Relative Insight reveals the statistically significant differences and similarities in your data. Many of these findings will be interesting, relevant and actionable while others may not be.

Insight cards serve as a tool for organizing the most insightful nuggets from the analysis that you can then use to craft narratives and bring measurable metrics to your text data.

This guide will help you make the leap from data to insights with ten tips from our experts to help you put together powerful, high-impact insights.

Always think back to the first stage of the project and focus on the discoveries that are most relevant to your research questions and the objectives of your stakeholder(s).

Relative Insight's comparative approach to text analytics reveals ‘unknown unknowns’ - insights you wouldn't have thought to search for. Try to avoid going into the data looking to prove or disprove a hypothesis. Instead, keep an open mind and prepare to be surprised!

Insight cards help keep your discoveries organized. You can easily add to or create insights from the comparison explorer and view them from the ‘Insights’ tab. Be sure to give your insight cards a title and description to summarise the insight.

You'll find interesting nuggets of insight diffused across the five linguistic categories: topics, grammar, phrases, words and emotion. We recommend starting with topics, before going through the other categories looking for evidence to support your existing insights or to create new ones.

Add anything that stands out as interesting to an insight card, even if it doesn’t seem particularly strong initially. As you create more insights, distinct themes and narratives will become clear. You may find certain themes begin to link together. When this is the case, you can merge insights together or remove any that are not relevant.

Verbatim examples help you understand the context in which a linguistic feature is being used. Context is key for building insights, as you need to understand not just what people are saying but how they are saying it. Verbatims can be viewed by clicking into a linguistic feature in the comparison explorer, or when viewing insight cards.

It may be tempting to focus on the biggest relative differences when building insights. It is important to note that all results surfaced in the platform are statistically significant, and thus you can be confident in highlighting even 'small' differences to your stakeholders.

You can also focus on the metrics that are most applicable to your data set. For example, word percentage is particularly useful for social listening when you need to know the proportion of times that a particular word or phrase is used. Or, verbatim frequency is great for surveys as it can tell you, for example, the exact number of individuals who have given a certain rating or used a particular word or phrase.

When you see a relative difference value of (∞), this means the linguistic feature is significantly present in one data set and completely absent from the other. Be cautious when you see this symbol, as an infinite relative difference does not always mean that the insight is relevant. For example, names and source-specific words will often be shown to have an infinite relative difference. Consider a comparison between a Harry Potter book and The Great Gatsby. ‘Gatsby’ would almost certainly surface as having a ∞ relative difference, but this is hardly surprising or insightful.

Once you’ve completed your comparison and built insight cards, the next stage of the project is taking action. In other words, this is the stage where you take your insights out of the platform to drive results. To build actionable insights, be sure to ask yourself 'What can be done with the information?' and 'How does it contribute to our ability to achieve our overall objective?' as you explore the analysis. You can record your ideas using the ‘actions’ field on the insight cards.

When building insights and defining actions, ask yourself ‘So what?’ or ‘Has this surprised me?’ to help yourself move beyond the obvious. This can help you take your analysis a step further, moving from what an audience is saying, to understanding why they are saying it. For example, if healthcare practitioners are talking about ‘rapid testing’, perhaps they are saying this because they feel that current testing is too slow or that there are so many patients that they need to be able to move faster. The ‘description’ field on an insight card is a great place to capture this thinking.