Quick note: and I were invited onto the Marmind podcast to share some stories about brand marketing. Check out the video below, or listen to the episode here — or wherever you get your podcasts (I’ve always wanted to say that!). 

Hooray, you just surveyed 1,000 people to find out what they think of your brand. You now have reams of data, sitting in a file, waiting to be translated into meaningful takeaways to share with your team. 

Or, you just completed a batch of interviews with your ideal customers. They told you so many amazing things, dozens of specific requests and insights, which you’ve added into a research database that’s bursting at the seams.

Now you’ve just got to figure out what to make of all of it. 

Gulp! 

The origin of the phrase “reading the tea leaves” is connected, unsurprisingly, to actual tea leaves. Tasseography, aka tea leaf reading, began in 2700 B.C. in China, shortly after the discovery of tea itself. People would finish their cup of loose leaf tea, observe the patterns and shapes of the remnants that were left behind in the cup, and divine the future. Reading tea leaves spread from China across Asia and the Middle East and gained popularity in Europe in the 17th century, where it took root as a metaphor – one that today’s business world has gladly adopted. 

We are reading tea leaves all the time at work: 

• What am I supposed to do with this peer feedback I received? 

• How can I even make sense of these customer interviews?

• What does all this data even mean? 

It’s hard to know what are the right answers when it seems like there are so many answers!

We are so often called on to be storytellers, whether we purposefully signed up for the job (hello, fellow brand and content marketers) or we simply need the skills of story and executive presence and talking good in order to do our jobs well. Combine this with the fact that we live in an era of Big Data and a world of information overload, and it becomes all the more critical that we are able to pull meaning from ambiguity, to parse a signal from all that noise. 

Typically, noise will manifest in one of two ways: 

1. Mountains of qualitative feedback – surveys, reviews, interviews, etc. 

2. Gobs of quantitative data – metrics, KPIs, spreadsheets, etc.

There are techniques in both arenas that can help you find the signal and tell compelling stories. Here are some of my favorite techniques for making sense of both buckets: the qualitative and the quantitative.

Qualitative “Noise”

Qualitative noise most often arises after conversations or interviews, be they research calls or peer feedback or open-ended survey responses. You may often find yourself looking at paragraphs of text, delivered in someone’s unique voice and vocabulary. This makes it all the harder to suss out takeaways quickly and at-a-glance. Everything looks and sounds so distinct!

But there is hope. Here are a few techniques for making sense of qualitative, open text. 

1. Do not presuppose

One of the core values of good customer research is to avoid asking pointed questions. Similarly, when analyzing qualitative responses, don’t try to back into an answer you’ve already determined. Confirmation bias can be real tempting as a shortcut to putting in the actual work of pulling the signal from the noise. As a simple thought exercise before getting started, it can be helpful to write down all the biases you’re bringing into the analysis so that you can be aware of them if you sense them cropping up later. For instance, if you’re interviewing customers about the product roadmap, your biases might be: 

• You hope they ask for a certain feature

• You orient toward the one customer journey that you happen to know best

• You presume the “why” behind their behavior

2. Word clouds

Take all the qualitative information you’ve been given, and enter it all into a word cloud generator; there are a lot of free tools out there.This will help you highlight some of the most common words that people are using, which should give you a better sense of any macro signals. A couple things to note: 

1. There may be a handful of synonyms in the list, which can mask a strong signal. To avoid this trap, don’t just look at the two or three most prominent words; look at the top 15-20 and keep an eye for any words that have similar meanings and could be grouped together. 

2. When telling your story to a group of peers or company leaders, a word cloud makes for a great visual. Just be sure that you give the graphic a little TLC by removing filler words like “the” and limiting the total number of words you show in the graphic, both of which your word-cloud tool may do automatically for you.

3. Make it quantitative: Categorize

It doesn’t seem like qualitative responses belong anywhere near a spreadsheet, but the spreadsheet can actually end up being a signal-seeker’s best friend in some cases. This technique is often used in NPS reporting, when you need to figure out actionable takeaways from those who love your product and those who don’t. But the technique can work on any qualitative feedback. 

1. Enter all your feedback into a spreadsheet. One sheet or section for each question. One row for each response. 

2. Assign each response to a particular, high-level category. At first, you’ll be guessing at the categories, but as you go along, you’ll begin to sense patterns and familiarity within the answers. 

3. If your dataset feels particularly unwieldy, you can add another column for sub-categories where you can get to an extra level of granularity.

Once you have all the feedback labeled, then you can create all sorts of charts and graphs to visualize the key takeaways from your data.

4. Get AI involved

No, it’s not cheating. It’s resourceful! 

These days, you can rely on AI to be your research assistant by giving your qualitative data over to ChatGPT and asking it to summarize any key learnings. It’s important to use AI as a tool, not the only tool. I’d suggest combining the takeaways from AI with your own takeaways from one of the above categorization or word-cloud techniques. The more that you can triangulate the signal from the noise, the more confident you can feel that you’ve found the true signal. 

Quantitative “Noise”

Quantitative data can take many forms, depending on your role on the team and your seniority at the company. Sometimes you might be looking at data for a channel, sometimes for a campaign, sometimes for a funnel, sometimes for the entire business. But in all cases, it can be a game changer if you’re able to take boring old numbers and turn them into a compelling and insightful story, full of clear signals. 

Here are a few techniques to get at the heart of what the data is telling you. 

1. Understand the methodology 

Before you go drawing your conclusions, you want to make sure that you understand the rules of your data universe. Methodology is key, especially when you’re analyzing other people’s data. For instance, if you are doing competitive research on your category, you may be relying on studies and reports from others. Or if you’re pulling funnel metrics for your marketing team, you’re likely borrowing from a number of different sources of truth across the department. 

Make sure that you know the definitions for your metrics, the timetable for your data pull, and the characteristics of the audience that you’re analyzing. 

2. Five Whys

This technique is an exercise used for retrospectives within product and engineering teams but has since gained traction for post-mortems all across the business. I like to adapt it to signal vs. noise storytelling. 

The way Five Whys works is relatively straightforward: 

1. You begin with an inciting incident. In a typical product/engineering case, something has gone wrong. In our storytelling case, the inciting incident can be a new data insight.

2. You then ask Why? Why did this happen? Why did this datapoint occur? 

3. You then query your answer with the same question: Why? 

4. You do this five times.

By the end, you will have dug deep to understand the true root cause of your data discovery, which often leads to a more robust, confident, and informative story. 

For example, let’s say we’re analyzing some acquisition data from our marketing channels, and we notice that last week was a particularly good week. Our first question would be, “Why was acquisition so high last week?” The answer might be that we had great performance from our organic social channels. Then we would ask, “Why was organic social so strong?” The answer here might be because reach and impressions had a really big week. Why? Because a new video went super viral. Why? Because we started a new video strategy based on trending news.

3. PPDAC

A popular scientific research method, the PPDAC is used by researchers to perform statistical investigations. The acronym works like this: 

• P - Problem: This is a hypothesis you begin with, which often leads to data collection in order to prove or disprove the hypothesis

• P - Plan: The methodology for how you’ve collected the data

• D - Data: The data itself

• A - Analysis: Using some data best practices to dig into the numbers relative to your hypothesis

• C - Conclusion: Returning to the original Problem statement to prove or disprove it

The real magic with PPDAC is in the consistent focus on solving a particular problem. One of the tricks with the noise of Big Data is that you lose sight of your target very easily. This PPDAC technique makes sure that you’re constantly returning to that original premise and using the data to answer the story that you started with. 

4. Questionstorming (Bonus: This works great with qualitative “noise,” too!)

If you spin ahead to you presenting your signal to your boss or your peers or your leadership team, you can imagine you are going to get asked some tough questions. So in this technique, in order to future-proof the signal you think you’ve found, you try to think up all the questions that someone might ask you once you’re presenting your final story. All questions are okay here; the point is to exhaust all question possibilities, no matter their quality or altitude. Afterward, you may choose to go back and refine your takeaway or simply prepare an answer to any questions you anticipate might need some practice. 

Over to you

When have you had to find signal from noise at work? How did you go about it? I’d love to hear your stories and to learn what worked and what didn’t!

For more…

Follow us on LinkedIn and on Instagram. Stay tuned to our Substack space for new community features and ways to meet your fellow subscribers.

And let us know if you’d like to go deeper with us to talk about you and creativity. We do:

• 1:1 coaching and mentorship

• Team workshops and consulting for marketing and leadership

• Speaking and appearances on podcasts and at events

Say hi anytime.

Kumbaya,Shannon & Kevan