15 Minute Read
In almost every conversation and interview we've done to date, researchers describe highlighting and tagging data as manual and tedious. One researcher even compared this step in her workflow (now famously) to a chore like cleaning the toilet. She said that while it takes effort to muster up the energy to start highlighting and tagging, knowing there are richer insights (or in her analogy, a clean bathroom) on the other side of the work is what motivates her. Coding qualitative data was a labor intensive but necessary step to structure data and build patterns, to evolve data into higher order meaning as information.
While there will always be the need for researchers to read and label data for the high degree of accuracy and quality some projects require, there is also an opportunity to leverage technology to make this process faster and easier.
We’re excited to announce an instantaneous way to code your data: Automatic Highlights & Tags on Summaries.
Now, you can automatically highlight and tag AI-generated Summaries on data files, reducing the friction to generate Insights in Notably almost entirely. In the status quo outside of Notably, what used to take weeks to process interviews or notes, now can happen in a matter of minutes, going from raw data files, to structured patterns, to generating insights.
We love to think about that story I led with because, as moms and fellow toilet cleaners who have logged hundreds of hours highlighting and tagging research data (before Notably the product existed!) and while we love a clean bathroom… we know there’s more than one way to get there.
So that’s why we always prioritize putting the researcher in control. Anything you can accomplish automatically or with AI in Notably, you can also do manually, just as you’ve always done. Auto-highlight & tag or take a slower, more rigorous pass and do it manually.
Reading and marinating in the data is an important step. And what I love about our AI summaries is that they dig way deeper into the richness and nuance of the convo than typical AI summaries bc they were written for research.
Tags are applied automatically based on how researchers historically used them and the keywords and descriptions of the tags provided by the researcher. Based on our convos, researchers generally have confidence in their quality control abilities. They prefer having the assistance of an initial coding pass that they can review, edit, or correct, rather than starting from scratch, doing everything manually, or, even worse, not having the time to code at all.
Create your own AI-powered templates for better, faster research synthesis. Discover new customer insights from data instantly.
When it comes to organizing and interpreting complex data, tagging is an essential tool, and there are two main strategies to master: inductive and deductive.
Inductive tagging is a ground-up approach where you derive tags from the data itself. Like assembling a puzzle without the picture on the box, you don’t start with preconceived notions of what the final image should be. Instead you allow patterns to emerge from the raw data itself without bias. This method is particularly effective in discovery research.
Conversely, deductive tagging is a top down approach where you develop tags based on your research questions or an existing research framework or theory. Like following a blueprint, you have a clear end structure in mind influenced by theoretical constructs or specific hypotheses. Tags are predetermined ahead of analysis. This method is ideal for structured research, like usability testing.
Suggested and automatic tagging in Notably combines these approaches, but uses a mostly deductive approach out-of-the-box. You start with a series of tags and then inductively come up with codes as you analyze your project data.
As a bonus for this release, we've improved our suggested tags algorithm. Now, the keywords added to your tag descriptions have more influence. And as a double bonus, we've written a prompt you can use with ChatGPT to generate keywords to enrich your custom tags at the bottom of this post.
This type of automation doesn't exist on any other research platform. We hope this new feature saves you a ton of time and helps more teams introduce a level of rigor, which might otherwise be skipped due to time constraints or lack of research training.
For more information on how to use this feature and tips for enhancing the accuracy of tags, read this help article.
Assume you're a qualitative researcher at [Company & Description], tasked with coding unstructured data from sources like focus groups, interviews, and surveys. Your objective is to identify words and phrases related to [Research Topic] that participants might use, formatted as a list.For the tags 'Tag 1,' Tag 2,' Tag 3,' and ‘Tag 4' provide a description in under 40 words. Then list 15-20 relevant, colloquial keywords for each tag.These should be in the first person and reflect a variety of regional dialects, income levels, ages, and accents. Keep each keyword to a maximum of three words and present them in a line, separated by commas. Do not include quotation marks around the keywords.
Try Teams for 7 days
Free for 1 project