Research is retreating behind its desk, at the same time as the insights function is worrying about whether it is trusted. I keep wondering whether those things are more connected than they look.
There is an insight buried in the most recent Greenbook GRIT Insights Practice Report that has been truly stuck in my head. The kind of stuck that I’m even thinking about it while I make tea/ or feed the dogs kind of stuck! Underneath all the noise about AI, the report picks up a quieter anxiety among insight professionals. It isn’t really will a machine replace me? It is closer to: does my organisation actually believe in what I do?
It’s a question most researchers will recognise, because we’ve heard versions of it from very good people over the years. Not generally because they thought the research itself was poor. Quite often it comes after a very good project, a strong debrief, plenty of engagement in the room, and then the slightly bewildering realisation a few weeks later that very little actually changed.
BUT… what has caught my attention in GRIT was sensing an anxiety sitting alongside another set of findings entirely.
Research appears to be moving behind its desk
In another part of the report, GRIT looks at the methods brand-side researchers are using. Interviewer-administered in-person surveys fell 35 percentage points this year. In-person qualitative research fell 15 points. Ethnography fell 15 too.
The methods gaining ground look quite different. Synthetic data is up 18 points, crowdsourcing is up 14, and online communities and bulletin boards are growing. Greenbook describes the broader pattern as a move towards approaches that can be run with minimal external dependency, while the methods losing ground tend to require physical presence or more external infrastructure. A lot more research can now happen without anybody ever needing to leave their desk.
There are obvious reasons for this, and we all understand them. Cost matters. Speed matters. Technology has made things possible that weren’t possible before. I also want to be careful not to make an argument I’m not qualified to make. I don’t have a position on the validity of synthetic data here, and this isn’t a nostalgic defence of ethnography or an argument that research only counts if somebody has physically gone somewhere to conduct it.
What interests me much more is what changes for the researcher when less of the work involves direct contact with the actual person whose behaviour we are trying to understand.
There really just is something about having been there
I’ve spent a lot of my career around qualitative research, and there is a particular authority that comes from having witnessed something yourself. Anyone who has done enough fieldwork will know the kind of moment I mean. You’ve spent time with someone in their home, or stood beside them in a shop, or watched them try to navigate a service that everybody inside the business believes is perfectly straightforward. Then you get back into a boardroom where somebody has a very polished explanation for why customers behave the way they do. And you know it isn’t quite true, because you watched the thing happen.
That matters more than I think we acknowledge. It doesn’t make your interpretation right, and it doesn’t remove bias. But it gives you something to draw on when the conversation gets uncomfortable. You aren’t only carrying a finding into the room. You have an experience of the thing itself.
That is what I mean by witness.
And I wonder whether witness has been doing more work for the authority of researchers than we realised.
Evidence is only part of what gives you standing
Research conducted remotely can produce excellent evidence. So can analytics, communities, social data, AI-enabled synthesis and plenty of other methods that don’t require a researcher to be physically close to a customer. What changes is the source of your standing in the room. Standing is what helps you hold onto an interpretation when somebody doesn’t like it. It is the confidence, and perhaps also the permission, to say No, I don’t think that is what is happening here when the easier thing would be to add another caveat and move on.
Direct exposure to people gives researchers one particular form of that standing because there is an asymmetry in the room: you have seen something the other people haven’t. As more evidence is generated through tools that everybody can access, that asymmetry starts to narrow. The researcher no longer has the same monopoly on access to the information. Someone else can use the dashboard, interrogate the data or run a version of the same tool. Which means the authority has to come from somewhere else.
I don’t think these are necessarily two separate stories
The GRIT report doesn’t draw a connection between declining in-person methods and the profession’s anxiety about being believed, and neither can I. There are far too many other things changing in organisations at the same time to pretend the data proves that.
But I personally find the coincidence difficult to ignore. At the same moment that insight professionals are asking whether their organisations believe in what they do, they are also doing less of the work that historically put them into direct contact with the people they were there to understand. I also don’t think that means the answer is “do more ethnography.” It does make me wonder whether we have underestimated what proximity contributed to the researcher’s influence beyond the information it generated.
There is a difference between presenting evidence that says something happened and knowing, from direct experience, that you watched it happen. If we remove more of that direct experience from the job, we probably need to be much more deliberate about where the researcher’s standing comes from instead.
Fieldwork was never enough. I said what I said.
There is an important counterpoint here because I’ve also seen plenty of extraordinarily rich fieldwork change absolutely nothing. You can spend weeks close to people, come back with an understanding of the problem that nobody else in the organisation has, and still lose all of that advantage in the debrief by refusing to take a view. We know the language.
“There are a few possible ways to interpret this.”
“These might be areas to explore.”
“One potential consideration could be…”
Everyone leaves the room with the evidence intact and no idea what it means. So I don’t want to romanticise proximity. Fieldwork might give you more to stand on, but it does not make the judgement for you.
In fact, this is the part of the conversation that interests me most, because I suspect it is where the future value of the researcher sits. The methods will keep changing. More research will happen without physical fieldwork. AI will make more forms of evidence available to more people, and increasingly the researcher will not be the only person in the organisation capable of finding information about customers. Judgement is harder to hand out.
The room was probably never that interested in our methodology anyway
I’ve banged on about this specific point for ages, and perhaps so loudly that folks stopped paying it any attention. So here I am again. Saying it again. With my whole chest: researchers care deeply about methodological quality because we should. It is our job to know where evidence came from, what it can support and where it starts to break. I remain less convinced than ever that this is what creates influence once we walk into a decision-making room. In my experience, the room is persuaded when someone knows something useful that it does not know, can explain why it matters, and is prepared to hold onto that view when the implications become inconvenient.
Historically, being close to customers gave researchers one powerful route into that position. If that route is becoming less common, then the other parts of the job matter more.
Question becomes about getting upstream of the brief and making sure we are solving the right problem in the first place.
Pattern is the judgement to know what deserves attention when there is more evidence than anyone can reasonably absorb.
And Meaning is what happens next: being prepared to say what you think that evidence means for the decision, particularly when the answer isn’t the one the room was hoping for.
Which brings me back to the anxiety in the GRIT report. If researchers want to be believed, I’m not sure the answer is to produce more evidence proving that they should be. I think we need to look much more closely at what gives a person standing in the room in the first place.
A question to leave you with
Think about the last piece of work where you had to persuade somebody of something they didn’t particularly want to hear. How much of your confidence came from the evidence in front of you, and how much came from having been close enough to the problem to know that you believed it yourself?
And perhaps more practically: when did you last watch a customer actually do the thing you are advising the business about?
If you want to know where your own influence holds and where it starts to slip, that’s what we built the HI Score to look at.
The HI Score is a short diagnostic for people working in insight and research. It looks at how Question, Pattern and Meaning show up in the way you work, and what happens to them once the work is under pressure, challenged by the room, or pulled into a real organisational decision.
Source: 2026 GRIT Insights Practice Report, Greenbook (16th edition, fieldwork Q1 2026). Brand-side researchers, n = 176.

