By Doug Friedenberg

If you’ve ever stood near a tree full of birds at sundown, you have some idea of the enormous amount of chatter occurring as the birds discuss their day and opine on the comfort of their chosen branch.   Although you might not think so, there’s likely to be some real information buried in the chatter.  As we recall, the best way to decipher bird chatter before computers was to listen to early Pink Floyd with pharmaceutical assistance.

But that was then, and this is now.  Amid the millions of tweets  from tweetyhumans that occur every day in the Twitterverse, there’s more than a little substance.  The trick is to screen the noise to reveal the substance underneath.  Which is where Topsy Labs comes in.  Their platform, www.topsy.com, is a leading realtime social search engine.

Our readers know that alpha is often found in the company of those investors who are earliest in seeing the first hints of future events unfolding.  Topsy may be the newest vision aid.

Topsy’s  Window into the Almost-Future

You can get a taste of Topsy by looking at AllAboutAlpha articles and clicking the number that mentions re-tweets.  There you will find a mini-universe of people tweeting article references. This universe has a hierarchy of influence, with some Twitterers ranked as highly influential, some merely influential, and quite a few others who exist but whose tweeted article references were noted numerically only.  Closer inspection will reveal that Topsy also features recent tweets of particular interest, and breaks the current crop of Tweets into various categories of higher interest.  And that’s the barest scratching of the surface.

Rishab Aiyer Ghosh, a co-founder and the chief scientist of Topsy Labs, explained some of the inner workings.  Since Topsy’s purpose is to screen out noise, it was somehow fitting to learn from Rishab that Topsy had been founded by people who were anti-spam.  He made the key point that a significant amount of the information that used to be derived from publications and newswires now comes from conversations on social networks. He thought this to be the most significant leap forward in news delivery since the advent of the news wire services. As Rishab put it, “Everybody in the world can be a reporter now, and our technology functions as editorial and quality control, and provides the metrics.”  Topsy’s algorithms process some 250 million social media posts every day and, in Rishab’s view, are effectively a coherent and relevant new stream.  We immediately decided this article had best be written quickly before Alpha readers started checking Topsy instead, although we knew that you, dear reader, would still have to figure out what to ask.

Rishab spoke of one of the more dramatic examples of recent note, back in early October when Apple announced the Iphone 4S.  Many Wall Street analysts became negative on Apple prospects and Iphone sales.  Topsy research, to the contrary, revealed that consumers were positive about the iPhone 4S the day it was announced, suggesting pent up demand.  Their algorithms were more directly in tune with what consumers were actually doing, as opposed to what analysts surmised consumers would do.  They got it right. Apple announced record sales two weeks later, leading to a 20% increase in the share price.

As Rishab noted, Twitter is excellent for on the ground information, providing real-time responses on the part of consumers, subject, of course, to Topsy’s filtration systems.   This is relevant when everyone can speak in the twitterverse, even if no one in particular is listening.  Within that world, Topsy is providing data quality control and metrics.  They are finding out consumer reactions not by a laborious interview process, but by being in position to catch a decent-sized random sample of humans chattering about the latest thing that happens to catch their fancy.

So Now That I Have This Pile of Steaming Data,  What Do I Do?

That question was the genesis of this article.  We were speaking with Erin Collard, a portfolio manager at Armored Wolf (a principal of the firm, John Brynjolfsson, is a frequent contributor on these pages).  Erin also happens to be an advisor to Topsy, charged with sorting the mass of filtered data into meaningful, usable information.  The advent of Google all those few years ago brought to the fore a skill that humans always had, even if they didn’t use it much.  That skill is the ability to ask questions.  Even if you know that you can find out anything you want, you still have to devise the magic question.

Erin’s role as an adviser to Topsy is to determine how best to coax meaningful information out of Topsy’s processed social media data.  As noted, Topsy provides a pretty clear way to perceive early response in the consumer sector.  Erin noted that apart from the Apple case, they have been able to get some clear sense of what was going on with Black Friday, as one example.  Anecdotal evidence, he also observed is there is more tweet activity among farmers than one might have thought, and he had the sense that there might be some clear information to be derived for the commodity markets as well.

Erin thought that a second, broader area of relevance was in a larger, geopolitical sphere.  He mentioned the signals they had gotten on the Arab spring, comments echoed by Rishab.  Erin went on to observe that there had been significant Tweeting about mortgage-related banking problems in the early stages of the housing crisis.  There was on-the-ground evidence of the mounting foreclosure crisis.   

A New Form of Technical Analysis?

It seemed to us that there were some parallels between the work being done by Topsy and the chart-based technical analysis which is a mainstay of Wall Street analysis particularly as regards trading and timing.  Erin thought this was partly true, although he thought that Topsy was predictive more on the behavioral side.  Topsy could “see” certain specific bits of human behavior regarding specified product introductions, for instance, from which an investor might infer a subsequent action in the stock price.

Nonetheless, there are other similarities, as Erin agreed.  Just as technicians screen for a rise in stock volume as an indicator of meaningful market action, Topsy’s search engine is aware of rises in word volume as certain topics become of greater significance, and volume breakouts in certain topics.

For the universe of investors, Erin thought that Topsy’s information was not necessarily useful for the extremely long term and short term investor.  However, for the investor with a horizon of a few weeks to a few months, he felt that some very good data is generated.

We wondered if it would be possible for people to game the system.  After all, there is some reason to believe that the Tea Party in the US was a bit manufactured, rather than being the spontaneous uprising that was sometimes reported.  Erin’s assessment was that there was a way, but it would require a concerted effort by hackers engaged on a campaign of what might be called social identity theft.  Something like that might overload servers and cause data distortions.

We also wondered if Topsy had similar data availability in emerging markets.  Erin thought that the flow of raw data was not yet sufficiently robust to get a handle on most matters of investor interest.  Nonetheless,  a data eruption such as occurred during the Arab Spring would provide a large enough sample to get a clear picture.

Erin summed up their efforts by saying that his focus with Topsy was to develop simple ways for investment managers to derive usable data from the system, but it would be driven largely by investors’ specific information needs.  He mentioned that research products for the investment community would be released this year.  It seems to us that the possibilities are startling.  Investors should be able to monitor immediate consumer response to meaningful new product introductions. You would have had a terrific heads-up when the Edsel came out, when New Coke hit the market, or other products from the crypt like Harley Davidson perfume or Bic brand underwear made their brief (no pun intended) appearances.

We don’t recall where we heard this story, but it has a bit of a point.

One day a wise man told his young disciple:  “Take this chicken, and go kill it where no one can see.”
The disciple came back several days later with the non-deceased chicken under his arm.
“Why didn’t you kill the chicken?” the wise man asked.
The disciple replied, “Well, everywhere I went, the chicken could still see.”

Today, “chickens” everywhere can see, and they can tweet, even if, as we hear, they generally use the hunt and peck method.