WASHINGTON, Dec. 29 (UPI) -- Whatever your view of Wall Street analysts, there's no question that some are better than others. StarMine Corporation tracks the earnings and stock prediction capabilities of Wall Street's analysts, and adds value for institutional investors by doing so.
When Wall Street analysts make a forecast on a company, of the next quarter's earnings, the next year's earnings or the stock rating, it is collected into a database by Thomson Financial, and produced on the "First Call" system. First Call issues a "consensus" estimate on the next quarter and next year's earnings, based on analyst forecasts received during the last three months. This "consensus" is then regarded as Wall Street's "base case" prediction for the company. If the company's announced earnings beat the "consensus," the stock price rises, if they miss "consensus" it drops.
Since Regulation FD, propounded late in 2000, prevented companies from giving privileged access to information to particular analysts, companies whose earnings are expected to outperform or fall below the current "consensus" significantly have taken to issuing "guidance" in mid quarter -- which in turn causes analysts to revise their forecasts and the "consensus" to move.
First Call's "consensus" rating is however un-weighted; it takes no account of the particular analyst's past track record in assessing the company in question, nor of how recent is the analyst's estimate.
StarMine, founded in 1997, is a computerized service that rates the analysts, looking at the past accuracy of their earnings records and stock ratings, and assigns them a "star" weighting according to their past performance on the sector in question. StarMine then calculates quarterly earnings estimates, "SmartEstimates," based on a weighted average basis, where the analysts with the highest "star" rating are given the most weight, as are the most recent recommendations, whereas old recommendations are rejected from the sample.
The StarMine earnings estimates can, by comparing them to the First Call consensus estinmates, be used to predict "earnings surprises" -- cases in which a company's earnings exceed the "consensus" estimate -- with an approximately 70 percent success rate.
The "StarMine Indicator" rates stocks on a 1 to 100 scale, based on their likelihood of producing an earnings surprise for the current quarter and the next 12 months.
StarMine's ratings of analysts can be used by investment banks to track their own analysts. This is useful since the securities industry's settlement with the New York regulators requires analyst compensation to be determined according to their analytical success, objectively measured, not according to corporate finance business gained.
StarMine also alerts institutional investors to any "bold" (substantially off-consensus) earnings estimates issued by 5-star analysts on stocks in the institution's portfolio, or on its watch list.
StarMine currently has around 160 fund manager clients, including 26 of the top 100 firms; it also supplies analysts' ratings to 17 sell side and 6 buy side firms, covering about 1,500 of the 6,000 analysts included in the Thomson worldwide database.
StarMine's primary products are provided by electronic data feed, but the company also operates a retail investor-oriented website, www.starmine.com, offering free "Earnings surprise" alerts and information on recent "bold" earnings estimates issued by selected 5-star analysts.
I spoke with Joe Gatto, founder and CEO of StarMine, not only about StarMine's service, outlined above, but also about how he got the company started, with no substantial initial capital.
Q: You started with a couple of engineering degrees and no money, right?
A: And I've still got a couple of engineering degrees and no money, because the business hasn't gone public yet! (I of course have shares which will presumably be worth something.)
I got an engineering degree here in DC, at Catholic University, and then I worked at the Naval Research Laboratory, trying to work out how to map 7-dimensional data on a 2-dimensional screen -- so I got into the question of the visual representation of quantitative information.
I then opened up a new chapter by going to Stanford and studying in a program called "Engineering in Economic Systems" -- focused on decision analysis, probability theory, and optimizations, giving a set of quantitative tools, like the math you learn in electrical engineering, that can be applied to all kinds of interesting business problems such as yield management for Hertz, or transportation planning for the government of South Africa.
I was then a management consultant for five years, three years with Strategic Decisions Group and two years with a company I co-founded called the Decision Company. We helped really big clients -- oil companies, pharmaceutical companies, Top 50 companies -- to make billion dollar decisions. When you're talking about drug development or oil field development, there's no market study you can do; you need to look at the breadth of options and the breadth of uncertainties, and come up with an answer. It's a different approach, more of a comprehensive than a narrowing analysis.
Around this time, after 5 years, I'd paid off my debts, made some money and thought "How can I apply what I know in the world of the stock market -- naïve little me -- and see if there's any way I can make money in the stock market with my above average amount of math? At the same time, I was humbled by efficient markets, and so on, and realized it was a difficult undertaking -- there are professionals out there, so let me tread carefully here.
I wanted to come to something that's fundamentally defendable, so I started looking at valuation models, price-earnings ratios, and came to the question of which "earnings" you should use. Forward earnings made more sense -- you're looking through the windshield, not the rear view mirror. But of course we don't know forward earnings. I had this example of Dell Computer in 1996 -- with maybe 20 analysts, the estimates for next year's earnings ranged between $4.14 per share versus $1.98. I thus realized that if I could take a directional bet that say the $4.14 estimate was from an analyst with a good track record, then maybe I would know something that wasn't in the "consensus" estimate of around $3 per share.
I then talked with I/B/E/S (International Brokers' Estimate System) -- the rival to Thomson Financial, now taken over by Thomson Financial -- which at that time had the best detailed history data, analyst by analyst, got some sample data and it looked interesting. I tried to talk with some money managers, but I really didn't know what I was doing here, not being from the securities industry, so I looked them up in the Yellow Pages and didn't get very far!
Then through I/B/E/S I attended a quantitative money manager conference, and started talking up the idea and eyes popped open -- all the feedback that a clueless but aspiring entrepreneur wants to hear! At that point, January 1997, I quit my day job!
By mid February I had a business plan. I shopped it around, and got some yawns from West Coast venture capital guys. However, one potential investor said he'd invest if I could get letters of interest from potential customers. So I went back to the attendees at the conference that I'd attended, put 20 calls to answering machines, got 12 callbacks and, after sending out a PowerPoint presentation, got 2 letters of intent to purchase a product if we could develop one.
By October 1997 I had obtained funding of $250,000 from investors from my previous consulting firm -- the "angel" round of financing. That $250,000 lasted 14 months, if you can believe it -- we're talking shoestring! I implemented the first Microsoft Exchange server that we had; we worked out of our homes until we got to 3 people, and then we got this humble office south of Market Street (in San Francisco) which we're finally moving out of now into much better space, for which we're paying only 50 percent more rent for 5 times the space -- tells you what the real estate market's done since the dot-com crash!
During this period we perfected our first product, went to some more conferences, and sold our first product, which was just a tool to allow investors to mine the I/B/E/S data themselves -- we were not in the business of answers, just tools, at that stage. Quantitative analysts are tough to sell to -- they always want new features, they ask you tough questions to show you how smart they are, and they don't want tools, they want the answer.
So after a time we developed SmartEstimates, and started developing the StarMine Indicator. We signed the first major sales client in spring of 1999, got a couple of million in venture capital in July 1999 from Hummer Winblad, then after a little over a year American Century Investors led the next round and finally Instinet led a financing round that closed in January 2002, which was the last round we needed to take because we're now cash flow positive. A little under $15 million all told through 3 professional and one "angel" round.
Revenues are up 5-fold in the last 2 years, and we have presences in Australia, South Africa and Europe. We're thus pretty unique in the analytic space, in providing both "black box" indicators and software tools on an integrated basis.
Q: Plans for the IPO?
A: You typically need $30-40 million in top line and $10 million in cash flow to go public, so that might be a 3 year plan, or another company may come along and acquire us.
