Where is market research reporting, top-line writing etc in the area of AI and generative text?
AI it seems can do a decent job in many areas of text creation - type in the challenge “explain the principles of qualitative research” to ChatGPT and it produces pretty impressive output. No idea how many languages it manages well, I only checked English and German - however, both were written clearly, the content was pretty well informed, it was well structured, with numerical bullet points.
It was written like a text book.
Moving to the other extreme, Chat GPT leaves much to be desired in the area of creativity. I typed in “write a paragraph in the style of Ernest Hemingway” and it churned out a lot of tosh. This was the first sentence: “The sun scorched the arid landscape as the lone figure trudged through the desert”.
Having just re-read The Sun Also Rises this gets a poor rating, language and style wise way off.
WRITING FOR INSIGHTS… LIKE A HUMAN - HOW?
Where do “normal humans” fit in between text-book correctness and creative aspirational? And market researchers in particular, whose job is to make sense of data and help it ideally come alive for people making a business decision.
A few observations from recent reviews of:
Market Reserach conference submissions
LinkedIn
Comments sections of mainstream media.
Starting with LinkedIn - a business portal. Peter Totman recently picked up on the over-use of phrases on LinkedIn such as “super-excited” or “amazing”. I agree with his short piece wholeheartedly. Jargon and laziness is killing the impact of business language.
Many contributions I read on LI are salesy, but samey. Including those “this is my struggle and how I came out stronger” stories which quite often smack of “curated honesty”, with an undertone of sales-intent.
But for many more simply “celebratory posts”, there is very often a lack of anything original on the emotional side. Lots of hugely positive but totally generic sentiments accomanpanied by smiling faces. Sales wise I’d say that’s sub-optimal.
AVOIDING JARGON
Businesses projecting confidence is understandible - but when it’s jargon-plastered, thumbs-up cookie cutter stuff, lacking a credible human touch, or is transparently one-sided, with a poorly concealed ambition, it jars - and can be forgotten quickly.
Moving on quickly to conference submissions - two things that struck me from reading 50 or so recently.
Firstly, a lack of clarity was surprisingly frequent, where it wasn’t clear what the submitter wanted to say. Complex thoughts that hadn’t been crystallised into simple sentences with a logical structure. Time is short - moving on is tempting.
Secondly, the overuse of business jargon as per above, but also hyperbolic tendencies or over-claims without a sense of the substance underpinning them. “Real-time insights” for example. “Future proofing” is another. There are plenty more buzzy-like words or phrases that just don’t resonate unless they are supported with evidence - “overarching trends” might make sense if it’s contextualised, but not if simply thrown into a sentence.
Finally, on comments allowed by e-newspapers. These are often the opposite of business-like language, have a huge spread, from the polite, well-articulated to those containing **** signs redolent of strong emotions and suppressed expletives. Whilst often containing typos, they are however emotional. Very human.
WRITING FOR EMOTION IN RESEARCH
Where do we in research hope to emerge in our insight-driven narratives in an age of ChatGPT, business jargon and heated public debates?
I’d say our territory is in the human. Our language needs to reflect that. To explain:
The MR submissions that caught my attention, and stayed with me the most were those which were personal, honest, simply expressed. This has immense value - clarity allows a reader a quick access to what the author wants to say, so is a great time saver. And they were free of jargon,.
Strangely they were also the most original - individual experiences are fascinating, and at the heart of what we observe in research.
And being so individual they are not something that ChatGPT can replicate. Yet.
My call to action? Market researchers looking to make their read-outs or toplines more impactful can maybe vary between the learnings of ChatGPT - clear structures, zero typoes, quickly generated concise texts - and the emotively engaging stuff we read (and hear) from participants.
And mix it together with your own personal voice - where’s the problem ;)
Now, back to the Hemingway challenge….
Curious, as ever, as to others’ views.
GET IN TOUCH!
P.S. I’m available for help on storytelling or writing of all sorts - polishing a submission, writing a topline, ghosting a blog, even helping finding a story in various data sources. Get in touch if you’d like to chat at eappleton90@googlemail.com
(Visual: Photo by Brett Jordan on Unsplash)
Nice one. All the examples you share remind me of something my own experiments with GPT etc ... the lack of "ORIGINAL VOICE" ... saying , writing, proposing with some originality.