Take the Pain out of Pain Point Analysis with ChatGPT
Updated: Apr 24
Every entrepreneur we have worked with has a love-hate relationship with the Customer Discovery Process (it's mostly love, we promise). The love happens during your conversations with potential customers, uncovering pain points, ideation sessions around product pivots, and validating your concept's resonance with potential customers. The frustration kicks in the moment that your team sits down to analyze those 20+ hour-long conversations.
Qualitative data, while essential to conducting robust customer discovery, can be very time-consuming to analyze.
Let's first explore a low-tech way to analyze customer data, then see how we can use tools like ChatGPT to increase the efficiency of the process.
Method 1: Dump and Clump
From a low-tech, easy-to-understand perspective, The Scientific Startup is a big fan of the "dump and clump" methodology. In fact, this is the only method we use when founders are kicking off their first round of customer discovery. The methodology uses a tool such as Miro (or good old-fashioned sticky notes) and a few hours of your time to sort through qualitative data and arrive at a set of customer insights that you can apply to your business.
Here's how it works.
Create your discussion guide.
Have 45-minute calls with your customers and take notes (by hand or with a transcription too). Capture as much data as possible.
Create a miro board with boxes for each customer call. Within the box for "Participant 1," copy each complete thought from your notes into a single sticky note. Do this for every customer. (You'll probably have around 40 stickies per person!)
"Dump" all those sticky notes into a new spot on the miro board (don't ever mess with raw data).
Start "clumping" them by theme. Did the word "cost" come up ten times from different people? Place them all together. Do this with every major theme.
Here's the simple truth behind the "dump and clump method"- the more people mention things, the more important they are. By the end of the process above, you'll have a sense of the "most frequent" pain points, themes, unmet needs, and jobs to be done for your business and product.
From there, you're ready to apply customer insights to your business, messaging, pitches, and wherever else a customer perspective adds value.
Method 2: ChatGPT Enhancements
While every entrepreneur should become familiar with the dump-and-clump methodology, we're all about the efficiency and scalability of customer discovery. Since we encourage a round of discovery with every major pivot of your business, it's important to have additional tools available to keep your business moving quickly.
Let's explore three ways that ChatGPT, inspired by an actual prompt to the platform and modified for our community of startup scientists, can support qualitative research analysis for Customer Discovery.
While the use cases are pretty broad, we agree with three ways that ChatGPT itself suggested that it can support qualitative data analysis.
Text Summarization: ChatGPT can summarize large volumes of text into a shorter, more concise version without losing the essential information. That 3-page document of notes can quickly be turned into 1 paragraph for a quick review and share with your team.
Sentiment Analysis: ChatGPT can analyze the sentiment of text, allowing researchers to understand the emotions and attitudes of participants. This can be particularly useful in qualitative research, where understanding the emotional response of participants can provide valuable insights.
Topic Modeling: ChatGPT can identify the topics being discussed in a large volume of text, allowing researchers to understand the themes and patterns that emerge from the data.
In its initial response, the tool also highlighted is that it can support "deeper dives" into the data by adding additional insights or perspectives that might not be in the dataset. While we do believe that the technology might advance to a point where you can "fill in the gaps" in a qualitative dataset, we caution researchers against using this feature of ChatGPT given the age of the data (a lot has changed since 2021), and the fact that it would not represent the authentic voice of your customer.
For example, you might introduce incorrect insights if you're launching a real estate or property technology company in 2023 and "fill in the gaps" with ChatGPT data from 2020-2021 before the major market shift. As one might imagine, a customer with an interest rate under 4% behaves very differently than a customer with a 6% rate.
All things considered, we highly encourage entrepreneurs to start testing some of these tools in their customer discovery processes. If AI tools help to keep you more accountable and increase your speed and agility, go full steam ahead! Just remember to spend some time reviewing ChatGPT's output on your own or with your team - a human touch makes the Customer Discovery process even more effective.
Measure Twice, Launch Once. 🚀 Join Our Community for exclusive content