Twitter and market research: demographic dangers
June 8th, 2009 | 
Author: 
Verify user demographics -- one tweet only
Slate today has a story on the topic of “orphaned tweets” — user accounts that consist of one tweet only. While Swansburg and Singer-Vine take the article in a hilarious direction by classifying the various kinds of orphaned tweets, the article points to a critical statistic relevant to managers considering social media market research.
A major Harvard study reveals that 10% of Twitter users are responsible for 90% of the site’s traffic. Add to this an older Nielson study revealing that Twitter has a month-over-month retention rate of just over 40% and the prognosis for creating a wide societal groundswell on twitter is bleak.
This creates an imposing challenge for those looking to create valid demographic samples from Twitter. One of the most popular objections to SMMR is that it is disproportionately limited to certain demographics.
The relevant question is whether this difficulty will be limited to Twitter or if it creates problems for the idea of SMMR itself. My answer has been that it is indeed possible to capture more diverse demographics, but there are three additional considerations.
1. Twitter might not be the best vehicle for capturing insight. Given the low insight available in 140 characters and the difficulty understanding demographics, other social media sites may be more appropriate for market research. However, if Twitter continues to grow and gains mainstream traction, it’s search functionality will make it an invaluable tool.
2. Use SMMR before target demographics are finalized. Smart market research practitioners will integrate social media market research into their research plans in a way that leverages it’s advantages — casting the widest possible net for insights. This knowledge should help inform the thinking as you progress to tighter demographics.
Properly understood, social media’s difficulty in expensively isolating demographics is in fact an advantage.
Posted in demographics, John, market research, Twitter | 
Tags: market research, social media, Twitter
I like looking at Market Research insights in two dimensions
#1 – Breadth of the Need (How much of the population has it)
#2 – Intensity of the Need (How “strong” of a need is it)
I think Twitter can help a lot with #1 – Breadth but not so much with #2 – Intensity. You can use Twitter to find out a lot about breadth. Once you have that, follow up with some in-depth SMMR to understand the details behind the intensity.
I’m not sure I would argue that Twitter can actually offer breadth though. It can offer a slice of the population but in general you have no real idea as to who these people are other than their location, maybe their vocation, and possibly their gender. What slice of the population do they represent? Probably late 20 to early 40 somethings in general. But most likely, based on my experience, people who live on the web all day and probably heavily overpopulated by software / web people, and PR, Marketing, and Consultants who cater to these people. There’s no question a certain percentage include bloggers of all stripes and hucksters using the latest shiny megaphone to broadcast their message – but it still is a fairly narrow strip of the population. You can easily find a lot (tens, hundreds, thousands) of people discussing certain topics or brands, but how broad is their representation? Are they truly lead users of the product / service area you focus on?
Josh Bernhoff coined the phrase, “Objects in the groundswell may be smaller than they appear.” And by that he meant, “people who congregate online are not a representative sample.” http://blogs.forrester.com/groundswell/2009/04/what-should-nbc-and-subway-do-about-the-save-chuck-campaign.html
In my mind, John has it right on this one.