01 See also #TiSDD 5.1, The process of service design research, and method description Building a research wall.
02 “In contrast to this abundant data, insights are relatively rare. […] When they are generated, though, insights derived from the smart use of data are hugely powerful. Brands and companies that are able to develop big insights – from any level of data – will be winners.” Kamal, I. (2012). “Metrics Are Easy; Insight Is Hard,” at https://hbr.org/2012/09/metrics-are-easy-insights-are-hard.
03 #TISDD 5.1, The process of service design research, provides more information on indexing and how much data you need to collect during your research until you reach theoretical saturation.
04 See #TiSDD 5.1, The process of service design research, for more information on peer review and co-creative workshops, as well as chapter 6, Ideation, on how to use key insights for ideation.
First insights are often generated based on patterns you find while you are collecting data, building your research wall, or codifying your data. It helps to write down initial assumptions, hypotheses, and intermediate insights at any stage of the research process and then critically reflect on them using your collected research data. If you don’t have enough data to critically reflect on an assumption, use this as a starting point for another fieldwork session and collect more data. Design research is iterative! 
Key insights help researchers to summarize and communicate their main findings. They should be built on research data and supported by raw data, such as quotes, photos, and audio and/or video recordings. Use indexing to keep track of the raw data that supports your key insights. Key insights should be carefully phrased as they will serve as points of reference for the further design process. You might use them as the basis for ideation or later on to evaluate ideas, concepts, and prototypes. 
There are many ways to formulate insights, and which framework makes sense will depend on the research data and the aim of your project.
For example: “Alan eats chocolate because it makes him feel safe, but it makes him fat.” Formulating insights in such a way is particularly useful when your research is followed by an ideation stage to improve a given situation. The structure of this key insight framework allows you to tackle the issue on three different levels:
- Activity/action/situation: Looking at the activity/action/situation level (“eats chocolate”) could lead to a design challenge like “Which alternative or additional activities could Alan do so that he still feels safe, but that positively affect the given friction of the original activity?” (This opens up the opportunity space to think about, e.g., offering additional sport activities so that he can still eat chocolate, but also achieves his goal of not getting fat.)
- Aim/need/outcome: Looking at the aim/need/outcome level (“it makes him feel safe”) could lead to further research questions like “Why does Alan not feel safe?” or to a design challenge like “What other things might help Alan feel safe?” (This opens up the opportunity space to offer alternatives that might help make him feel safe, like self-defense courses or anything else that might affect his self-confidence, but also help him achieve his goal of not getting fat).
- Restriction/obstacle/friction: Looking at the restriction/obstacle/friction level (“makes him fat”) could lead to a design challenge like “What other food could Alan eat that doesn’t make him fat, but still makes him feel safe?” (This opens up the opportunity space to offer other food options, like low-carb chocolate or fruits or vegetables, that still make him feel safe but also help him achieve his goal of not getting fat).