One of the joys of working in-house is seeing the cumulative buildup of your efforts over time. You’re able to invest into something bigger than individual projects; you can build a body of work.

I’m reviewing past studies and in the process I’m creating a list of the topics and research questions I’ve worked on to see how they have evolved in the last couple of years. By itself, the list mainly shows how the organisation’s focus has gradually shifted. Interesting, but obvious. What’s not apparent is how these projects have come about: which questions needed more ‘shaping’ than others? Which ones are verbatim requests, and which ones are unrecognisable from their initial ask? The invisible stories of what happened during the scoping sessions.

This is one of the most strategic parts of the work as a researcher in industry: manoeuvring problems into useful levels of granularity.

In other words: collaboratively defining the lens with which you’ll investigate the problem. In a low maturity research environment, this typically means that as a researcher you’re trying to push research questions towards a bigger picture problem space in order to gain higher-level insights (a common struggle). In a high maturity research environment it can mean the reverse: ensuring a broad question is sufficiently actionable by trying to pull it into specificity (a bit more uncommon).

These granularity levels roughly correspond to Stewart Brand’s pace layers: while the big picture insights typically have a long lifespan – as nature and culture change slowly – the product/feature/concept-specific insights expire at a much faster rate – as they’re more directly affected by things like competition and innovation.

At the extreme ends of the scale we have:

None of this is as absolute as it sounds, because you can incorporate a variety of questions into one study. You’re simply manoeuvring parts up or down the granularity layers to create a more useful whole. This calibration is fundamental to collecting a range of data with both short-term and long-term value. Rather than a single shift from one granularity layer to another, reality is more about actively moulding the overall shape of the research: creating a blob that sits optimally across granularity layers. Here are two examples from real studies, with questions from my discussion guides mapped to the granularity layers they address:

In Study 1 interview questions focused heavily on the feature/product/concept layer – and as a result the data collected in the study has had a short lifespan: while it was useful at the time, it is now rarely looked at and it is doubtful whether anyone would notice if the data and insights would vanish. Study 2’s questions were distributed more equally across the granularity layers, with a bigger focus on the context/needs layer and the mental models layer. The data collected in this study is proving to be valuable over and over again: the higher-level insights remain relevant and useful across multiple products.

Moulding and shaping

Let’s look at a practical example:

A few years ago, I was asked to investigate why some customers paid late – or had not paid at all. The focus of this request was on the potential need to change the payment process within the product, so the request sat very much at the high end of granularity. But why stay there?

After a few productive scoping conversations with the stakeholders we were all in agreement that this was an opportunity to learn more about customers’ payment workflows, how their organisation’s procedures around reimbursement worked, and how they viewed their own role in the payment process. We moulded the problem space into something beyond tactical.

I investigated the problem through a variety of data sources: a large volume of support tickets, a detailed case study on a specific late paying customer, operational data, a series of in-depth interviews, and previously gathered survey responses. Collectively, the analysis painted a vivid picture of a complex landscape of problems at a macro, meso, and micro level: from global currency challenges, country- and organisation-specific policies and processes, payment delegation, and mismatched mental models. To this day it remains one of my favourite examples of the depth of insights even a seemingly tactical problem can reveal – with a bit of proactive shaping.

Reshaping so drastically is not always within reach, but in my experience it is always possible to include some bigger picture workflow and context aspects. I’m currently benefiting from the data I’ve collected through the handful of contextual introductory questions I have asked in almost all my interviews in the last few years. Over time this has become a highly reusable and insightful source of data. The small but consistent efforts add up to something strategically meaningful.

Consciously moulding and shaping research so that it sits across granularity layers (yes, even – especially – purely evaluative research) is valuable for many reasons. Let’s assume you’re working towards incorporating more of the bigger picture. Your moulding efforts can have:

• Immediate impact: the study itself becomes richer by digging into the underlying reasons for people’s feature, product, or concept usage. This is where the benefits of bigger picture research become most tangible to stakeholders; I’ve seen many readjust how they think about their problem spaces in future conversations.
• Broader impact: the collected data has broader relevance so can be shared with different audiences beyond the original research requester. This data is the main source of my daily Slack feed of verbatim quotes and the quarterly sessions where I broadcast recent research clips to a wider internal audience.
• Long-term impact: by continuously incorporating at least some bigger picture questions, you’re building knowledge rather than just answering requests. The longer lifespan of the collected data and insights makes it ideal for reuse in, for example, meta analyses. This not only benefits the organisation’s efficiency (both in terms of money and time), it is also ethically preferable: we’re optimising the data research participants give us. Thinking ahead about what data might prove valuable in the future is a critical part of this long-term impact.

Upfront influence

Oddly enough, this upfront work of moulding the research – or rather: reshaping the organisation’s questions – across meaningful granularity layers is not what most of the discourse on strategic influence is centred around. Instead much of the focus is on the communication of findings, the structured formatting of needs, maintaining research repositories, the automation of curation, etc. All stuff that happens towards the end of research. All stuff that is only worth doing if you’ve collected valuable data; if (at least some of) your data has a longer lifespan.

The biggest moment of influence takes place at the very beginning: during the scoping of the work, when the granularity levels are calibrated. After all, this will determine everything that follows:

Organisational goals ➔ project’s objective and scope ➔ research questions ➔ research method(s) ➔ type(s) of data you collect ➔ the analysis and synthesis method(s) ➔ depth, reusability, and lifespan of collected data and insights.

It’s less about whether your research is predominantly evaluative or generative, and more about the continuous effort of moulding and shaping the organisation’s questions across granularity layers. You’re building a knowledge foundation, with use far beyond the scope of the initial research request. A foundation from which you, and your organisation, will be reaping the rewards thanks to your past self’s bigger picture thinking.