Jason Hardman's recent articles
Going granular: human analysis gives detail automated tools can't
One of the biggest complaints we hear about automated tools is their lack of detail. They simply aren’t able to give the insight researchers need to answer burning business questions. However, consumer posts online are filled with opinions that provide the detail needed to help businesses. By using human analysts to extract granular detail at the individual opinion level, WaveMetrix's pioneering approach overcomes this problem and gives researchers insight that makes a difference.
We frequently hear of the shortfalls of automated tools in giving granular details. Because their algorithms look at consumer posts on an overall level, they are unable to truly tell you what consumers are talking about and what they think.
This lack of detail is frustrating and a key part of the social insight iceberg automated tools miss. You know that the conversations being made online are detailed. It’s this detail that provides actionable insight and meets your business' needs.
WaveMetrix’s pioneering approach enables us to extract unprecedented detail. By using humans to read and tag conversations into a comprehensive and detailed categorsiation scheme, vital insight can be gleaned into what consumers really think. Researchers are finally able to utilise the conversations consumers have online to understand what consumer attitudes of specific, individual topics and understand their perceptions in-depth - rather than the generalisations automated tools make.
Automated tools run algorithms across a whole post, missing the nuances of human conversations:
- Because automated tools use algorithms to analyse online posts, they miss the vast majority of detail contained within them and only look at a generic, overall level - not the individual opinions
- This means sentiment is often be classified as neutral, when this is simply not the case. If a post has both positive and negative phrases, a tool may class it as neutral
- This also means automated tools are unable to classify posts into a detailed taxonomy, instead giving a generic view in to what consumers are talking about
Example coding of a smartphone review by an automated tool:
It's all about individual opinions - by having humans tag granular topics and sentiment based on individual opinions within a post, all the insight can be extracted:
- Individual opinions: human analysts are able to interpret online posts at the individual opinion level, meaning that all the information contained within can be extracted
- Understanding all the topics in a post: humans are able to understand that a post may reference several different aspects of a topic and tag accordingly in a detailed categorisation scheme. This means a detailed picture can be built
Example coding of a smartphone review by human analysts:
The human approach pioneered by WaveMetrix means a detailed taxonomy can be built, with data lending itself to easy interpretation and understanding:
Example data based on coding consumer opinions of a coffee machine:
- WaveMetrix's coding of invidividual opinions into a comprehensive categorisation scheme means that a detailed picture can be built up of consumer opinion, allowing for every single facet of a topic to be deep-dived to extract all relevant business insights