The potential business benefits of micro-blogging site Twitter have been undermined by new research which found nearly half of all tweets are "pointless babble".
US-based market research firm Pear Analytics studied 2,000 tweets from the public timeline over a two-week period, capturing tweets in half hour increments.
The firm then categorised the content of the tweets into six buckets: News, Spam, Self-Promotion, Pointless Babble, Conversational and Pass-Along Value.
The firm classified Self-Promotion as any corporate tweets about products and services, Pass-Along Value as any re-tweeted messages, and Pointless Babble as the "I am eating a sandwich now" type tweets.
The results put Pointless Babble as the largest category with 40.55 per cent, Conversational a close second at 37.55 per cent with Pass-Along Value a distant third with 8.7 per cent.
Only about five per cent of the tweets studied came from businesses trying to market or promote products or services, while a little more than three per cent came from news.
This tends to pour cold water on suggestions that Twitter can be used as an effective business tool and news source, or as an efficient way for firms to engage with customers.
"Many people still perceive Twitter as just mindless babble of people telling you what they are doing minute-by-minute; as if you care they are eating a sandwich at the moment," wrote Pear Analysis founder Ryan Kelly.
"With the new face of Twitter, it will be interesting to see if they take a heavier role in news, or continue to be a source for people to share their curr ent activities that have little to do with everyone else."
PUBG news and updates: November's Update #23 to bring new Skorpion pistol and changes to blue zone visibility
Genuinely useful side-arm coming to PUBG in Update #23
Asda, Morrisons and Tesco in the frame for checkout facial recognition technology
Research opens up new possibilities for structural batteries, where the carbon fibre forms part of the energy system
Another shape could have indicated hard-to-detect particles