Who we are

StakeholderWatch works for large organisations whose performance largely depends on how stakeholders feel about them and act towards them. Insurers, pension funds, banks, retailers, energy, retail and food companies, industry, governments and healthcare providers, they all live by the same fact. When customers, employees, investors and the public trust a firm, they buy from it, work for it and defend it, and when they stop trusting it, they will stop exchanging resources with it. Since 2013 we have measured that relationship for over 80 large organisations.

What we measure

We measure the traditional scores and numbers: reputation, brand, sentiment, satisfaction, trust, buying intention, willingness to work for or invest in the firm. But we do that in a way few providers do. Every survey is designed for one client and fielded for that client alone. Surveys runs every day of the year among real, verified people, so a change can show on the day it happens.

The model behind the numbers

Such scores on their own tell you what happened. To also know why they change, these numbers need embedding in a theoretical model of how people come to feel and act as they do. Ours was developed with researchers at VU Amsterdam. It treats a firm and its stakeholders as a network of exchanges, in which macro trends, events and contact channels change what people believe about a firm. Those beliefs change how they evaluate the firm and the resources it offers. Those evaluations finally predict if stakeholders will engage with the firm, or talk about it. This model is grounded in social exchange theory and network theory and has been in daily use since 2022.

The question everyone asks and nobody answers

Every board, every marketing team and every communications department asks the same question: what is moving our reputation, our brand, our sentiment, our sales? Academic research answers it with care, but one study at a time, for one question, years after the fact. Industry research mostly reports outcome scores and lists the drivers that ought to explain them, but scores alone do not explain why outcomes change.

Causal Intelligence

This is where Causal Intelligence helps and goes beyond outcome scores. A structural equation model, estimated every night on the client's own daily data, measures the causal effect of more than fifty macro, social, event and channel factors on the client's outcomes. These systems only show an effect when it reaches statistical significance. A reputation, brand, marketing or HR team opens the dashboard and reads which forces are pulling or lifting its numbers, how hard, and since when. There is no report to commission and no analysis to wait for. The answer is on the screen, and it is updated every day.

What stands behind it

Thirteen years of daily fieldwork across nine sectors. A behavioural model built with VU Amsterdam. A structural equation model that estimates causal effects. A research programme on reputation dynamics during organisational crises and the founder's upcoming doctoral dissertation scheduled for defence at VU Amsterdam in 2027. And a rule that governs all of it: no synthetic respondents, no imputed answers, and a fieldwork budget spent on finding real people to collect data, because what goes into a model is what comes out of it.

The team

StakeholderWatch works from one office in Amsterdam. The team is small by design and built around the three things the product needs. The IT engineers who run the daily pipeline, the estimation and the dashboards, and who keep the data the client's own. Social scientists and econometricians, the people who design the surveys, build and test the models, and read the results before a client does. The others run the firm and its client relationships, so that every client has one person who knows their data and their questions. The founder leads the research programme with VU Amsterdam and is a doctoral candidate there.

The combination

Other providers offer parts of this. To our knowledge, StakeholderWatch is the only firm that offers academic-grade methods and daily estimation and single-client fieldwork and verified human respondents and a theory-based behavioural model and a causal engine for fifty or more factors and full access to the raw data as one product.