Boolean who? How To Set Up A Social Listening Query.
This year I’m testing a few new tools. Most of them are social listening tools, including YouScan. I liked what I saw and I teamed up with them – I’m currently testing it to assess its strengths and the opportunities it can bring to digital marketing efforts for my clients and me.
In May I also joined a Social Listening Query writing competition sponsored by them. The challenge was not easy, it read like this:
- Research question: As they enter the workforce, how do GenZ in the US feel about the impact of AI on their current and future career prospects?
- Date range: 1 year
- Markets: US
- Language: English
- Age range: Gen Z (under 27)
- Focus: The focus is on GenZ talking (not other people talking about GenZ)
I did not win but I learned a lot, especially about GenZ. And even though my query is far from perfect I wanted to show you the thinking behind a project like this, the methodology behind a Boolean search and why it takes so long to get a tool properly set up for a social listening project.
But first things first. Here is the query I sent in:
Pretty? I do not think so, it’s one of the longest and most convoluted queries I wrote. I tried to come up with a way to layer in keywords that could help me zoom in on GenZ talking because I feel like most of the online chatter comes from people talking about them, not from them.
Methodology: The Thinking Behind The Query
The rules of the competition asked us to document our query logic and send it in alongside the query I shared above. Here’s what I sent.
INITIAL KEYWORD RESEARCH
1) Asked ChatGPT to come up with a basic query on work/career-related terms and artificial intelligence-related terms.
2) Expanded on the initial query with relevant branded terms for the top players in AI. I took a look at these initial results to understand if I was headed in the right direction. Identified some keywords that were pulling in spam mentions. Excluded Gemini as it was pulling in a lot of horoscope mentions.
GENZ RESEARCH
1) Identified GenZ relevant slang AND top influencers.
I did try to identify both of these with the help of AI, but it was not 100% helpful so I resorted to the wisdom of the crowds. I asked my followers on LinkedIn (link here) and a Romanian community of marketers (link here) for their support. I Received more than 100 comments on both posts, I got to have a 1-hour long chat ith a GenZ marketer about the slang and what it means. I danced to this millennial rap about Gen-Z ing.
2) Took all this unstructured data and fed it to ChatGPT, asking it to make bullet lists and boolean queries for GenZ slang terms and influencers.
3) Also read Kantar articles on GenZ:
a. The ABCs of Gen Z
b. How well do you know Gen Z and Millennials?
c. New ways of working from Gen-Z to Boomers
4) According to all my research GenZ is most active on social media. Also, in the second version of my query, I concluded that news/blogs tend to include meta-analysis on GenZ behavor/attitude towards AI and their career prospects. These are the two reasons why I decided to filter the query to only include social media, forums, and reviews to try to improve relevancy.
WRITING & EDITING THE QUERY
- I got some errors with the query so I tried ChatGPT to identify the issue, it pinpointed the exact operator I was missing and identified some duplicate keywords. I cleaned that up.
- Check data quality:
a) First check: analyze the demographic data in YouScan
b) Second check: reviewed sources, Twitter was pulling in most of the mentions. I reviewed hashtags and added irrelevant ones to the exclusion list.
c) Third check: a manual review of the April and May mentions, excluding more keywords that pulled in spam mentions, a lot of them related to the two wars (Russia-Ukraine, Israel-Gaza), as well as some political spam hashtags (#trump2024) or games-related keywords.
This is one way of doing things. I am sure the other contestants found other ways of approaching this challenge, better ones. I’m constantly learning so this one right here helped me go back to the drawing board.
This competition was a great way to learn about different methodologies, grow the community of social listening professionals and talk about the work that we do and how we do it. I believe it is an important step towards educating our clients and our peers about the value social listening can bring to our marketing efforts.