What can researchers learn from entrepreneurs? Part 1
By BWA Research Fellow, Peter Slattery*
This is the first part of a series exploring what researchers and practitioners can learn from entrepreneurs. All views entirely the author’s.
What is research and why do we do it?
Research can be described as “creative and systematic work undertaken to increase the stock of knowledge … and the use of this stock of knowledge to devise new applications”.
The purpose of research is usually to inform and optimise a particular course of action. This could be anything from choosing the best place to eat lunch, to figuring out the best intervention to change a behaviour.
As behaviour change researchers, we do research in different ways and for different reasons in our personal and professional lives.
In our personal lives, we minimise research efforts. For example, we generally won’t create a report with properly cited references to persuade someone who we disagree with on Facebook.
We know that writing that document is unlikely to justify the effort involved. We definitely wouldn’t consider creating a report if, for example, we were unsure of the exact topic being debated, or whether our report would even be read – that would seem irrational (and we, of all people, would never be irrational, at least not predictably so!).
By contrast, in our professional lives, we conduct rigorous and effortful research; detailed and inflexible processes that culminate, sometimes after multiple reviews, rejections and revisions, in a carefully-formatted report or paper that cannot be easily revised or extended.
These efforts are justified by our belief that doing research this way is best for us as researchers and for the wider world as research consumers. But why?
Where do our research norms come from?
Behaviour change research emerged from social science research, which inherited its research norms from older branches of science, such as chemistry and physics.
Our belief that relatively intense, effortful and rigorous research methods are the best way to understand our phenomena of interest (e.g., target audiences and behaviour) and comes from modelling ourselves, to some extent, on what appeared to be the best practices within hard sciences research (e.g., physics, math, and chemistry).
However, there are several important differences between behaviour change and hard science research. In the hard sciences, researchers are often working with universal causal laws and static attributes. By contrast, behavioural science deals with much more inconsistent causal relationships and entities.
Effects are contextually bound by many transitory phenomena. The personal and social realities that we try to map are emergent, chaotic and unpredictable, changing while being examined (and because they are being examined).
One implication of this is that the chaotic and complex nature of social and psychological contexts generally make it much harder to determine ideal research questions and methods.
Of course, the hard sciences have difficult questions to answer and uncertainty to overcome. However, the stability of their contexts, causal laws, interactions and desired outcomes make picking the right research question and method much easier.
Physics questions like ‘what are the spin dynamics of this particle?’ are narrow, stable and answerable in a way questions like ‘how can we make people exercise more often’ never will be. Indeed many social science problems are ‘wicked’ in nature – difficult to understand, define and resolve.
A second implication is that behavioural science findings are also much more context-dependant, local and temporary. For example, an apple falls just as fast from the tree now as it did when Newton calculated the speed of gravity and the effect of gravity can be calculated and accounted for anywhere in the world. By contrast, the best way to change apple eating behaviour is very different from what it was in Newton’s time and probably differs extensively by time and place.
Because of these differences, behaviour change research norms may not always serve their owners well. Often, their rigour and effortfulness is based on a presupposition that the research has attributes such as clear goals, law-like relations and timeless and universally-relevant output; attributes that don’t usually apply in social science research domains.
Given these challenges, I’ve often wondered whether behaviour change researchers can learn something from the field of entrepreneurship.
How do entrepreneurs do research?
A lot of entrepreneurship is applied behaviour change research focused on resolving problems like how to get people to buy a service or product, use a website, or recommend an organisation.
While researching these behaviour change problems, entrepreneurs act with very different incentives to academic researchers (e.g., increasing business performance vs publishing) and optimise for very different goals (e.g., getting actionable insights efficiently vs meeting a journals’ standards).
Studying entrepreneurial research can, therefore, provide academics with useful insights into practices in a behavioural research context where efficiently applied behaviour change research and practice is particularly well rewarded and incentivised.
Just like behavioural researchers often do in their personal lives, entrepreneurs try to conduct research in the most efficient way to achieve the contingent goals.
There are no standards to uphold beyond doing what is most useful; if a research problem appears relatively unimportant or of uncertain future value, the effort invested is appropriately minimal.
If the behaviour change problem is unclear, they reduce uncertainty and risk by exploring parts of it in stages, for example, engaging with a target audience in iterations to explore i) the problem, ii) the solution desired by customers, and iii) the ideal platform and business model for that solution.
Unknowns, updates in understanding and full pivots on initial plans are both expected and encouraged. It is assumed that knowledge, contexts, problems and potential solutions are always evolving, often to the extent that the research could be outdated by the time it’s done.
Just as with problem determination, research is usually done in short iterations. Impact, through the extension of past work, is prioritised over originality. Templates are identified and followed, and if a process or output can be copied for efficiency gains, then it is.
Insights are captured in a flexible and extensible format; for example, in a Wiki or Google document, with the expectation that they will inevitably need to be revised and updated.
The research output is often entirely unstandardised and unrefined to maximise the ratio of useful insights to effort; there is certainly no-one making sure that it is all in Harvard referencing so that a journal will accept it.
How does this apply at BehaviourWorks?
At BehaviourWorks, we have explored ways of achieving academic and business goals more efficiently by using entrepreneurial methods.
One approach that we have had considerable success in is the ‘Hackathon’.
These are short intense sessions where we, as behaviour change experts and researchers, look at a proposed intervention or process through a ‘behavioural lens’ and suggest solutions and improvements, which are captured in a short and simple report.
As an approach, it is well suited for cases where the research question is not yet clear or important enough to justify a more intensive exploration, or deadlines are too tight to permit a more traditional research process.
We have also had considerable success with Rapid Reviews, a type of accelerated review process focused on increasing review efficiency, for example, by reviewing reviews rather than primary studies.
Rapid Reviews enable researchers to provide detailed research within reduced timeframes, but also to meet most academic requirements by being standardised, reproducible and publishable.
Recently we have also developed a comprehensive behaviour change research database.
Research databases can build and share knowledge much more efficiently than more traditional approaches (e.g., sharing bibliographies).
Affordances like shared access, multi-record upload and machine learning-based recommendations make them better for building knowledge. Similarly, the fact that they can store information in formats that can be instantly imported into reference software significantly increases the ease with which researchers can use them in their work.
Most behaviour change researchers default to using rigorous and effortful research as the best hammer to ‘nail’ all behavioural problems.
By contrast, I have argued that researchers should also consider the rapid, innovative and low effort research approaches common in entrepreneurial contexts.
Do you agree? I’d love to hear your thoughts!