How to Measure Brand Consideration: A Guide (2026)

Brand consideration is the share of people in a defined market who would put your brand on a shortlist for a specific need, and you measure it by asking that question directly, then supporting it with behavior and choice-model evidence. The hard part is not collecting the number. It is defining the metric tightly enough that the result survives a skeptical finance team.

Most brand tracking gets how to measure brand consideration wrong in one of two directions. Either it reports awareness and calls it consideration, or it publishes a single percentage with no benchmark and no way to act on it. Both produce a slide nobody can use.

This guide walks through the whole job: defining the measure, choosing between aided and unaided prompts, sizing the sample, calculating the result, reading it by segment, and connecting it to revenue. Most teams can run a first study in three to four weeks, but only if the decisions in step two happen before anyone builds a questionnaire.

What You Need Before You Start Measuring

Before touching a survey platform, write down six things. Teams that skip this step end up with data that measures something other than the decision they actually face.

A defined target market. Consideration has no meaning without a denominator. Quarterly buyers of running shoes in the United States is a market. Everyone is not.

A decision context. Which occasion are you measuring? The next purchase, a purchase in the next 12 months, or replacing a specific current product. Each answer produces a different number, and they are not comparable.

A research objective. Establish an overall level, compare against named competitors, diagnose why consideration is low, track change over time, or predict purchase. Pick one primary. Studies that try to do all five usually report none of them cleanly.

A sample plan. Decide the population, the target sample size, the quotas, and the fieldwork dates before you see a quote. Reeling sample back afterward is how budgets disappear.

Comparison brands. Consideration is relative by nature. A brand considered by 22 percent of the market might lead the category or trail badly, and the list you show respondents decides which. Include the brands buyers realistically shortlist, not the ones you wish they would.

Tooling and a reporting period. A survey platform such as Qualtrics or a specialist panel provider, an analysis tool for weighting and cross-tabs, and a fixed cadence for reporting. Quarterly tracking with the same question wording beats monthly reporting with a question that changes every wave.

Step-by-Step: How to Measure Brand Consideration

1. How to Measure Brand Consideration: Define the Metric

Write the definition as a single sentence with a numerator, a denominator, and a time frame. For example: the percentage of people in the target market who say they would actively consider buying from the brand the next time they shop for this category in the next 12 months.

That sentence settles four arguments in advance. It excludes passive liking, excludes people outside the market, fixes the horizon at 12 months, and ties consideration to a shopping occasion rather than a general attitude. Specify the journey stage too, since B2B buyers in an active evaluation behave nothing like someone hearing the name for the first time.

How to tell it worked: two people reading your sentence independently write the same questionnaire item from it. If they cannot, the definition is still too soft.

2. Set the Research Objective and Benchmark

Decide what decision the number will feed. If the answer is a budget conversation, a before-and-after benchmark is the priority. If the answer is a creative or positioning problem, a competitor comparison with diagnostic follow-up questions matters more.

Then set the threshold before fieldwork. A three-point decline or a ten-point gap behind the category leader is worth acting on; a one-point wobble is noise. Writing that number down in advance is what stops the meeting where someone declares victory over a 0.4 point gain.

Sources of benchmark: your own prior wave, a competitor set measured with the identical instrument, or published category norms from a panel provider. Your own trend is the most reliable of the three because it uses a stable sample frame and stable wording.

3. Select the Right Measures

No single measure answers the question. Use one direct measure and one supporting measure, and add a modeled one when the decision carries real money.

Aided consideration. You name the brand and ask whether respondents would consider it. Reliable, comparable across waves, and flattering to unknown brands because you did the work for them.

Unaided consideration. You ask which brands come to mind first and code the open answers. Far more honest, and consistently lower. The gap between the two numbers is itself a finding, since it measures how much of your recognition is doing real work unaided.

Prompted shortlist and salience measures. Ask respondents to pick the brands they would keep on a short list, or to allocate 100 points of preference across a brand set. Good at ranking relative strength; heavier on respondents.

Behavioral proxies. Branded search volume, direct traffic, product-page visits that arrive without a click-through, comparison-shopping behavior, branded keyword growth, and share of voice. These measure actions, not intentions, which is exactly why they make a useful check on what the survey claims.

Modeled estimates. Conjoint or discrete choice modeling infers the attributes that move consideration and predicts share under different propositions. It is the most expensive option and the only one that answers “what would happen if we changed this,” so reserve it for positioning and pricing decisions.

Platforms increasingly sell consideration as an optimization target, and marketers keep asking what those objectives actually optimize for. A useful answer is to treat platform-reported lifts as a directional signal and validate them against your own survey instrument, not as a replacement for it.

4. Design Survey Questions and Response Scales

Write neutral items, randomize brand order, and keep response options symmetrical. Double-barreled questions produce unusable data because you cannot tell which half moved the answer.

A usable aided item looks like this: “Which of the following brands, if any, would you actively consider buying from the next time you shop for running shoes?” Options: all listed brands marked as considered, plus “none of these” and “not sure.” Add the none option or you will inflate the result by forcing a positive.

An unaided item: “Thinking about running shoes, which brands come to mind first?” Capture the verbatim response and code it blind, so the coder is not steering the result toward a hoped-for brand.

Two tests decide whether the item measures shortlisting. First, run a split sample where the brand list includes a plausible decoy; if responses barely move, the question is not sensitive to real alternatives. Second, compare against familiarity: respondents who have never heard of a brand should not score high consideration, and if they do, the response scale is too loose.

Rotate the order of brands and options between respondents. A brand that always appears first wins a share of the answers it did not earn, and that bias stays constant across waves, which makes your trend line quietly wrong.

5. Recruit the Right Sample

Define eligibility with category involvement and purchase timing screens, then size the sample. As a working rule, roughly 1,000 completes gives you about plus or minus three points of margin on a single percentage at 95 percent confidence, and enough base to compare a handful of brands. That is a starting point, not a guarantee; calculate properly for your design.

The bigger risk is who you survey. Existing customers, subscribers, and fans will report high consideration for your brand and low consideration for competitors, which flatters exactly the number you most want to prove. Screen out current customers where the question is about category buying behavior, or analyze the category buyer and the customer as two separate groups.

Apply quality controls that are actually built in: attention checks, a minimum completion time, response-pattern screening, and open-text review. Then apply weights for age, region, gender, and category spending so the results match the market you defined.

Segment cuts shrink the base fast. Cutting a 1,000-response sample into six segments leaves roughly 165 responses each, and a ten-point swing inside that cell means little. Publish segment results only when the base supports the claim, and describe the base size next to every number. This is the most common statistical complaint about DIY brand studies, and it deserves more attention than it usually gets.

6. Calculate and Report the Results

Aided consideration is simple: respondents selecting the brand, divided by all eligible respondents in the target market, multiplied by 100. Keep the denominator clean, since dropping incomplete or disqualified cases from the base inflates the result.

Worked example. You field 1,200 completes. After screening, 1,050 are eligible. Of those, 312 select your brand. Aided consideration is 312 divided by 1,050, or 29.7 percent, which rounds to 30 percent. With a 1,050 base, the margin of error is roughly plus or minus 2.8 points at 95 percent confidence, so the defensible statement is 30 percent, plus or minus 3 points.

What that result proves: about three in ten eligible buyers in this market would put your brand on a shortlist for this category within the stated window. What it does not prove: that you will win a quarter of the category, that the figure will hold in another market, or that these 312 people are the same people who convert.

Then build the confidence interval, rank against the comparison set, and cross-tab by the segments that matter to your decision. Report the point estimate, the interval, the base size, the fieldwork dates, and the exact question wording, in that order. A results pack missing the wording is not reproducible.

7. Diagnose Drivers and Turn Findings into Action

A score without a driver is a report card with no teacher. Use the survey to explain the number: test awareness against consideration, perceived difference, trust, relevance, availability, price acceptability, and any barrier the category has, such as a fit concern or a switching habit.

The gap pattern does most of the diagnostic work. High awareness with low consideration means reach is not the problem and relevance, credibility, or proposition is. Low awareness with high aided consideration means people like you once you say the name, so the work is unaided memorability and discovery. Both low points to a category position problem. High everything with low sales points to a conversion or availability problem, not a brand problem.

Segment comparisons sharpen it further. A 30 percent overall figure that is 55 percent in one region and 14 percent in another is a targeting problem, not a brand problem, and the action list looks completely different.

Assign one action per driver, set a target date, and retest on the next wave with identical wording. If the result has to prove causality for a big spend, run a brand lift study through an ad platform with a control group, then use the survey to explain the mechanism behind the lift.

Common Mistakes in Measuring Brand Consideration

Nearly every questionable consideration number I have seen traces back to one of these six errors.

Treating awareness as consideration. Knowing the brand exists is a different question from choosing it, and a famous brand can score high on one and low on the other. Fix: report both, always as separate lines, and describe the gap rather than letting the higher number stand alone.

Asking a leading question. Wording that argues for your brand, or a response scale with no neutral or negative option, inflates the result. Fix: neutral phrasing, randomized order, a none option, and a decoy brand to test sensitivity.

Mixing purchase intent with shortlisting. “Would you buy this in the next month” measures a much smaller and much later group than consideration does. Fix: keep them as separate items with separate scales.

Using a sample that already likes you. Panels built from your own email list or buyers of a competitor’s product bias the comparison. Fix: recruit from the category, screen customers out or analyze them separately, and weight to market profile.

Reporting a single percentage with no context. A 30 percent figure without a base, a margin, a benchmark, and the fieldwork window is not a finding. Fix: number, interval, base, dates, wording, every time.

Declaring success from one period. One wave cannot separate a real effect from sampling noise or a moment of cultural attention. Fix: plan three waves before the campaign ends and treat the first as a baseline.

Two smaller traps worth naming. Measuring the rejection set is useful, since a brand that appears on an explicit avoid list behaves differently from a brand nobody mentions, so ask it as its own question rather than inferring it. And a growing number of buyers now start with an AI assistant instead of a search engine, so a repeatable prompt set checked across ChatGPT, Perplexity, and Gemini is becoming a supplementary signal for whether your brand is surfacing at the moment of asking.

A few reporting habits that make results stick: keep one file with every wave’s questionnaire and codebook, store results as base sizes rather than percentages only, write the decision each wave was meant to inform at the top, and set a review date before you need the number rather than after.

Frequently Asked Questions

Conclusion

Start with the sentence, not the survey. Write the definition with a numerator, a denominator, and a time frame, pick one decision it feeds, and set the benchmark before fieldwork begins.

Then pair one direct measure with one supporting one, since the survey tells you what people say and the behavioral data tells you what they do. Read the result by segment, diagnose the gap between awareness and consideration, and assign one action per driver. Treat how to measure brand consideration as a repeated cycle rather than a one-time score, because the only number worth defending in a budget meeting is the one built on identical wording three waves in a row.