Mary Meeker's YouTube Math

From Silicon Alley Insider: Morgan Stanley's Internet analyst Mary Meeker was a good deal more optimistic than we and most others about the revenue impact of YouTube's new overlay ads.  Specifically, Mary concluded that the overlays could immediately add $4.8 billion of gross revenue and $720 million of net revenue to Google's annual results.  This compared to the tiny $12 million to $360 million of gross revenue that we projected.

Well, we were baffled at how Mary could be so amazingly bullish, so, on a tip from a reader, we checked her numbers.  And, Mary, it may be time to scream at yet another research assistant.  Why?  Because, in advertising lingo, "CPM" means "Cost Per Thousand" not "Cost Per One."  When Mary updates her model to divide by 1,000, therefore, we expect she will wish to revise her conclusions.

What happens to Mary's estimates when you do the math right?  Well, that $4.8 billion of gross revenue becomes $4.8 million, and the $720 million of net revenue becomes $720 thousand.  So if, as Mary suggests, Google can float ads on top of 20 million streams a month, secure a $20 CPM, and keep 15% of the gross revenue, the overall impact will actually be, as we suggested yesterday, immaterial.

Here's Mary's note (PDF) Meeker's YouTube Math.  Here's a page laying out Mary's math and the correct math.  And here's our own YouTube math.

(Update: And let he who is without sin cast the first stone... As Charlie Wood points out, the original version of this post had its own numerical typo. Mary, I'm feeling your pain!)

"We estimate that Google will generate $100 trillion of revenue in 2010. Or maybe $10 billion.  Whatever."

Comments

[Charlie Wood]

Mr. Blodgett,

Ms. Meeker's math may be off, but it looks like she's not alone. According to my calculator, one thousandth of $4.8 billion is $4.8 million, not $48 million. Then again, what's an order of magnitude between friends? Dude, this is Google!

Regards,
Charlie

[wren]

I guess video-on-demand really doesn't scale. $1.6B too late for GOOG.

[Neal S. Lachman]

Amazing! For Mary Meeker's sake, I was hoping you were wrong. So, I took the time to read through her report and checked out what her fundamental research pointed to. But Henry, you are right (as almost always).

This is a blunder of the first order. Did MS and/or Mary correct this, or did they offer a rebuttal?

[kumars]

Including cost of delivering those streams, this is what it may look like:

Using Henry's data, to generate $400K/mo:

monthly video streams: 2,000,000,000.
assume duration of video streamed on average: 6 minutes
assuming 300kbps, amount of data transferred a month =
(2x10^9) X (300kbps) X (6x60 seconds) / (8x1000x1000) GB = 27,000,000 GB.

for breakeven, cost of delivering 1 GB has to be $400k/27000000 = $0.014 or 1.4cents/GB.

Market rates for BW are atleast 15X the above number!

GOOG's internal BW rates may be lower but still...

(i) Users are going to want better than 300kbps - so data demands will increase over time and youtube will have to keep pace.
(ii) Bandwidth costs do not scale down like processing/storage costs do.
(iii) CPMs are not going to increase much more than $20. With fragmentation of online video market, possibly CPMs may go down.
(v) Calculations above dont take in account that 15% of revenue goes to partners but they dont share the BW cost.
And there are lots of other operating costs....

[greentea]

Also, Google's *net* revenue is not 16.4B, that is the gross revenue. So the percentage computation is strange - the net YouTube revenue as a percent of the gross Google revenue. Had she used the real net number for Google, the percentage would have been even more absurd.

[Neal S. Lachman]

Kumar,

First, I believe Google only keeps 15% and 85% goes to the content partner; you got it the other way around.

Second, you are right about bandwidth requirements. This is something that not only Google, but all online content aggregators/distributors, will face. The solution for Google is to own their own backbones, which will cost billions of dollars, or to partner up with companies that own broadband systems and backbones.

[Tessinge]

Great story, Henry. My experience in the internet biz is that there is quite a bit of diversity in revenue and cost metrics, but CPM is what it is!