Method Series: Leveraging Your Highlight Data & Insights 101

Congratulations! You've completed your project on Highlight - and now you're onto the fun part - the Insights experience in the platform. On more longstanding, built-out, or enterprise teams, you may have a turnkey, structured process for turning Data into Insights and your next steps might be clear.

Then there's the rest of us. You want to tell a compelling story that retailers can't resist, or that your internal stakeholders can make decisions from, or that your consumers will gravitate towards. But if you're not an expert, how do you leverage your Highlight data and insights to tell the right story, and have the most impact?

Here's a set of starter tips to help you leverage your data in the right ways - no matter the audience or use case.

Reporting rules of thumb

Get started with these best practices for reporting on your data:

USE “TOP 2 BOX” REPORTING

Best Practice: "Top 2 box," often annotated as T2B, means summing the percentages of the top 2 responses to a question. It's the most common way of reporting on 'high scoring' metrics, so we'd recommend using this aggregation method to report on Likert or hedonic scale questions - especially 5 to 7 point scales.

Examples: 

• In 3rd party research with 100 consumers nationwide, 69%* loved our product when rating overall enjoyment.
• *Question: How much did you enjoy this product? Scale: 1 (Not at all) to 5 (Extremely)

USE BAR CHARTS TO VISUALIZE COMPARISONS

Best Practice: Bar charts are ideal for visually representing discrete data points or comparing data points, such as between products in competitive testing or when testing multiple prototypes. You can export these directly from the Explore page in your platform.

Examples: 

PULL STANDOUT QUOTES, AND ATTRIBUTE THEM

Best Practice: In your data, you can pull qualitative quotes from open ended question responses, to highlight standout quotes, or representative perspectives. This can be incredibly rich, illustrative data to add color to quantitative reporting. In the Highlight platform, you can access and dive into verbatims for all open-ended responses.

You can also attribute a quote to a respondent, for whom you have specific profile data: choose from a range of demographic, psychographic, and shopper data points, to help describe the respondent in the most helpful way. While research and privacy standards prevent us from sharing respondent names, it’s not unusual to assign a persona name to bring quotes to life while keeping them anonymous.

Examples: 

• “It's exactly as you'd imagine a double chocolate chip cookie but even better. It's guilt-free indulgence.” - Chris, 24-year-old flexitarian man in NY

• “Wonderful new carbon neutral, non GMO, TASTY cracker made by a climate friendly company. You simply have to try them!  You will love them.” -Cindy, 37-year-old Target shopper in CA

ALWAYS SOURCE YOUR DATA!

Best Practice: 

• Research source (Highlight data, 3rd party research, consumer testing)

• Sample size/scale of test
• Time of testing
• Region
• Any additional descriptive data

Examples: 

• *3rd party consumer research conducted in February 2025, across n=100 consumers ages 18-66 nationwide

  OR

• *Source: [Brand] x Highlight custom online survey of [x US adults], letshighlight.com  Product Research, [January 2025]

Excel data exploration tips

Highlight's platform offers a robust analytics suite that allows researchers and brands to see powerful insights with just a few clicks. Curious to dig deeper and explore a bit more of the raw data in Excel form? Here are a few beginner tips for exploring your data to build hypotheses - primarily if you're using our Raw Data Download to poke around.

IDENTIFY OUTLIERS

Best Practice: Use the Excel dataset to sort specific columns (questions) from high to low to understand both qualitative (open ended) responses to why they felt that way, and dig into the respondent info to build hypotheses for who is rating certain metrics extremely high or low.

Examples: If only 3 people selected an NPS of 0 for your product, dig into that data - what were the open ends? Why 0? Is there any consistency in type of person who answered this way such as where they live, their dietary preferences, or other lifestyle factors?

SORT YOUR DATA

Best Practice: Similarly, use the Excel dataset to sort specific columns (questions) from high to low and scan demographic data to see if you can build hypotheses for trends in people characteristics that might be driving scores up or down.

Examples: If you’re looking to understand how to improve texture, sort by high to low scores and look at the lowest score open-ended responses. Skim the demographics to see if there’s a type of person that particularly doesn’t like your product texture. Beware of sample sizes being too small, but use this as directional insight!

HOW TO THINK ABOUT FILTERING & LOW BASE SIZES

Best Practice: Before filtering or segmenting or cutting your data too much, consider the sample size and base size of any subgroup you analyze. Digging into "what to moms think" or "what people in Texas think" may be interesting, but extracting conclusive insights from too small of base sizes is not recommended. You can start to build hypotheses but just remember - the "base size" of that audience is too low to draw validated conclusions.

Examples: You can always search the dataset for “Whole Foods” or “Sprouts” to pull out specific shopper insights. You can definitely pull quotes for retailers or shopper decks, but quantitative stats (80% of Sprouts shoppers would recommend this to a friend) are not recommended if we're only considering a handful of Sprouts shoppers! In this case, n=50 or greater would be worth reporting on.

Here's an example of how one brand created a presentation for the retailer Sprouts by using bar charts and annotated open-ended quotes from their Highlight testing.

Frequently asked questions

What’s a good score? What’s a bad score?

This is relative and can be observed through comparison. If all scales in your questionnaire are consistent (e.g., a 9-point Interval Scale), you can quickly observe the product attributes that are underperforming relative to the rest. Talk to your Highlight team member or book a demo to learn about Highlight's Benchmark Builder Solution and more robust ways of benchmarking your product performance.

How should I think about sample size & statistical significance?

Statistical significance determines that a relationship between two or more products is caused by something other than chance. Highlight's Scorecards and Crosstabs allow for stat testing at several different confidence levels.

Confidence level is the probability that the observed result cannot be explained by sampling error alone. If a result is significant at the 95% confidence level, it means there is a 95% chance that the difference is real and not just a quirk of the sampling. If we repeated the study 100 times, 95 of the samples drawn would yield similar results. This is important because when we state that a comparison is “significant,” we are trying to extend our limited survey results to the larger population.

Read more about how to choose the right sample size for your next product research product.

How do I feel confident making decisions using my data?

Using statistically significant differences to drive insights and action is a great place to start. Talk to your Highlight team with additional questions! The Highlight software and team is here to ensure you have the data you need, to answer your key questions and enable your team's success.

Create products people love