If you bought LinkedIn (LNKD) before they announced quarterly earnings on Thursday February 9, you made 18% in a single day. News outlets were abuzz with how the company doubled its subscribers in a year and beat analyst revenue estimates. To be fair, you probably didn't buy LinkedIn but alas, how could you have known that they would release such stellar numbers?
Actually, you could have - all you needed was a little math and a basic understanding of LinkedIn's business. Let's start with two dirty little secrets you should know about the finance biz:
- Companies lie
- Analysts are lemmings
Companies lie. Fine, maybe they don't straight-out lie exactly (unless you're Angelo Mozillo), but they definitely shade the truth, especially when it comes to guidance for the future. Especially for companies who are still growing, there's not much upside in being completely honest about what they expect next quarter. If they say exactly how much they think they'll earn and the economy turns downward, it's an "earnings miss" and their stock gets hammered. So instead, growing companies are conservative with guidance, sometimes overly so.
Analysts are lemmings. Do you know where Wall Street analysts get their quarterly estimates of how much companies will make? From the companies themselves. A company is usually covered by about 10 analysts, each of whom makes $500,000 or more a year. If an analyst significantly deviates from the pack and is proven wrong, there's always a hungry junior associate ready to take their spot. There just isn't a lot of upside in being aggressive with estimates, unless the analyst has a truly compelling insight.
So with that introduction, let's rewind the clock and imagine that you are looking at trading LinkedIn on Thursday February 4, before their earnings report. From a macro perspective, you know that:
- The biggest part of LinkedIn's revenues comes from employers who pay to post jobs.
- Economic indicators show that U.S. economy is in recovery mode and, most importantly, companies are hiring.
- There aren't any popular alternatives out there yet (otherwise you'd have heard about it).
You also have access to LinkedIn's past results and the guidance they published in the 3Q 2011 earnings report. In fact, they even put it in a nice, easy-to-read presentation.
Finally, in high school you probably learned a technique called statistical regression that can be used to predict what happens based on past trends.
Since we know that nothing in the market happened to negatively affect LinkedIn's business model (and the spike in hiring may actually have helped it), let's apply a simple regression model to LinkedIn's 2010 and 2011 (through 3Q) numbers to try to predict its 4Q numbers.
Members: LinkedIn's member growth fits an exponential growth curve very well, with R-squared of 0.9994 (This is just a fancy way of saying that the line is close to what the actual points were. The closer to 1 the better). The equation shown predicts that LinkedIn's 4Q members will be 147 million.
Revenues: Similar to members, LinkedIn's revenue growth fits an exponential equation well. Prediction for 4Q revenues based on this equation: 171 million.
Adjusted EBITDA margin: This is LinkedIn's percentage of Adjusted EBITDA relative to their revenues each quarter. Adjusted EBITDA is LinkedIn's internal proxy for the profit they actually made, so this margin can be thought of as their profit margin.*
A linear regression here isn't terribly informative, since the margin has varied a lot. However, note that the quarter with the lowest margin (14% in Q1 2011) was right before the company filed for IPO, when it may have had larger IPO-related expenses. In any case, the equation predicts margin rate for 4Q is 17.6%. Multiplying that the predicted revenues number of 171 millions leads to Adjusted EBITDA of 30 million.
*I know adjusted EBITDA is LinkedIn's own non-GAAP measurement and I should probably use standard EBITDA instead. Since the only real difference was their inclusion of stock compensation in adjusted EBITDA, it doesn't make a big difference to the prediction.
So let's compare our basic regression model's prediction versus analyst estimates and LinkedIn's own guidance in 3Q 2011:
Members: The model's 145 million prediction was hypnotically close to the actual 147 million number.
Revenues: Compared to actual revenues of 168 million, our model's 171 million prediction did a lot better than LinkedIn's guidance of 154-158 million. As for the consensus Wall Street analyst estimate of 160 million, it's amazing how close they are to LinkedIn's guidance, huh? Lemmings.
Adjusted EBITDA: I have to tip my hat to LinkedIn here - they definitely had a great quarter from a margins perspective. While 34 million looks great compared to their 19-21 million guidance, our model was pretty decent with 30 million. And if you take out the 14% IPO quarter outlier and reran the regression, our model would have predicted 32 million. Not too shabby, huh?
So with a little math and common sense, you could have predicted that LinkedIn would probably exceed expectations from a numbers standpoint. But does that mean that their stock would jump 18%?
Not necessarily ... LinkedIn also vastly exceeded its overly conservative guidance in its 3Q 2011 earnings announcement on November 4th, with 139 million revenues (versus 121-125 million guidance) and 25 million adjusted EBITDA (versus 9-11 million guidance). However, its shares opened 7.5% lower the next day!
Was there a risk that the same thing would happen again? Of course. However, overall market sentiment has improved markedly since last November, and the Facebook (FB) IPO halo still envelopes social media stocks. Betting that the market would react differently this time would have been a smart play, in hindsight of course.
Look, I have no idea if LinkedIn stock is a good buy or not. It certainly isn't cheap from a valuation perspective, but social networks are indeed changing the world rapidly and LinkedIn is the only player in its market right now. However, I will definitely be looking at a short-term trade the next time it releases earnings in April.