Researchers have struggled for years to identify, classify and quantify themes in their data.

That is why we are excited to share that you can now accomplish this in just seconds in Speak!

This is a game-changing solution that is only available on our platform.

Want to identify, classify and quantify your first themes? See exactly how to do it:

In this video tutorial, we walk through the entire process of analyzing a synthetic Spotify cancellation survey response and then identify, classify, and quantify themes from it automatically using Speak's AI:

To begin your theme analysis process, you need to upload the data you want to analyze.

With Speak, you can upload and analyze audio, video and text data.

You can do that through the software, Zapier integration, CSVs and APIs.

You can easily upload from your dashboard or anywhere in the app by selecting the "Quick Actions" dropdown (the image above is a good reference).

We have documentation on all of these options so please review in our help center!

We recommend putting all the data you want to analyze together in one folder.

You can easily do this on upload or edit files after upload to send them to your desired folder.

Once your data is uploaded and organized in your desired folder, you now want to create the fields you want to map the AI analysis responses to.

You can easily create as many custom fields as you want in Speak on this page.

If you don't have any pre-determined themes, a great option is to create a text field called "Exploratory Themes".

In this process, we are working through "Theme Identification".

After identifying themes, we can easily standardize the themes for consistency across the data and quantification using the same process we are detailing.

If you already know your themes, then you may want to name the Fields "Themes" or "Standardized Themes".

In this case, in the AI analysis step we are showing you next, you will be providing the AI with some context and the list of themes and forcing the AI to pick one or several of the most relevant themes. This enables standardization and easy quantification across your data set.

This is known as "Theme Classification".

Now that you have created your fields, you want to be able to view the results from the folder level.

This is easy to do! Just visit the folder, select "Columns" on the top right and move the columns you have created into view by dragging them from the right side to the left.

Speak will store these preferences in the session and they will remain as long as you don't reset your cache. If they do disappear, you can easily re-add them into the view again.

Visit the folder where you have organized the data you want to analyze.

You can see the folders you have created on the left side of Speak.

If you don't select any files, Speak will analyze all the data in your folder.

Alternatively, you can select an individual file or a few files using the checkboxes on the left side. If you do this, Speak will only analyze those selected files.

With your desired files selected, select either "Prompt" or "Chat" and a pop-up will appear.

Once the pop-up appears, select "Map Response to Field" and you will be asked to select the Field you want to map the response to. That means when you run your analysis, the answer will be added to the field.

In this example, we chose "Exploratory Classification" for the mapped response field.

It is now time to write your prompt to help the AI run its analysis.

Below, we have provided prompt structure ideas for you based on whether you are identifying themes or classifying themes:

Please review this _______ data and provide the main theme for why _______. Please be concise and only provide the Theme Name in the output.

Please review this _______ data and give me only one theme from this list:

Respond with nothing but the selected theme from the list above. Do not include the number. For example, you should only return: "EXAMPLE THEME". Consistency is crucial as we have standardized themes we need to quantify.

Once you are happy with the prompt you can run the analysis.

The data length and volume will determine how quickly the response comes back.

You will see the analysis results start to populate in your select columns.

Want to re-run your analysis with updated prompts or themes? No problem! You can easily follow the same steps above and update the fields with a new run.

If you have identified new themes and want to standardize them for better quantification, you also have a great sample prompt and process above.

Want to download the data for analysis? This is easy to do!

If you are on the 7-day trial or have the Premium Export Add-On in your subscription, just select the files you want to download, select "More", "Export", and "Media Details (.csv") and hit "Export All".

Speak will automatically generate a high-quality, nicely structured CSV with all your data combined including the columns with the themes.

Once downloaded, you can easily use Microsoft Excel, Google Sheets or other platforms to analyze:

You can automate theme identification, classification and quantification at the file level using our Prompt Automation system.

Book a paid customer success session with one of our experts today and we will make sure you are getting the insights you need to make better decisions:

Thank you so much! Have a wonderful day.