-time, with a few hours a week each. The people who will run the new way are the people who design it.
3. A clear decision point. Agree in advance when and how you will decide to continue, change or stop. A short challenge panel with the sponsor and one or two outside voices is enough.
4. Written roles. A one-page RACI for the first effort: who decides, who does, who must be consulted, who is informed. It removes most of the "whose job is this?" friction a department was meant to solve.
5. A record of what you tried. Keep a simple log of initiatives, decisions and lessons. After two or three cycles, a portfolio map of that log tells you what kind of central function, if any, you actually need.

Then, after six to twelve months, ask the department question again, with evidence. You may find you need a small central team of two or three people who coach and convene, not a department that owns innovation. You may find you need nothing more than a regular portfolio review. Or you may find there is enough repeatable work to justify a unit, and now you know exactly what it should do.

An illustrative example

Illustrative. A regional logistics company in the Netherlands wants to "get serious about innovation". The first draft plan proposes an innovation department with a director and four staff.

The chief operating officer asks for a different start. She sponsors one problem herself: late delivery slots booked by retail customers. A team of five forms from customer service, planning and the warehouse, each giving half a day a week. They agree a RACI on one page and a decision meeting at week eight.

By week eight they have tested a new booking confirmation by hand with twenty customers and changed one planning rule in one depot. The COO decides to roll the rule out to two more depots. Over the next two cycles, she notices the teams keep needing the same things: someone to facilitate, someone to run the decision meetings, and a shared log. That is what she funds: one full-time facilitator and a quarterly portfolio review. Not a department.

Do's and don'ts

Do

• Start with a sponsor, a part-time team and one problem.
• Put the people who will run the change in the team that designs it.
• Write down roles and the decision point before you begin.
• Revisit the structure question after six to twelve months, with evidence.

Don't

• Don't hire a head of innovation before you know what the job involves.
• Don't create a unit whose success depends on other units' goodwill and give it no authority.
• Don't cite a standard as the reason for a department without reading what it asks for.
• Don't treat a new org chart box as a signal of commitment. Protected time and acted-on findings are the signal.

Questions people ask

Should a company have a dedicated innovation department? Sometimes, eventually. A dedicated unit makes sense when there is a steady volume of repeatable innovation work, such as running challenges or managing a portfolio, that the rest of the organisation cannot absorb. It rarely makes sense as the first step.

Who should own innovation if there is no department? A named senior sponsor owns each initiative's outcome, and the line units that run the affected service own the change. A small coordinating role, often part-time at first, can keep the record and run decision meetings.

What is the difference between an innovation team and an innovation department? A team is formed around a problem and usually draws people from the units that own it. A department is a permanent unit with its own headcount and budget. Teams are cheaper to start and closer to adoption; departments offer continuity once the work is understood. For a wider comparison of structures, see innovation lab vs centre of excellence vs innovation unit.

Where to go next

To set up the light structure, use RACI for roles, a stakeholder map to see who must be involved, and a retrospective after each cycle. For the whole first month, read where to start with innovation in your team, and if you are working in the public sector, starting innovation in a government entity.