YouTube wildly profitable?

Froosh over at hipmojo.com argues that YouTube is "wildly profitable" and estimates the company pulls down $7.5 million a month in revenue. Venture capitalist Fred Wilson says Froosh's analsysis is not "crazy math," but I'd argue it is sketchy and decidedly optimistic math.

Let's go to the numbers.... Froosh's income statement for YouTube looks like this:

Home page ads: $175,000 a day x 30 days a month = $5.25 million in revenue
One ad per videostream served @ 100,000,000 streams a day @ a $0.50 to $2.00 CPM = $50,000 to $200,000 x 30 days = $1,500,000 to $6,000,000 in revenue
For a total of about $7.5 million a month in revenue.

Says Froosh: "Call me crazy, but that means that just with its main page alone, YouTube more than covers its bandwidth charge…"

Well, not so fast.

First, let's look at the numbers Froosh relies on. According to PaidContent, Froosh's source for the numbers on which he builds his projections, which does notYouTube is probably still running in the red.seem to have verified the financial terms of the home page placement conclusively (in comments, the reporter says it is "speculative"), the $175,000 generates about 400,000 viewers for the advertiser. This language suggests that it is may be a CPC deal, rather than CPM-based and that makes more sense since the CPM at $175,000 per day with 400,000 viewers would be $437.50, which is high-bubble pricing that, to my knowledge, no one pays for impression-based advertising today.

If it is a CPC deal and YouTube is generating one click in 50 impressions, it would take 20 million impressions to generate 400,000 viewers. Given that the site has about 40 million streams a day, it's not beyond the realm of possibility that they generate that many clicks, but one wonders whether the home page is seen even a small fraction of the time—many streams come from embedded players, so there is no actual contact with the YouTube site.

The second problem with Froosh's numbers is that he uses the figure 100,000,000 video streams a day to calculate other ad revenues. PaidContent reports that YouTube streams only 40 million videos a day—less than half the basis of Froosh's revenue calculation. And some of that inventory is presumably consumed by the front page placement (which would assume a home page clickthrough rate  of four percent, not beyond reason, but close), so let's say that YouTube has free inventory of perhaps 30 million impressions a day, not 100 million.

If that's the case, at CPMs of between $0.50 and $2.00, the potential revenue is only $15,000 to $60,000 a day, as little as $450,000 a month.

Now, let's discount all the revenue by 20 perent to account for the cost of sales. And, because no sale is unnegotiated, let's assume that many customers are paying less than the rate card. We can't guess what the actual fees are, because they are unpublished and "speculative" at this point, but we can assume that some of the days of the month are not fully paid placements on the home page, whether because of discounts or make-goods on inventory needed to meet advertiser expectations.

So, that's 20 percent off revenue for sales cost and let's say 20 percent off for discounts and make-goods, so the company is really only likely to be making $175,000 on its home page every other day or so. At a 40 percent discount for sales and discounts, the monthly revenue from the home page leaves $3.15 million a month from the premium placement on the home page.

Now, the CPM rate Froosh used for his calculation of other revenue is pretty optimistic. Much of the AdSense inventory may sell for pennies, rather than dimes or dollars, for example. I'm only going to say that a $0.50 CPM is very aggressive and, therefore, YouTube probably makes less than Froosh estimates. For argument's sake, let's agree they make the low end of Froosh's estimates, another $1,500,000 a month.

Now, it looks more like YouTube's monthly revenue is probably closer to $4.65 million. That still feels high, but I'm willing to live with it. Does that make YouTube "wildly profitable" and Froosh argues? 

Short answer: No. PaidContent does provide a figure for bandwidht used daily by YouTube, about 200 terabytes a day. I've looked into streaming and downloading bandwidth pricing and find that most providers are offering a Gigabyte of throughput for around $0.85. That would place YouTube's daily bandwidth costs at $170,000, or $5.1 million a month.

Let's be really aggressive and assume YouTube is getting a deep discount from Limelight Networks, its hosting provider (though probably in trade for equity, which investors need to be leery about, because it dilutes the value of future shares and may represent a big increase in real costs if the deal lapses). If Limelight is charging $0.45 a Gigabyte for throughput, YouTube's costs are $90,000 a day and $2.7 million a month.

But throughput isn't all the costs YouTube is paying for. Storage of all that data costs something, especially as you have to replicate data across the world to support high-performance playback. These costs are factored into Limelight's pricing. In other words, Limelight is probably charging something closer to $0.65 per Gigabyte than $0.45.

Let's revise the bandwidth costs up to account for storage and maintenance costs to $130,000 a day and $3.9 million a month.

I've never seen employee numbers for YouTube, but if it isn't 50 people I'd be surprised. At an average salary of $50,000 a year, the company is looking at about a quarter million a month for salaries. Add other overhead, such as office space, computers, accounting and marketing, and you're looking at about $400,000 to $450,000 a month in salaries and general and administrative costs.

So, finally, back to the numbers. Is YouTube "wildly profitable"? If we take the adjusted numbers and add in expenses, the answer is certainly "No."

Total Revenue is $4.65 million a month after sales costs and discounts.
Bandwidth/Hosting Costs: $3.9 million a month
Salaries, G&A and other costs: $450,000 a month
Profit = $300,000 a month or a net margin of about 6.4 percent.

That looks like YouTube may be breaking even, but we've been generous in revenue calculations and, perhaps, in the cost of people and operations. YouTube is probably still running in the red.