In recent months internet social network companies like Facebook, Twitter and social gaming firm Zynga are valued worth many times their expected revenue this year. With Facebook valuation being $50 billion, twitter $810 billion and Zynga $79 billion![i]
Facebook’s implied value has risen fivefold since mid2009, it is really intriguing and exciting to find out what made Goldman Sachs to value, a firm whose business model is unproven, worth more than established media giants such as Time Warner (see chart).[ii]
In light of these investment opportunities, it is very important to answer questions like: How to value social networking sites? Are there any relevant models? How can investors save themselves from overpaying for investments because of faulty valuations? Etc. This report tries to answer these interesting questions from the valuation world.
The article explains the inadequacies of traditional valuation models in detail with suitable examples. Traditional valuation models like DCF analysis, cost approach; market valuations etc. are applied with whatever data is available in the market about Facebook. The inadequacy of traditional models, learned through academic courses, will be explained by showing invalidity of underlying assumptions in the case of social networking sites.Finally the future outlook and forecasting is done through extensive secondary market research.
Standard Valuation Techniques:
The following traditional methods are available to value any companies:
Now let us assess the relevance and importance of each of these methods for the valuation of Facebook.
The Market Value:
In this method the market value is found by:
Total Market Value = Share Price * Total outstanding market shares.
But since Facebook is not Public company we don’t have data on its share price as yet.
However, Facebook’s shares are being traded on some secondary markets like sharespost.com. Most of the investors are High Networth individuals (HNI) and institutional investors.
The recent data on this trading gives following information about Facebook value:[iii]
Recent Contracts 



PRICE($) 
IMPLIED VALUATION($) 
TYPE 
DATE 
36 
84726000000 
Common 
07/25/11 
35 
82372500000 
Common 
07/14/11 
The drawback of this method is lack of sufficient liquidity in the market. And the value of the firm is decided by few contracts which are done by HNIs and Institutional Investors.
Liquidation value
It is the amount one can get on liquidating all of the Facebook assets (both tangible and intangible. But most of the Facebook’s assets are intangible assets for example, profile information of its users which are very difficult to value. But value of intangible assets will be reflected in high revenue Facebook will get by charging high prices for advertisers. Because for them user profile information is very important given the scope of target advertising.
Comparable value
Selection of comparable companies:
The following companies are selected as comparable companies for Facebook.
Company 
Business 
Why selected 

Most of the business is from search engine 
Source of revenue is similar (online advertising) 
Baidu 
Search engine in China 
Source of revenue is similar (online advertising) 

Social Networking site 
Similar revenue model although different user demographics 
Yahoo 
Online Portal , email service 
Source of revenue is similar (online advertising) 
Now consider some valuation ratios for comparable companies in Q1 of 2011:
Valuation Ratios as of March 2011[iv]
Ratios 

Baidu 

Yahoo 
Weighted Average 
EV/EBITDA 
13.79 
66.47 
254.16 
11.1 
152.2785 
EV/REVENUE 
5.39 
37.39 
32.92 
2.85 
30.228 
Price/Sales 
6.29 
37.35 
32.87 
3.19 
30.296 
P/E 
24.01 
84.43 
1183.56 
15.78 
624.5205 
Growth Rate Q1 2011(%) 
26.57 
80.22 
110.06 
23.25 

But which valuation ratio should we choose, for Facebook valuation? Can we choose average value?We shouldn’t select average value because comparable companies are on different growth path.
[v]
As shown above, ratios like Price/sales are high in the initial years of these companies to take into account future high potential growth. Also denominators like sales & revenue are also low, requiring high ratios. But as companies grow older their valuation ratios become lower because of low potential growth.
Now Facebook with an estimated growth rate of 100%, it makes sense to use valuation ratios of companies which are in the initial stages of growth like LinkedIn and Baidu. But can we take them directly? Or should we take average?
Taking weighted average with following weights:
EV/REVENUE of Facebook: 0.10*5.39+0.35*37.59+0.5*32.92 +0.05*2.82 = 30.3
The estimated revenue for Facebook is calculated in DCF analysis.

2010 
2011 E 
Revenue 
1983 
3640.06 
EV= (EV/REVENUE)*REVENUE 
59942.124 
110031.7 
Similar EV calculation using other ratios is shown in Exhibit 1
Discount cash flow approach:
Estimating discount factor
Excel file ‘Discount Rate’shows the calculation of discount factor.(Excel File attached in Exhibit a) )
I have used normal CAPM model wherein both daily and weekly share prices of comparable companies are linearly regressed with the corresponding index prices for the same duration.
The beta factor for different companies is as follows:

Daily Regression R square 
Weekly Regression R square 
Selected R Square 
Corresponding Beta 
Company 





0.506330247 
NA 
0.506330247 
0.941981284 
Yahoo 
0.000906708 
0.339667364 
0.339667364 
1.078715513 
Baidu 
0.000906768 
0.072252968 
0.072252968 
1.301120125 

0.001427371 
0.36064684 
0.36064684 
3.400069085 








Weighted Avg. 
1.970761204 
Since both Google and yahoo are growing at much lesser growth rate than other companies including Facebook; it doesn’t make sense to use their low beta values as an estimate of Facebook. Hence both have given weights of 0.15 and 0.05 respectively.
Whereas LinkedIn and Baidu are growing at reasonable faster rate but Facebook growth is faster than both of them. But beta value calculated for LinkedIn is from only 10 weekly data points hence a lower weight of 0.35 is given to reflect this. But weight for Baidu is 0.45 because of its high growth rate.
Now Risk free rate, approximated by US treasury bill 3 month rate is very low, given the economic scenario.
Risk free rate: 0.02 %
[vi]
Also the risk premium, according to analyst’s estimate is 5.70%[vii]
Discount rate 
risk free rate+ (risk premium)* beta 
11.2533388641554000% 

Approx. 
11 
Estimating growth
To estimate the future growth rate, I have analyzed the growth pattern of comparable companies since their inception. It was observed that Natural Log (Revenue) follows a pattern for almost all the selected companies. (Fig 1)
Rev. = Revenue
Historical revenue data[viii]
The pattern comes out to be:
Clearly there is some trend in In (Revenue) of these companies. The growth rate of In (Revenue) is large in initial stages but slowly there is an exponential decrease in growth rate of In (Revenue).
Using these patterns of comparable companies we can easily predict the most probable pattern of Facebook. Then by reverse natural log we can have an estimated revenue and corresponding revenue growth.
Clearly, the possible pattern would depend upon industry and competitive environment. Hence we are better off doing scenario analysis, in which different patterns are estimated and then NPV value for each scenario (Pattern) is found out using DCF analysis. Finally the weighted average of these NPV values is taken to get a final NPV.
To demonstrate how to calculate NPV from a pattern, let’s have most likely pattern of Facebook based on Google’s pattern.
The extrapolated pattern is:
Corresponding Revenue is:
Thus we have assumed that Facebook will reach its near steady growth rate in 2018 at 4%.
Now let’s calculate the NPV of Facebook using these values and discount rate. We can make a template for the same so that any changes in underlying assumptions can be immediately incorporated.
Valuation Template:
The entire valuation template is given in Excel file. .(Excel File attached in Exhibit b ))
Let’s understand what different terms stand for and sources of their values.
Operating Margin is assumed to be increasing and will remain roughly constant at 40% as Facebook increases monetization of its user. [ix]
Conclusion and Future additions
Conclusion: So we can conclude that Facebook in secondary market is overvalued and its value should come down once IPO happens. The premium investors are paying in secondary market might be because of ability to buy these lucrative shares well before most of others.
Future additions:
Scenario analysis:As shown in Technical model, we can use estimation of different patterns (given below) to calculate the NPV for corresponding scenarios.
Scenario 2:
The calculation of weighted average of NPVs for different scenarios (with assigned estimated probabilities) can then be used to estimate the total value of Facebook.
References
[i]http://www.economist.com/node/17853336
[ii]http://www.economist.com/node/17853336
[iii]https://www.sharespost.com/companies/facebook/valuation
[iv] http://ycharts.com/companies/GOOG/market_cap
http://www.wikinvest.com/stock/Yahoo!_(YHOO)/Data/P:E
http://finance.yahoo.com/q/ks?s=GOOG
[v] http://ycharts.com/companies/GOOG/ps_ratio#compCos=BIDU,YHOO,LNKD
[vi]: http://www.bloomberg.com/markets/ratesbonds/governmentbonds/us/
[vii]http://seekingalpha.com/article/263536surveyingusequityriskpremiumin2011
[viii] http://www.wikinvest.com/stock/Yahoo!_(YHOO)/Data/Revenue/2000
http://investor.google.com/financial/2008/tables.html
http://accounts.icharts.net/myicharts/app?service=external&sp=Y3jXySNC&page=TeamChartDetail
[ix] http://www.allfacebook.com/everythingyouneedtoknowaboutthegoldmanfacebookdeal201101
http://www.marketwatch.com/story/facebooksprofitabilitytopsyouthfulgoogles20110120
http://www.secondshares.com/2010/03/25/facebook50billionvaluation/
This article has been written by Nitin Bandgar from IIM Bangalore