Wednesday, October 30, 2013

Linkedin Q3 earnings results

This post is part of a series of documenting the eventual implosion of Linkedin's stock price, updated on a quarterly basis.  Last quarter's post

Q3 results

  • It lost $3.4M in the quarter.  About what it guided last quarter.
  • While it beat its own revenue guidance by $20M, it provided no benefit to its income.
  • Page views and visitors dropped despite a 17M increase in members
  • Its slideshare subsidiary had an even sharper drop in visitors and page views..
Q4 guidance
  • It is guiding only 40% increase in revenue over last year for the last quarter.
  • The lowered guidance from Q2 for Q4 is restated.
  • It expects a drop in page views and visitors again.
  • It expects another loss in Q4
  • Full year earnings per share will be lower than last year, despite a >50% full year revenue increase.
  • It will have increased total diluted shares to over 120M, from 112M last year (~8%)
Speculating on why it remains unprofitable no matter its sales.
  • Many of its new members may be ghosts.  
  • Lawsuit this quarter alleges what is already widely known:  Linkedin hacks into your contacts and spams (sends) invitations on your behalf to everyone there.  
  • The spam can trick children into signing up with some difficult red tape to removing the account.
  • It has lowered its member standards to target 14 year olds so that they may link with colleges.
  • Even with its poor reputation for email ethics and privacy and systems safety, it has launched a mobile email service that grants linkedin full access to all of its users incoming and outgoing email
  • That it beats its own guidance for EBITDA but never for income effectively proves that it is either padding its revenue, or just pays its employees larger stock bonuses as a function of that increased revenue.
The shady marketing practises exposed this quarter would be a likely explanation for an increase in members without an increase in visitors and pageviews.  Arguably, assuming the 17M new members all were unique visitors during each month, then linkedin lost 17M unique visitors from its previous members.

Revenue per user dropped to $1.52 from the previous quarter.  That further supports the premise that they are running out of marketable new users, and perhaps just padding their user numbers with shady practices.

Market cap of $30B is impossible
120M shares at $250/share is not sustainable, because it presumes an evental $3B annual profit.  It previously stated its addressable member market as 500M members.  It is over 50% penetration.  While many of its members view it as legitimate as viagra penis enlargement ads, the first 50% penetration is necessarily easier and more potentially profitable than the last 50%, because the first 50% include those that find the service obviously useful.

There is still no evidence that the recruiting fees market worldwide is over $4B in sales.  This also acts as a significant barrier to its future earning potential.  Until proven otherwise, its sales potential should be limited to 2 or 3x current levels.  Even if it stopped paying its management so much, and tripled its non-GAAP profit, as profit available to shareholders, $600M per year, is $5/share in EPS.  A $50/share price target.

For the first time in their conference call last night, they mentioned thinking about the full 3B workers in the world as opportunity.  They are not making a profit on the core 250M people that currently find their service worthwhile enough to sign up, and are generally (higher paid) knowledge workers.  Theoretically these members should be more profitable than the marginal new member.  They currently earn $6/revenue (annual) per member, and 0 profit.  Acquring 3B users will be a significant cost, and continuous reinvestment of any contributions from profitable members into storing new accounts that are likely to just take up server space, and be annoyed by emails filling their viagra box.  It expects $135M in depreciation costs this year.  This is mostly computer and technology infrastructure costs.  Increasing users 10 fold will increase those costs 10 fold beyond LNKD's already steep user acquisition costs, and by doing so, it will arguably be chasing users that are even less profitable than its existing user base.  Furthermore any future reduction in costs/performance of future computers is offset by the database operation (computing time per operation) costs that will typically rise 100 fold with a 10 fold increase in database size.

EBITDA and NON-GAAP earnings is not a useful metric
Financial analysts that place any weight on non-GAAP earnings or Earnings Before Income Tax and Depreciation for linkedin make a grave error.

Unlike capital intensive companies such as airlines,  mineral extraction, or car plants, which generally pay a large amount to start a project whose costs get depreciated slightly each year, Linkedin's depreciation expenses is for ongoing computer upgrades and expansion, and these costs are generally depreciated very quickly, and in some jurisdictions are allowed to be expensed.  Unless we expect it to stop renewing its computer infrastructure, their depreciation costs should be a core part of their operations and earnings.

Linkedin also adjusts EBITDA to exclude the significant stock based compensation it awards itself.  Almost $200M this year, leaving profit for the year of a little over $20M.  The dilution this stock compensation creates directly harms any shareholder who is not receiving compensation from LNKD.  Because of the dual share structure that keeps solid control of the company with insiders, there is no possible pressure that would change its compensation policies.

Other anaylyses on LNKD
Commenting after LNKD issued over 5M shares in a secondary offering this quarter, was this detailed long term projection for the company and a resulting company value of $12B and $106/share.

Most of the numbers are quite fair and optimistic at the same time.  The one major error in the analysis is assuming only 112M shares outstanding over the next 20 years.  LNKD is already up to 120M shares, and an optimistically low number of new shares would be 1M per quarter, and that results in 200M shares outstanding in 20 years.  So a $12B eventual company value is only worth $60/share.

The other major issue with the $12B valuation is relying on operating income that excludes depreciation and stock compensation, and so for the next 5 years overestimates earnings by 7x.  So the $12B valuation relies on income that is likely to never be available to shareholders.

The major issue with social media valuations
Assumptions of $10B-$30B valuations on social media companies like LNKD or Twitter assume a world where users are monetized aggressively, and the users don't mind being monetized.  There also tends to be an assumption of a world where advertisers want to pay very high prices, and sell themselves on spending their money, and the social media company just has to staff enough people to answer phones and return email in order to collect all of that money.  Analyst valuations should be more considerate of the uncertainty of these assumptions.

A followup with Facebook (FB) confirming the above point
FB reported earnings a day after LNKD.  Impressive sales growth of over 60%, based on the strength of growth in its mobile advertising.  Much better than LNKD, However, they warned that there is not much more advertising that they can show users, and so their future growth is limited.