I’m a market researcher using an AI (askpolly.ai) for story content and validation. Data is validated through probabilistic sampling on social media. Marketers know what this means (or they can Google it!).
I have endeavored that every day I will write a marketing article inspired by current events or articles with the added benefit of it being backed by real, statistically-validated market research. To assist me I am using an AI that can produce this validated research in under a minute. While anyone can do this part (AI is very affordable these days!), you still need to know the right questions to ask that will make for an interesting story. And story ideation is often the bottleneck for freelance writers and consultants.
My inspiration today came from Kaya Yurieff at The Information who published a piece this morning essentially saying that Elon Musk should stop unilaterally deciding which features to introduce on Twitter and just ask the hundreds of thousands of people who are tweeting about this every day what they would like to see. It occurred to me this is a perfect project for Polly™ (the AI).
So first, it turns out that Ms. Yurieff is correct. According to Polly, there really are hundreds of thousands of people discussing this — almost 300,000 in fact. So much for the belief that people are fleeing Twitter in droves! And while my AI uses generative text (just like chatGPT!) to summarize all those conversations, it turns out it’s still a complicated story. In short: there is no top 5 or top 10 list that will please everyone.
When I asked Polly what people’s favorite features were on Twitter, it really came down to demographic. For people over 65, for example (yes, they are on social media!) features that allow you to block certain people or certain content from certain types of people come first. Young people, however, not only don’t like this, but they find the potential of it downright offensive and against the entire ethos of “social” media. And while you may think the solution is “if you don’t like it, just don’t use it” apparently that doesn’t fly with many in this crowd. If you start advocating for this feature arguments ensue, perhaps something purveyors of other products don’t have to worry about when they introduce new features. People on Twitter like a good debate. Are there demographic exceptions? Of course, but the divisions by age are actually quite stark for just about all the most mentioned features. For instance, young people really like the idea of audio rooms, but we see little engagement on that from older customers.
When it comes to Twitter features for reporting misinformation, a first glance at the data would seem to indicate this feature appeals mainly to people in their 40s and 50s. But when I hone in on under-represented populations, I see that it’s not the feature the detractors are opposed to, but the sloppy way in which Twitter has deployed it.
We do start to see more cohesiveness when we look at marketing tool features. Marketers love in the following order:
1. The mixed media feature
2. RSS feeds tailored to interests
3. Threads for condensed versions of blog posts
4. Features to streamline workflows and customer engagements
What unifies people most seems to be what Twitter features they don’t like:
1. Paying for features
2. Twitter rolling out features too quickly, without quality checks
3. Features that allow people to save your content and make NFTs out of it
Ms. Yurieff is correct in that Mr. Musk is able to, and should, consult Twitter customers to get a better understanding of what they are looking for (and what they will pay for) on Twitter. It literally took me less than 5 minutes to get the above insights and Mr. Musk has many more AI tools than I do, presumably! The key is getting outside your echo chamber and ensuring independence between all the people in your sample. But it’s not a matter of coming up with a “top 5” list. It’s about understanding the nuance of the conversations. Do they dislike the feature or its implementation? Is it the right feature for that audience? The nuance makes all the difference.