Photo by Katrina.Tuliao & Blogtrepreneur
Young adults are increasingly wary of stock market investing – and considering how many came of age in the wake of the financially devastating Great Recession, this should come as no surprise. Of course, while it’s true that there’s inherent risk in investing, it’s also an invaluable tool for those looking to build wealth, and today’s interconnected and data-driven world has brought us tools that are uniquely suitable to the way 20-and-30 somethings process information.
It’s that concept that led to the formation of BUZZ Indexes, a company that seeks to utilize the big data created through social media to identify investment opportunities in the market. According to BUZZ Indexes, online discussion and social media posts related to finance and investment have increased upwards of 500% in the last three years, and their team uses this info to offer financial advice. Buzz Indexes has even created an ETF that’s based on this methodology of aggregating the social media investment sentiment of various companies.
So how did BUZZ Indexes and the NYSE/ARCA listed BUZ ETF come about? Check out our interview with their founder Jamie Wise below to learn more about the company’s data-driven financial strategy.
EQ: What exactly is BUZZ Indexes and where did the idea come from?
Wise: At BUZZ Indexes, we use artificial intelligence and natural language processing techniques to analyze millions of conversations, whether they are social media posts or tweets or blogs, to identify what people are saying around individual stocks and their investments in general over various social platforms. The goal is to identify the sentiment behind those conversations, and try to find insights from those conversations that can be predictive of future stock returns.
If you think back to companies like Gap Inc. (GPS) or Procter and Gamble (PG), they were, at the time, very actively listening to the conversation in places like Facebook (FB) to try and understand what their consumers were saying about their products and brand, and allowing them to make better product decisions or better marketing decisions. But there was a really big change in 2013. That’s when Twitter (TWTR) adopted their cash tagging methodology, where you put the dollar sign before the stocks involved. That allowed people to find a natural forum to talk about whatever stocks they were interested in, spurring an explosion of conversation online, where, all of the sudden, thousands of people were talking about their investment portfolios.
That was great news for us, because now we could monitor a couple of things. First, we had a much broader universe of stocks to look at. Generally speaking, the bigger the stock, the more widely held the stock, the more likely it is that people would be talking about it. There was this consistent flow of data around large-cap US equities that allowed us to measure sentiments around those companies, not only looking for changes or a positive sentiment, but just being able to measure baseline sentiment levels.
EQ: You’ve utilized pretty complex analytics and algorithms to filter data and pinpoint which stocks people are talking about, and what that conversation really means for market activity. Can you walk us through how those algorithms have evolved, and what makes them so unique?
Wise: We’re very fortunate in that we’re living in a world now where processing power and processing speed has increased dramatically in the last couple of years. Machine learning allows you to train models in much the same way that you’re trained as an investment analyst. When you read a blog or a news article or even just a tweet, you understand the context of what is being said, because you are familiar with the way people speak about stocks. You’re familiar with slang, you’re familiar with the nuances of the conversation. For a human being, it’s quite easy to be able to read something and understand if someone is positive or negative, and to what degree.
You can apply that same logic to computer programs and algorithms. By training a model, and through vast quantities of data, you can allow it to learn from experience. You can give it scenarios where you can help it to score baseline models, and you can train it that way, and then, by feeding it huge amounts of information, you can fine tune a model. It can actually fine tune itself to get better and better at understanding tone, context, sarcasm etc. when reading through things.
EQ: I understand BUZZ Indexes also has its own ETF. How did that come about?
Wise: Yes, the ticker symbol is BUZ, and it is listed on NYSE/ARCA. We created an index based on sentiment data, and we designed the ETF to track the performance of our index.
The sentiment that we’re trying to measure is not very short term. Instead, we look at sentiment from a longer-term perspective. We think about it much like an ocean has a natural current to it. If you come to look at any stock, and you can observe the conversation flow that is happening around that stock, what we’re trying to identify is the waves building in that ocean.
So, just like a wave will build and have momentum and carry forward before it crests, we look at sentiment the same way, and we measure the momentum of that wave in weeks, if not months.
EQ: What’s the process for choosing which stocks are included in the ETF?
Wise: There are 5,000 or so stocks that trade on US markets and people are talking about all of them to varying degrees. We start off and take that universe of 5,000 stocks and distill it down to the most talked about stocks. We take that list and distill it down to a list of about 150 or so names. The most talked about stock list is dynamic and is updated every quarter.
We only look at large-cap equities with a market cap of $5 billion or higher, and we do that for a couple of reasons. First, we want to make sure that the underlying stocks that we’re tracking are liquid and can support the product. Second, it’s a safeguard that you could have against potential bad actors online trying to influence stock prices. It’s much less likely that some of them will be trying to promote a stock with ill intent online if it’s a large cap, very liquid, very talked about security. By focusing on large-cap equities only, it allows us to make sure we know real people are talking about that stock, and we also know that our index is investible.
Once we have identified the 150 or so of the most talked about large-cap US equities, we rank them based on sentiment. We rank them from most positive sentiment to least positive sentiment, and each month we take the top 75 names most positive names for inclusion in the index.The stocks in the index are weighted according to sentiment with each stock subject to a maximum weight of 3.00%.
EQ: Your site makes mention of “BUZZ Influencers”, the people who are influencing the direction of stocks and where they’re expected to go moving forward. How do you decide who constitutes an influencer in this space?
Wise: There’s tons of data. Using Twitter as an example, you’re able to identify who was saying it, how often they talk about stocks, how many followers they have, and how their followers react to their posts. Is it retweeted often? It’s basically a measure of how fast some person’s comments can flow through social networks and perhaps influence other people. We track all that in our analysis, and we can also track forecasting accuracy.
EQ: How have you found the concept of influence – or influencers – has evolved in the age of social media?
Wise: Influence is a really interesting topic today, because we have a president who likes to tweet about stocks. You would think that there would be nobody more influential than Donald Trump. One thing that we found in our analysis over time, even before Donald Trump, is that the importance of influence changes over time.
We first started seeing this effect with Carl Icahn when he started tweeting about Apple (AAPL). It was shocking. It was new, and it certainly moved the stock. The same thing happened when Trump started first tweeting, whether it was about Boeing (BA) or Lockheed Martin (LMT) or General Motors (GM). Some of those early tweets certainly had an immediate impact on the stock.
Now, everyone is finding that when either Carl Icahn talks about Apple or Donald Trump says something about Harley-Davidson (HOG) for example, there’s much less of an impact when they say it. There’s a reason for that – we call it diminishing returns of influence. Now that we have either Carl Icahn tweeting all the time about Apple, or the president about various stocks, the shock factor wears off. What it does do however, is serve as a catalyst for a much broader conversation amongst investors.
So, if you have Trump talking about GM, the opportunity doesn’t lie in trading off of Trump’s tweet or being able to analyze it very quickly and get into the market based on if it’s positive or negative. The real opportunity is that in him talking about a particular stock, he sparks a much broader conversation which can be mined for insight.Thousands of investors are encouraged to go online and comment and share their own views, engaging with one another and sharing their own personal sentiments on that company.
EQ: How is the BUZZ ETF performing at the moment?
Wise: Performance has been really good. A lot of people will naturally compare us to the iShares Edge MSCI USA Momentum Factor (MTUM) platform, which is a price momentum ETF. I think we have beaten that index by over 5.00% since we launched. We have clearly demonstrated that sentiment momentum, captured in the BUZ ETF can outperform a passive price momentum strategy.Over the last six months, we’ve also beaten the S&P 500 index. The performance has really been a great proof of concept in validating our strategy, demonstrating that these sentiment insights can actually lead to better performance.
EQ: That’s great to hear. What are the plans for BUZZ Indexes moving forward? Any significant advances on the horizon?
Wise: We plan, at some point, to launch a broader suite of products that use these types of insights. We are always asked where there is a version of BUZZ for shorting stocks, because people talk negatively about stocks online, too. All of the work we’ve done so far has shown us that if we focus on the negative sentiment the portfolio not only underperforms the long sentiment BUZZ, but also underperforms the market. There are definitely opportunities there.
I think we have opportunities to do things that are sector-specific. The consumer sector is one that naturally stands out, that you could do a sector focused product launch or even a commodity or macro-themed approach. There are a lot of people discussing currencies and interest rates and commodities online. There is, of course, insights to be had from those conversations.
EQ: Before we end, is there anything else you wanted to share about BUZZ Indexes?
Wise: The whole mission of our business here is to really democratize access to technology. We all know that the world’s biggest and best hedge funds use these alternative data sources in their quantitative investing strategies. We think the time is right for us to bring these same concepts to market and allow all investors who are quite used to not only living their lives on social media, but benefiting from insights of the crowd on social media. Whether you’re trying to figure out where to go on vacation, what restaurant to eat at, or something to buy, we all look at product reviews and TripAdvisor reviews and Amazon (AMZN) reviews. We have recognized that there’s a lot of value if you can crowdsource that information and make it predictive for you, specifically. You have the opportunity to do that now with your investment portfolio, and we’re excited to be able to deliver that.
