Tuesday, 18 September 2012

Inside Bloomberg's secret financials

Blind men and elephants

As I'm wrapping up my week-long series on Bloomberg, I just wanted to put up a little bit of analysis on Bloomberg's financials, as I haven't seen anyone else do this before.

As I said in my first post on the subject, Bloomberg are notoriously cagey about their financial data. It's interesting that their official About page tells you lots about what they do, but very little about the company itself.

But occasionally you get little peeks under the Kimono. For one tantalising year Merrill Lynch (at the time a 30% shareholder) disclosed more detailed financials about Bloomberg in their annual report (albeit aggregated with a couple of other equity stakes - see p116). A Fortune article just before the credit crisis got access to revenue numbers (which by and large corroborate with the ML data) and made some guesstimates about profitability. And there were a few more tid-bits in this Vanity Fair piece. More recently revenue numbers (which seem to be quite widely circulated internally) have come out as scuttlebutt in the NY Post (note I do not generally use the Post as a source for investment advice!! But needs must...).

Overall its like blind men feeling the elephant.

Anyhow, here's what the elephant looks like from where I'm sitting:

Note - Table updated 20/09 (see comment below). Original available here.

The small print

Okay so a few obvious disclaimers.

  • A lot of this is based on assumptions and educated guesswork. I can't claim the source data is consistent or, for that matter compliant with Generally Accepted Accounting Principles. (So its basically as reliable as an average Wall St Analysts' estimate then! :-p )
  • Cells highlighted in orange are ones when I have had to plug in a blatant guesstimate - normally revolving around details about profitability. Generally I have tried to interpolate the data I have, or extrapolate from the most recent hard numbers.
  • In contrast cells highlighted in blue are ones where I have reasonable confidence about the source data - either becomes it comes from more official channels or because a couple of different sources triangulate together.
  • As I mentioned, the biggest issue is about figuring out Bloomberg's profitability. As I said they sometimes open the kimono about revenues but have NEVER talked about profitability. The only hard data we have is the ML data for 2005-07. This is cleanest in 2005, partially impacted by the consolidation of a Blackrock stack in Q406 and wholly impacted by this in 2007. I have tried to strip out the impact of this stake. From that I can at least figure out opex /employee and use that to make educated guesses about margins in other years (number of employees is ocassionally disclosed).
  • Most of the data generally looks sensible, with the exception of the jump in revenues between 2007 and 2009. Partly this would be due to the acquisition of Businessweek, but given revenues for this unit were only about $130m that can't make up all of the increase. Maybe there were other acquisitions bulking it up, or maybe there was just a big bounce back in 2009. [EDIT 20/09 - found some more reports talking about Bloomberg's 07/08 revenue growth so I feel more comfortable with these numbers now. Have updated the table to reflect this.]
But even with these issues I think the underlying shape of the beast is clear. You have a $7bn+ revenue company running at an operating margin above (and most likely significantly above) 30%. It employees 15,000 people and has over 310,000 users of its terminals in circulation.

Not bad for a cloud company. Remember cloud poster-child Salesforce.com is worth $21bn on a mere $2.2bn of 2011 revenues (albeit growing at a stonking 37%). Salesforce has 8,000 employees and reported zero operating profit (because two thirds of revenues were spend on sales costs to drive that stonking growth). As CEO Marc Benioff wrote in the annual report And I’d like to thank all of you, my fellow stockholders, for your confidence in salesforce.com as we strive to reach our goal of becoming the irst enterprise cloud company to achieve $3 billion in annual revenue.

Someone clearly forgot to tell him that Michael Bloomberg got there eight years ahead of him.

Analyzing Bloomberg

Putting my analyst hat on, a few things occur to me:

  • Customer concentration is a risk. With only 310,000+ individual users at a rev/user of $24,000 its extremely concentrated. Note that when Lehman went down they had over 3,000 Bloomberg terminals, roughly 1% of Bloomberg's entire revenue. Even a Amaranth, a mid-sized hedge fund, had over 200 terminals. And of course the fortunes of all those customers tends to be extremely correlated. Although Bloomberg seems to have grown through the crisis (helped by acquisitions) it certainly must have been dicey at times.
  • But customer concentration also helps margins. However there's also a bright side to that. Sales and marketing costs are normally a significant part of the cost base for enterprise software companies (20% of revenues at SAP, 22% at Oracle and of course a whopping 67% at Salesforce.com). For someone like Bloomberg however it will be a lot less intensive - the larger customers like investment banks will already be fairly captive, and only need a finite number of S&M resource. In contrast smaller start-up funds will already know that they need a Bloomberg. I would wager there is some cost involved around marketing new services and offerings, but I doubt it will be anything like Salesforce.com levels.
  • Bloomberg should be highly profitable. Mature cloud computing companies are some of the most profitable businesses out there. Like transaction processors, one you have dropped the initial capex building the data centre (or renting it from Equinix, as Bloomberg do), the marginal costs of running the business (both in terms of P&L cost and capital intensity) are very low. As one of the more mature cloud vendors out there I don't think Bloomberg will have any problem producing cash margins of over 30%.
  • Bloomberg has a huge cash pile: Even if you assume Bloomberg were paying a full whack 30% tax rate on their profits (its seems to have been lower in the 2005-07 disclosures) they have generated nearly $12bn of cash over the last ten years (if you assume net income = cashflow, which is a reasonable assumption for a software company. Note the last ten years analysis - this is the time in which Michael Bloomberg has been in politics and has had his Bloomberg stake locked up in a blind trust - i.e. he wouldn't have been taking out dividends on that. Less whatever M&A Bloomberg has done, it should all be sitting on the balance sheet. This gives them enormous strategic flexibility.

Valuing Bloomberg

To finish I just want to update the valuation of Bloomberg I put up in my first post.

To recap I'm not doing anything fancy - I thought about using a bunch of carefully picked comps and/or extrapolating the relationship between software company P/E ratios and operating margin. But that's all a bit complicated (remember more complexity can often mean less accuracy not more. Occams Razor FTW!).

I simply assume a Price/Sales multiple of 10x the operating margin (more explanation on that here) and a trailing P/E ratio inline with the S&P 500. I would wager good money that Bloomberg's growth prospects and cash-rich business model makes it a higher quality than the average S&P 500 stock, but lets be conservative. I'm not trying to nail down precisely how much Bloomberg LP is worth, but trying to make the point its worth a helluva lot more than you think and that is why you should be aware of it.

Similar to what I wrote before, Bloomberg is worth something in the region of $27-31bn Enterprise Value. But what I forgot to add in was the value of the net cash sitting on the balance sheet - even assuming a 30% tax rate on historic cash generation that gets us to a total Equity Value of more like $38-42bn.

Interestly Facebook's currently worth $48bn.

So Bloomberg isn't just the biggest cloud company you've never heard of. It's one of the biggest cloud companies - period.


  1. Bloomberg has a very "un-official" policy of raising prices every 2 years. the next increase for 2012 has already been announced.
    The 2007 - 2008 increase in revenues (in a decreasing termianl sales environment , most likely was due to the introduction to all the terminal base of its new prices.

    Reuters used (maysbe still does) inlation-linked contracts. It's normal practice in the software world yet Bloomberg users seem to forget and are always shocked when the new price increase is announced. Puzzling indeed.

    On a separate note, I strongly doubt they posted a +16,000 net termianal sales in the period 2009-1010. I would review net terminal base downwards by at least 10,000 units.

    For all your other assumptions...I agree

  2. Heya!

    Thanks for your comments. Very interesting. Yes I always thought Bloomberg were a bit slack on pricing (SAP are similar - they have historically had an inflation kicker in their contracts but not enforced it). I remember back when I started in the City a BBG terminal was about £1000 /month. It doesn't actually seem to have changed that much in the last decade!

    re: the 09-10 terminal sales there may well be some discrepancies (so hard to model a company without having full disclosure!). Businessinsider reported 286k 08 terminals. The NYT reported a 2% volume decline in 09. The NY Post (I know, I know) reported 13672 net adds to 310,000 in 2011. So yes I am unfortunately working a system with multiple potential points of failure!

    On the positive side it has dawned on my that I might be able to calculate the increase in Bloomberg's book value (i.e. its retained earnings) from changes in the book value of the Merrill Lynch equity stake pre-07 (provided there aren't too many other investments lumped in there). Which might actually make it possible to calculate Bloomberg's RoE. Let's see!

  3. Bluehost is definitely the best hosting company with plans for all of your hosting requirments.