This year, UK management scholars will publish roughly 12,000 peer-reviewed articles. A handful will be read outside the academy. Fewer still will change how an organisation operates, or how a manager decides anything. This is not a complaint about academic life. It is a diagnostic question, and I think it is the most important question a doctoral researcher can ask before choosing a topic.
What does research worth doing actually look like, in a field where the supply of research has outpaced the demand for it?
I have been asking candidates this for fifteen years. The answer that survives the longest tour through my office sounds like this: research is worth doing if it can change a mind that is already thinking carefully about the thing you are studying. That is a tougher test than it sounds.
Three tests I run before I read the methods
When a proposal arrives on my desk, I run three quiet tests before I look at the research design, before I read a single citation.
1. Will someone act differently within a year?
If the answer is yes — if a CFO reads your study and runs a slightly different audit, or an HR director adjusts their onboarding, or a supply-chain lead reconsiders a supplier — the research is worth doing. If the answer is "it contributes to the literature", I have a harder question: contributes how, to whose decision, under what circumstances?
This is not anti-theoretical. Much practice-changing research is deeply theoretical. But the good ones carry their theory lightly, in service of a specific human doing a specific thing. Granovetter's embeddedness, Thompson's rules-in-use, Edmondson's psychological safety — these earned their theoretical weight because they survived contact with a practitioner who had to decide something.
2. Would the finding surprise the practitioner community?
Not the scholarly community. Not the average reader of the Academy of Management Review. The practitioner community. If a seasoned supply-chain director reads your abstract and says "we all know that" — and they are not being dismissive, they just already know — then you have confirmed something, and confirmation is a real contribution. But it is a different contribution than surprising a well-informed reader.
The best doctoral theses I have supervised land just ahead of the practitioner's intuition. The finding is something the practitioner almost knew, but had not articulated, could not have defended, and did not have evidence for. The research clarifies the intuition into a defensible claim.
3. Can you defend the evidence against a sceptical reader who knows the field?
This is the hardest test. A sceptical statistician will find problems with your design. A sceptical practitioner will find problems with your sample. A sceptical scholar will find problems with your framing. You do not need to satisfy all three. You need to satisfy one of them completely, and not lose the other two.
What passes, and what does not
Let me be concrete. Three proposals I have seen in the last twelve months.
A study of resilience in UK food-manufacturing supply chains after the 2022 energy shock. Passes all three tests. A procurement director who reads it will change how they weight energy exposure against unit cost. It surprises, because the common assumption is that diversification solved the problem — the proposal shows it mostly did not. The evidence is eight firm case studies with matched supplier data, and the sceptical reader has something to grab onto.
A study of "meaningful work" among UK knowledge workers, using existing survey items. Does not pass. The question is important. The design adds no new evidence. It will refine coefficients rather than change a decision.
A study of how FTSE-250 boards govern AI adoption, based on twelve interviews with chairs. Passes, narrowly. Practitioners will read this because they are still working out how to govern AI. The design risk is that twelve is a small sample — but the twelve chairs are doing something no one else has yet studied, so the sample is defensible.
What this means for a doctoral researcher
If you are planning a DCUK doctorate, I would ask you to spend the first three months of your candidacy on the question, not the method. Most candidates rush to method, because method is where research design can be learned quickly. But a well-designed study of an unimportant question is worse than a rough study of an important one — because the first finishes nothing, while the second starts a conversation.
Three practical suggestions.
First, spend a day talking to the practitioners in the field you think you want to study. Not interviewing them. Talking to them. Ask what they would want to know if someone spent a year researching it. The answers will surprise you.
Second, read three of the most-cited papers in your prospective area from the last five years, and ask: who acted differently after reading this? If you cannot find three, that tells you something about the state of the literature.
Third, write your research question as a single sentence before you write your proposal. If it requires two sentences, you probably have two research questions, and the proposal will quietly become a weaker version of both.
The longer game
I have been at this long enough to see the alternative. Research that was not worth doing gets published, contributes to the citation count of scholars who were also not doing research worth doing, and is read by no one outside that circle. It is not malicious. It is the rational response to an incentive system that rewards publication volume.
But we do not have to participate in it.
DCUK researchers work mostly part-time, mostly alongside a professional life that makes them chronically aware of what a real decision looks like. That is a structural advantage — you are unlikely to forget what a practitioner sounds like, because you still are one. Use it. Choose a question your professional self would want answered. Do the research your academic self can defend. Write for the person who has to act on what you find.
The rest of the field can look after itself.