Consumer Insights Research

What is consumer insights research?

Consumer insights research is the practice of understanding the motivations, associations and attitudes behind what people do. With the reason in hand, a team can act on it instead of guessing from the behavior. Google's AI Overview for the term draws the usual line, citing Qualtrics, Market Logic Software and Trustpilot among others: market research answers the “what,” and consumer insights answer the “why.”

The standard toolkit for the why is surveys, interviews, focus groups, social listening and observation. Every one of those except observation depends on people describing their own reactions in language, and language is where a reaction gets edited. Ask someone whether a campaign feels trustworthy and they'll consider what a sensible person would say before they answer. Our research page has a name for that state: people answer “in personal PR mode.”

A visual perception study changes the instrument. Participants see a board of images and place green dots on the ones they associate with the prompt and red dots on the ones they don't. Constellations aggregates those dots into an Audience Perception Map (an APM): a heatmap showing which images resonate, which draw resistance, and which are contested, with a percentage on every image. This is audience perception mapping: an audience responds to images instead of describing them, and the pattern shows where the group agrees, where it splits, and why. Nobody is asked to explain anything. Every question teaches people what kind of answer it wants, and a picture gives them nothing to aim at.

The method sits inside a longer research process. On a brand project we start by learning the company, its market, its audience and what matters to its employees, then assemble the images worth putting in front of people, usually 24 to 48 of them across a few boards. The direction comes out of what the audience converged on. Nobody brings two finished options and asks people to vote. The method itself, board by board, is in Concept Testing With Images.

Consumer insights research vs. market research

Market research measures what's happening in a market, and consumer insights research explains why the people in it behave the way they do. A share report can tell a beverage brand that its campaign imagery isn't moving anyone. It can't tell the brand which images make people want the drink.

Our Coca-Cola study can. We asked 467 people one question about 24 Coca-Cola images: which of these make you want one right now? They placed 4,683 dots, and even the illustrated Santa took 78.54% red. Every image that cleared the study's 60% threshold showed physical evidence of cold, like ice, frost or condensation, and every image that failed put the brand where the drink should have been. That's a why, and it comes with a number a creative team can put on the table when the direction gets questioned.

The two kinds of research need each other. A share number without a reason sends a team guessing, and a reason without a number can't tell you how much it matters. A perception map works for both sides of that table. The creative team reads it as a picture, and the analysts check it as a set of percentages.

Why surveys and focus groups miss the why

Both ask people to turn a reaction into words, and the turning is where the honesty goes. Harris described the mechanism plainly: “If I were looking at a certain type of language being presented to me, I probably would feel compelled to answer in a tone that's related to that language. Here, I felt I can just basically answer from the heart.”

Three things shape an answer before it reaches you:

• The question. Wording signals what a reasonable answer looks like, and people answer toward it.
• The room. In a focus group, the first confident answer pulls everyone after it. We've written about alternatives to focus groups, and the same dynamic runs every internal review. Design by Committee is about that room.
• The history. People answer a direct question at work through everything that's happened before it. Express Newark's staff had been through, in Harris's words, “a whole bunch of consultants,” and a direct question got an answer filtered through that history: I know what they think, I know where this is going.

Figure 1 · The work answer and the dot

Words make people careful. Images are faster than careful, and the reaction lands before anyone has had time to polish it. That's the gap visual research works in, and it's also why participants finish quickly. In our college study, prospective students typically finished in a minute or two, on iPads, while they waited in line at an admissions event.

When your team and your audience see the brand differently

You find out exactly where the brand's meaning is unsettled. Both views are data. The trouble starts when a company treats its own people's view as the audience's view without checking.

The check is to run the same boards past both groups as separate segments, then read where they agree and where they split. Where both cluster on the same images, that material can carry the brand, because it works inside the building and outside it. Where they split, the team gets a choice to make on purpose: lean toward the group that carries the buying decision, or hold the core steady and let individual campaigns lean. We wrote up that rule in When the Segments Disagree, along with the practical habit behind it: if you're even slightly curious whether a group sees things differently, segment it. Segments blend back together in a click. A pooled result can't be split apart afterward, and an average of two groups pulling in opposite directions looks like consensus and hides the fork.

Figure 2 · The same boards, read as two segments

Sometimes the inside view is the right stand-in, and the job is to use it deliberately. On the InEight rebrand, covered in the next section, the target buyers were capital construction executives, who are hard to survey at scale. So the maps went to the company's own staff, many of them years into talking with those executives. A proxy audience works when it knows the real one, and it helps that employees are the people who have to live with the brand afterward.

How do brand teams use consumer insights research?

Brand teams use it at the three moments when a brand decision is about to get expensive: before a rebrand picks its direction, when an audit has to say whether the brand is being read the way it was meant, and before packaging commits to a print run. Each one sits on the edge of what we call the Approval Cliff. Approval means nobody blocked the work. Alignment means the team believes in it. A why the audience supplied is what closes the distance between the two, because it gives the room something other than taste to agree on.

Rebrands. InEight, a global capital construction software company, rebranded at its ten-year mark, with a large group of stakeholders who each wanted something different from the new identity. The team ran the usual research first: stakeholder surveys, internal interviews, customer insight and competitive audits. Then, instead of synthesizing all of it into one recommended direction, they built four Audience Perception Maps of tonal directions and sent them to the whole organization. The result settled on two tones, which the team named Authority & Control and Wisdom & Counsel. “The trust that we were going in the right direction was built in through that process,” said CMO Melissa Esbenshade. Full disclosure: the creative firm on that project was Dark Square, the firm our founder Joseph Fioramonti leads. The case is at InEight Rebrand: Evidence-Backed Identity for Enterprise Software, and if the rebrand you're facing is already stuck, Why Rebrands Fail covers the pattern under the famous ones.

Audits. A standard brand audit inventories everything the brand has produced and grades it against the style guide. That's useful, and it measures the work against a document. The measured version adds one step: sample the material so it represents the full spectrum of what the brand puts into the world, then test that sample with the people closest to the brand and, where it matters, with the audience. Consistency with the guide is a filing question. Whether the brand is being read the way you intended is the audit question. See What a Brand Audit Should Actually Measure.

Packaging. A pack gets reviewed at desk distance in good light and bought in a second from four feet away, next to a wall of competitors. Testing finished comps is too late to be cheap, so the useful test puts elements in front of buyers before the comps exist: category cues, material and finish, how the product itself is shown, typographic register, and color against the category. Existing buyers, new buyers and the retail buyer run as separate segments, because they read a shelf differently. The Coca-Cola result above is the packaging lesson in miniature, since how the product is shown decided more of the response than the brand's own icons did. Testing Shelf Perception Before the Print Run has the full approach.

How do universities use consumer insights research?

Mostly for two jobs: learning what prospective students respond to, and finding out where a staff already agrees. A university is a hard approval room, with large committees, long memories and an audience already on campus, so it's a good test of whether the method holds up.

Admissions. A liberal arts college tested campus imagery with high school juniors and seniors who were already in contact with admissions. The first round mixed academic and student-life scenes on the same maps, and the energetic scenes swallowed everything. Not one academic image drew a response on either map. The college learned that teenagers prefer concerts to classrooms, which it knew already. The second round gave each map one theme, and the same audience suddenly had clear preferences about which academic environments felt welcoming. The test structure decided what the test could see. Both rounds of maps are in The College Use Case.

Alignment. When Harris, co-director Nick Kline and deputy director Natasha Dyer took over Express Newark, Rutgers University–Newark's center for art, design, digital storytelling and music, the staff were in what Harris called “let's-wait-and-see” mode. Working with consultant Wendy Watson-Hallowell, they used a visual study so the organization's values could surface without being steered. On one long-standing exhibitions issue, the vast majority of staff turned out to be on the same side, and that changed who could raise it. “It transitions us from ‘the leadership needs to change this’ to ‘as a community, as an organization, this is not what we believe in,’” Harris said. Watson-Hallowell watched it in the room: “I saw decisions being made more easily, instead of the personal stuff.” Read how Express Newark built alignment during a year of transition.

Universities buy through department, faculty and course licenses. The structure is on research and institutional pricing.

How public agencies use visual research before positions harden

Agencies use it to see where residents stand on an issue before the wording of a proposal gives everyone a side to defend. Words carry a lot of freight in public work, even two that sound interchangeable. We asked 26 people to sort 36 images into “health” or “wellness” with no definitions given. They drew one consistent line: health is what the body does, and wellness is what it receives. A sprinter in pace graphics took 11 of 11 dots as health. The massage took 89.5% as wellness. The study is small and the report says so, resting its weight on the patterns that repeat across many images. For an agency naming a program, that's the difference between promising people something to do and promising them something to receive. The full map is in how a U.S. public distinguishes health from wellness.

Public campaigns tend to miss in a predictable way. The message gets written at a distance from the people receiving it, the approval group is large and unlike the audience, and the money commits early, often at the grant proposal stage. A perception test at the front produces something a funder or board can read: “we put a range of material to the people this serves, here is where they converged, and here is what we built on it.” Testing a Public Campaign Before the Money Is Committed walks through the sequence.

What a finished consumer insights study looks like

A finished study is a set of maps with a percentage on every image, plus a written read of the pattern. Ours are published, so you can judge the output before you talk to anyone. These are on the site now:

Read across them and a pattern shows up that none of the studies was designed to find: audiences sort images by what the image is doing. The drink in motion, the symbol doing a civic job, the body doing or receiving. Style and palette explain much less of the result than function does. For a designer, that's a useful thing to know before the review, because it's an argument about what the work does rather than whose taste it suits. Five Studies, One Pattern: Audiences Sort on What an Image Is Doing makes that case study by study, with its limits stated, since several of these samples are small and a 22-person study is a directional read.

How many people does a consumer insights study need?

About 30 responses for a pattern to start holding, and 300 to 500 before a brand decision rests on it. Our research page puts the point where patterns typically start to hold at around 30 responses, and below that a result is directional. Our pages on testing brand direction before launch and on alternatives to focus groups put a brand decision at 300 to 500 people from the target audience. The Coca-Cola study, at 467, sits inside that range.

Figure 3 · How many people a study needs

Results come back in under 48 hours once the test is live. The slow part is everything around the fielding: learning the organization and the question, and choosing the images. The test can go to your own audience by link, or to a recruited audience that matches them.

Two habits make a sample worth more. Segment anything you're curious about, because segments can always be blended back together. And structure the boards around the comparison you need, because the college study showed that a loud category on the same map as a subdued one measures volume.

How to buy a study from Constellations

Brand and creative teams buy a done-for-you study: we design the boards from your material, field the study, run the analysis and walk you through the result. Teams that want to run studies themselves get certified.

Every company study starts with a 30-minute scoping call where we frame the question, settle who the participants are and set the timeline. After the call you get a payment link or an invoice, and work starts on payment. Studies at $20,000 and up can split in two: half to start, half at report delivery. The full comparison is on pricing, and you can book a scoping call from there.

Universities and public agencies buy on institutional terms: department licenses from $1,000 a month, and a single Engagement Study for an agency decision at $9,500, quoted for procurement or a grant application and invoiced on terms aligned to the fiscal year. The detail is on the research pricing page, and you can talk with us to scope one.

A good place to start is the direction you expect to be challenged on, or the decision your team keeps getting work answers about. Write down the question, and look at whether it's a question about how a group of people feels. If it is, it can be put on a board, and the answer can go into the room with you.