When impact is not what we imagined

In the first season of The Impact Corner, we talked about our understanding of impact, the various ways to achieve it, co-creation, impact culture and the role of institutions in facilitating it. A key idea has emerged throughout these Corners: impact is neither linear nor immediate, nor fully predictable.

In this new series of articles, we want to take this idea further and address a question that often falls outside the usual narrative: what happens when the outcome of our research is different to what we expected? What if an innovation, an evidence base or an application has unintended —or even negative— consequences?

The research system tends to prioritise what fits the 'success' narrative. In this corner, however, we want to highlight the unexpected and accidental impact. We also want to recognise negative impact as an inherent part of scientific work in real and complex contexts. For example, CFCs (chlorofluorocarbons) were discovered in the 1930s because of their 'safety', they solved the problem arising from the use of toxic and flammable gases in industrial refrigeration processes until then. However, a few decades later, a study published in Nature1  revealed that CFCs were responsible for depleting the ozone layer due to their persistence in the system. CFCs were officially forbidden in 1987 by the Montreal Protocol. 

Impact pathways can take many different forms, some of which are unexpected. This idea is not new. Back in 1936, the sociologist Robert K. Merton was already thinking about the unplanned results of planned social action.

One of the key insights we can take from this, which remains fully relevant today, is that an unintended outcome is not necessarily negative, although it can be. In complex contexts, unexpected outcomes can arise from structural dynamics or the serendipity of action itself, as discussed in more recent work on anticipation and theory of change (Andrews, 2024; Sørensen et al., 2025).

Thus, the unintended consequences of a research project or innovation can range from unexpected benefits to drawbacks that emerge during implementation, or even negative outcomes that produce the opposite effect to that intended. Working towards impact necessarily involves accepting this uncertainty.

A recurring issue is the risk associated with combining pressure to deliver results with a lack of opportunity to reflect on what happens after a project ends. If we do not actively ask how the knowledge generated will be used, in which contexts and with which limitations, a wide range of medium- and long-term consequences may emerge.

This is not necessarily due to bad intentions, but rather because the complexity of the real world often exceeds our ability to predict, especially when there is no explicit follow-up on the impact pathway (Sørensen et al., 2025). Incorporating this perspective is part of a more realistic and responsible approach to impact.

In this context, the specialised literature has started to use the term grimpact, introduced by Derrick et al. (2018), to describe any negative consequences of research. Through various examples, these works remind us that research is not neutral and that it can reduce social well-being under certain circumstances, whether intentionally or not.

Talking about grimpact can be uncomfortable, but it is necessary. Paradoxically, discovering that a result is not beneficial to society can itself be a valuable form of impact. Making errors or unintended effects visible helps to prevent repetition and contributes to collective learning.

Admitting mistakes and unintended consequences is part of the scientific method. Being transparent about this is also an act of responsibility towards society. In some cases, addressing a grimpact properly can indirectly generate positive impacts through learning, mitigation or follow-up research. 

In previous Corners, we emphasised the importance of engaging with stakeholders to promote the use and mobilisation of knowledge. However, these interactions are not only key to generating a positive impact, but they are also essential for identifying risks and unintended consequences.

Often, the first signs of unexpected impact emerge outside of academia. Local stakeholders, managers, administrators or affected communities are often the first to identify tensions, side effects or unforeseen consequences. Maintaining continuous, open and trusting relationships with these stakeholders helps us to anticipate problems, facilitate dialogue and gain a better understanding of the variety of effects that research may generate. 

Promoting a healthy impact culture means being able to explain what worked and what did not go as expected, as well as what we learned from the process. It means having a culture that does not limit itself to showcasing positive results but rather integrates critical reflection and collective learning as part of the process.

If we want our research to be truly connected to society, we must accept that this connection involves responsibilities, dilemmas and sometimes uncomfortable consequences. Rather than trying to avoid any negative impact, which is almost impossible, perhaps the challenge lies in learning to anticipate them better, recognise them when they appear, and manage them responsibly.

Ultimately, discussing unintended impacts means asking difficult but necessary questions: Who benefits from an impact? Who may be excluded? Which voices are we hearing … and which are we not? Incorporating these questions into our research is a way of viewing impact as an ongoing commitment to understanding the complexity of the world we aim to transform.

• Merton, Robert K. (1936) “The Unanticipated Consequences of Purposive Social Action.” American Sociological Review 1, no. 6 (1936): 894–904. https://doi.org/10.2307/2084615
• Sørensen, OK et al. (2025) “And then a miracle occurs—a review of theory of change models for societal impact of research”, Research Evaluation, Volume 34, rvaf057, https://doi.org/10.1093/reseval/rvaf057
• Derrick, G. et al. (2018). Towards characterizing negative impact: Introducing Grimpact. 1199-1213. Paper presented at 23rd International Conference on Science and Technology Indicators 2018, Leiden, Netherlands.
• Andrews, Clinton. (2024). Better Anticipating Unintended Consequences. IEEE Transactions on Technology and Society. pp. 1-1 https://doi.org/10.1109/TTS.2024.3403412