Warren Buffett, 'Sage of Omaha' and world-famous investor, is no fan of the wisdom of crowds: 'A public-opinion poll is no substitute for thought.' Perhaps with good reason - under his leadership, holding company Berkshire Hathaway has grown its share price from $66 in 1970 to more than $100,000 in 2011, which is probably better than if he'd listened to us lot.
Yet evidence is mounting that there might just be something in what people think after all, as a number of studies are indicating that social media is a lead indicator of share-price performance.
Before looking at these studies and their implications, let's address why I'm writing about this in Marketing. First, because if you're struggling to get the top floor to stump up for social-media activity, the likelihood that it might influence the company's share price should make them more amenable to overtures of a budgetary nature. Second, because increasing shareholder value is what we're all about; it doesn't get more important in corporate terms.
Players of the stock market are always seeking bellwethers: magic indicators that give a sign of things to come and a lead on rivals. So it comes as no surprise that buzz monitoring has been seized upon as a potential forecasting tool; people who 'like' a brand, become a fan, discuss it in forums, tweet or blog about it are expressing their interest and approval in dozens of ways.
Plenty of black-box services claim to follow this commotion with varying degrees of accuracy, but the question is whether a statistically significant relationship exists between the stock price and the buzz that's monitored.
Management consulting group Altman Vilandrie conducted a proof-of-concept study in the US, looking at the iPhone 4 antenna problem in 2010. It found that social media was a good proxy for consumer sentiment and acted as a lead indicator both for the stock-price decline and news sentiment; it was more reliable than online news sources, which lagged and then tended toward sensationalisation.
Meanwhile, a 10-month study was taking place, ending in February 2011. A researcher at Pace University in New York collaborated with Famecount, a social analytics company, to examine the relationship between the stock prices of Coca-Cola, Starbucks and Nike and the number of fans they had in social media. All three brands were reported as displaying a significant correlation: importantly for stock-pickers, that relationship worked on a 10- to 30-day lag of stock price to social media, making it a useful predictor.
Comments on these findings focused on whether the relationship was causal or correlational, but one pertinent comment from a trader stood out: 'Any relationship at all that is statistically reliable is going to make me money.'
Now, start-up market-information service Sntmnt.com, based in the Netherlands, has taken the logical step of applying machine learning to pick out the correlations between its 10,000 buzz sources and the stock prices of 25 funds on the Amsterdam exchange. Early results from its beta claim 56% accuracy in its predictions; and it's this artificial intelligence-based approach that is likely to gain ground.
Moreover, as marketers, there's a further point: these things become causal by their nature, regardless of their validity. If the market starts to believe that buzz is a predictor of stock performance, then whether this is real or not ceases to matter; buzz spikes will start to be reflected by jumps in stock prices as traders follow the metrics.
Before we get too excited and start making bets, though, we should remember one other Buffettism: 'Beware of geeks bearing formulas.'
Andrew Walmsley is a digital pluralist
30 SECONDS ON ... STOCK-PICKING
- A stock's value is represented by the discounted value of its future cash flow. While markets have intrinsic values, prices can fluctuate above or below these, due to a range of factors.
- Not averse to data-gazing, stock-pickers assess performance using a variety of tools, including FTSE graphs, Bollinger Bands, Relative Strength Index and Fibonacci Retracements.
- Investors can use a top-down or bottom-up approach to choosing stocks. This equates roughly to the 'outside-in' (analysing global and macro-economic factors) and 'inside-out' (internal metrics) of a business.
- Those investing in BP would rue the day they had, following a catastrophic spill from the Deepwater Horizon oil rig in April 2010. In its biggest one-day shares-fall for 18 years, £12bn was wiped off its stock-market value.
- The best long stock-picker for Japan in July 2011 was Citigroup Global Markets' Mattia Ciancaleoni, who achieved a relative return of 16.11% versus the MSCI Japan index in FactSet's AlphaStars competition.
- A popular category is green energy, especially wind and solar companies.
- The most profitable stock in the US last year, according to NASD100.com, was Food Technology Service, which gained 52.9%. Meanwhile, eBay gained 27.4%, The Coca-Cola Company 20.7% and Diageo 18.5%.