Are You Designing Your Research Backwards? Here's Why You Should
We've all been there. Sitting in a presentation where someone unveils a beautifully produced market research report, complete with colourful charts, complex data tables and elegant transitions. Twenty minutes in, you find yourself asking the question that haunts so many research presentations: "So what?"
If you're nodding in agreement, you're not alone. And the solution might be suspiciously simple: design your research backwards.
The concept isn't new. Harvard Business Review published an article by Alan R. Andreasen in 1985 titled "Backward Market Research" that introduced this approach. Yet nearly 40 years later, it remains surprisingly underutilised in our industry.
Too often, research follows a linear path: take the client brief, translate it into a questionnaire or discussion guide, collect the data, analyse the findings – and then figure out what actions to take.
All seems logical, until it doesn't. By the time you realise the data doesn't answer any of the questions you needed to address, it's too late. The budget is spent, the respondents are gone, and your stakeholders are waiting for insights that simply aren't there.
I witnessed this first-hand when working with a retail client who commissioned an expensive customer satisfaction survey. The questionnaire covered every conceivable aspect of the shopping experience – from store layout to staff friendliness to product range. When the results came in, they showed high satisfaction across most metrics, yet sales continued to decline. Why? Because the research never asked about the one thing that was driving customers away: the emotional tipping points that influenced their shopping behaviour. The research answered questions that weren't relevant to the business challenge.
The backwards approach flips this research process on its head: start with what you want the data to tell you. Some proponents advise going as far as sketching out the final report, complete with dummy tables, charts and key findings before you've collected a single data point. While this might seem extreme, it forces clarity about the decisions that need to be made with the research.
Realistically, few of us have time to create entire mock reports. But the principle remains sound: begin with the end in mind. Before writing a single question, ask yourself and your stakeholders: "What decisions will we make with the data from this question?" If you can't answer that, you probably shouldn't be asking it. It's just taking up valuable real estate when you could either have a shorter survey (respondents will thank you) or ask a better question.
The challenge often lies in the overwhelming blank page when beginning questionnaire design. We instinctively bookend it with the familiar: screening questions at the start, demographics at the end. These are comfortable and safe.
But then comes the laundry list of "nice to have" questions from various stakeholders, and we end up simply giving them structure and logical flow rather than ruthlessly questioning their purpose. We lose focus on the end goal of each question. It's like building a plane whilst flying it.
I recall a financial services client who wanted to include questions about every product feature in their portfolio. When I asked what they would do with the data on their mortgage protection insurance options, they paused and admitted, "Well, nothing really. We can't change those features for regulatory reasons." That question, and twelve others like it, we consequently deleted from the draft.
The backwards approach doesn't just improve design, it focuses your analysis later. Data can be overwhelming, but when you know exactly what business questions you're trying to answer, you know where to direct your analytical attention.
The biggest barrier to adopting the backwards research approach isn't technical, it's cultural. It requires more upfront investment in strategic thinking and planning, which often deters busy teams who just want to get the research started. The conventional approach feels more efficient in the short term, even if it produces less useful outcomes.
3 things that helped me focus during the design process:
If you work visually, sketch out what you expect the key findings might look like, but leave enough wiggle room to change direction when the data surprises you, as it inevitably will.
Work collaboratively with your client or stakeholders early in the process. It can help to share something visual, even if it's not perfect. A "no, that's not quite right" is a better starting point than a blank page.
Rather than asking "what does this question add?", ask "what decisions are we unable to make if this data is missing?" This subtle shift focuses everyone on utility rather than interesting-but-useless information gathering.
Backwards research isn't a panacea. There are legitimate concerns to address.
Don't lock yourself in too early and become blind to unexpected insights or emerging trends. Deliberately build in a "fuzzy front end" to your research to allow this wiggle room.
Be wary of how backwards research can be weaponised to confirm biases. In corporate environments where leaders sometimes need data to justify decisions they've already made, starting with the desired outcome creates the risk of designing research that only reinforces existing beliefs rather than provides true insight.
The value of research isn't measured by the volume of data it produces, but by the clarity it brings and the decisions it enables. When done right, backwards research doesn't just produce better questionnaires, it produces better business outcomes.
Next time you're staring at a blank questionnaire template or wondering why your last research project didn't quite deliver the insights you needed, consider working backwards. Start with the decisions you need to make, then determine what data would support those decisions, and only then design your questions.
Your respondents will thank you for shorter, more focused questionnaires. Your stakeholders will thank you for more actionable insights. And you'll never again have to sit through a presentation wondering "so what?"
