Focus group interview (FGI) transcripts refers to data that has been audio recorded and transcribed. It is challenging to understand consumer discussion and interaction. Researchers are required to read the document iteratively. Topics need to be examined in detail and in depth. Some researchers physically “highlight” certain sections of document to find similarities and differences across cohorts.
FGI data is based on consumer experience and richer than quantitative data. While reading FGI, intricacies and complexities about consumers are discovered by Researchers. Researchers also need to find similarities and differences between various cohorts such as life stage and users. It is tough to generalize from a specific Focus Group Interview cohort to a larger population.

Infopickle recognizes that each FGI data must always be understood in relation to the client brief and category context. Outputs from Infopickle helps qualitative researchers organise, notice and represent their data in 3 ways: Data Sorting, Auto Summarisation and Perceptual Maps.

Legacy vs Contemporary

Data Sorting for focused reading

Having just conducted a focus group or an in-depth interview, researchers have a key hurdle before they can start interpreting and reporting their study. Transcripts need to be converted from audio to text, cleaned and collated for analysis. Each project constitutes 12-25 focus groups or in-depth interview transcripts in word (*.DOCX) or txt formats. Each document is 50-100 pages long and has different activities such as self-introduction, attitudes, interest, opinion, market-mix and purchase intention. Browsing through multi-doc, multi-activity is very tedious. Some research firms manually rearrange the data in excel format (*.XLSX) to make it navigation friendly. This manual transfer of data still takes 40-80 hrs per project depending on the analyses heads (rows) and number of FGI (columns).

Infopickle uses Natural Language Processing (NLP) to perform the following tasks
– Read by Analyses heads: Infopickle automatically classifies moderator questions and corresponding responses into various analyses heads such as introduction, word associations, concept exposure, product experience, attitude towards ad, message comprehension, and personification.

– Easy to Navigate: Infopickle arranges responses in meaningful order to read and analyse effectively in MS excel.

– Data Management: Infopickle removes irrelevant filler words such as ‘M: R: Ok Thank You’ and long moderator sentences are compressed. Difficult to analyse text is repaired. For example, repetitions such as “I am ready” spoken by 3 respondents to same moderator question is represented as “I am ready (3)”.

Benefits to Qualitative Researchers

  • Focused reading as per analyses heads.
  • Similarities and differences between different cohorts such as demographics, users and life stages.

Auto Summarisation of recurring themes

As researchers read and re-read transcripts, they develop hypotheses that needs validation. They discover relationship between various “tabs” or contexts that needs generalization. They draw out relevant quotes to support their key story. Wading through 1000+ pages of words to test multiple theories or code data is a daunting task.

Thematic Analysis: Infopickle extracts verbatims directly from the transcripts. One form of extraction is by each theme across all responses. Other form of extraction is by each analysis head or rows. Themes are extracted automatically using text analytics.

Prioritised Themes: Each analysis head is paragraph long which is compressed to support reading and inference of FGI. Ideas are easily collated in one place saving reading time. Researchers can categorise themes against actual verbatims in different analyses heads.

Sentiment Analyses: Each sentence is coloured with sentiment words. Researchers can decide if the sentiment is meaningful to study objectives or not.

Online vs Offline
Customised vs off the shelf

Perceptual Maps

Focus Group Interview (FGI) usually administers several projective techniques: ranking, word association, and mood boards. These techniques generate lot of attributes. Attributes that differentiate across studies, brands, market are deemed to be important. Differentiation seeks to distinguish one brand from others on any basis that is important to the buyer. Importance of attributes can be further linked to associated brand preference and behaviour.

Infopickle creates brand perception map, mind maps and onion peel maps directly from focus group interview transcripts.