Using Observations in Analytics

Observations can be explored in Analytics either on their own or in the context of a Theme.

Exploring them on their own is useful when you want to surface the top issues or praises across all your data.

Exploring them within a Theme is often more focused, as it shows which specific issues drive sentiment or NPS for that Theme.

This guide explains how to work with Observations in Analytics: where to find them, how to use them as a breakdown or filter, how to combine them with Themes, how to show descriptions for easier interpretation, and how to use Quick Actions to focus your view or dig deeper.

You can work with Observations in two main ways in Analytics.

We generally recommend starting from a Theme and then drilling into its Observations. This approach gives you the broader context first, then lets you see the specific issues driving that Theme’s performance.

For example, you might start with the “Product Packaging” Theme and then look at the Observations beneath it to see whether complaints such as “too much packaging” or “packaging is not recyclable” are contributing to a rise in negative sentiment.

You can also explore Observations on their own - without a Theme breakdown - when you want a cross-sectional view of the data. This is often useful for spotting emerging issues or opportunities across all feedback, regardless of Theme.

A common example is creating a table that shows the top issues across all feedback by volume of responses.

Observations can be used either as a breakdown or as a filter.

When you use them as a breakdown, you can choose whether to display all Observations or only a particular type - issues or praises.

Because there can be many Observations, you can also limit the table or chart to show only the top set (for example, the top 10 by volume).

If you want to see both the broader category and the granular issue in the same view, first apply a Theme breakdown, then apply an Observations breakdown as the second breakdown.

This shows how specific issues contribute to the performance of each Theme.

You can also display short plain-language descriptions under each Observation name to make them easier to interpret.

Turn on the Show Descriptions toggle in the table or chart to reveal these summaries. This is especially useful when there are many Observations or when you’re new to the data.

The same description is also shown when you open that Observation in the Feedback view.

Quick Actions let you focus on what matters or investigate an issue without changing the underlying data.

They can be used alongside Breakdowns, Filters and Descriptions to move smoothly from a broad view to a focused analysis.

You can exclude an Observation to remove it from the current table or chart. This is helpful for cutting out noise - for example, one-off or low-value items - while keeping the Observation available elsewhere.

You can drill down by an Observation to narrow the analysis to that single issue. The rest of the Analytics view adjusts to show only the feedback linked to it. For instance, you can drill down on “Sauce breaks when water is added” to see how that particular problem behaves across segments or over time.

You can also filter by one or more Observations to compare groups. For example, you might compare NPS or sentiment between customers who mention a certain issue and those who do not.

Finally, you can use Zoom In for instant context. When you zoom in on a charted data point, the Insights by Lyra panel opens on the right with an automatic summary of what drives that point. When you zoom in on a single Observation, a dedicated panel opens showing its description, key metrics and a few representative snippets.

A good practice is to use these tools in combination.

Start by breaking down by Observations, exclude the ones that are irrelevant, drill down on the one that matters most, and then zoom in to get a concise explanation - all without leaving the Analytics page.

Remember that Quick Actions only affect your current view.