The Future of the CDN Market

Pricing Pressure Is Exaggerated
I continue to see a lot of analyst reports still talking about the "pricing pressure" in the CDN market. But I still don’t see anyone giving examples of what the "pressure" is. I think it is important to keep that in perspective as many in the market make the pricing declines sound bigger than they actually are.

As an example, while some public companies mentioned pricing pressure with some of their largest customers during their latest earnings call, afterwards they told me the pricing reduction for these customers was typically fractions of a penny on per-GB delivered contracts. Fractions of a penny are not a huge drop in the market. And customers who are already pushing a ton of volume and who continue to grow their traffic expect to get some discounts based on growth.

The lowest price I saw in 1Q 2009 was 2.5 cents per GB delivered for more than 500TB of traffic a month. When I questioned many of the major CDNs about this price, nearly all of them told me they don’t price delivery that low––but the contracts say otherwise. That price is not the norm, as 500TB a month in delivery is a very large customer. But there are now more than a handful of those customers on the market. Bitrates are larger and the video is higher quality (some is in HD). The major content owners are pushing some ridiculous volumes of traffic as well. That said, my guess is that the 100 top content owners probably push more than 60% of all the video traffic among all the major CDNs. So the CDNs are still getting the vast majority of their video-related revenue from probably 20% of the customers in the market. This will shift over time; it’s already beginning to. But the bottom line is that the major content owners are still the ones pushing the vast majority of traffic among the CDNs.

While it looks like pricing dropped by 50% at the 500TB level, keep in mind that there are very few deals that are that size. I saw only two deals in the 500TB range for all of the first quarter. The vast majority of deal sizes are in the 250TB range, which remained pretty stable (see Figure 3).

For many content owners, traffic is not growing as fast in 2009 as it did in the first half of 2008. But it’s still growing. That’s not to say all content owners are seeing slower growth; many––such as Netflix, MLB.com, Hulu, and others––are seeing a lot more growth already in the new year. However, since the CDN vendors won’t tell us how much video traffic they push quarter to quarter and what kind of traffic patterns they are seeing, it’s too hard to make a general statement on what the overall market is doing. No doubt the traffic growth is down overall, but the percentage compared to previous years can only be compared on a customer-by-customer basis later in the year.

Many people are still asking me about Akamai potentially cutting its CDN pricing for video in the hopes of getting its media and entertainment business to grow again––I don’t see that happening. I do, however, see Akamai doing more noncommit contracts with content owners, which is great for the customer but bad for the CDNs. Even content owners who aren’t that big––for example, those doing 150TB of delivery a month––have contracts with Akamai in which they don’t commit to any traffic on a monthly, quarterly, or yearly basis. While this is something Akamai and many of the CDNs have done in the past, I saw Akamai doing more of this in the last 2 quarters. I think this is a sign that Akamai is trying to be more flexible, which is good. But its pricing on most video deals I have seen is still coming in at more than 35% higher than Limelight or Level 3.

In 1Q 2009, I saw more contracts from companies based outside the U.S. in London, India, and Singapore. Clearly, the pricing environment is different in those regions. It is also clear that many times content owners do see a difference in performance in places such as India and Singapore among Akamai, Level 3, Limelight, ChinaCache, and Tata Communications. Akamai’s strength still lies in its performance in these regions of the world, and I do see some content owners paying more for that service. There is nothing wrong with Akamai charging more in the market, as long as it can show a measurable difference, which it seems to be doing in India and China. It’s really in the U.S. that Akamai is charging more for video delivery but not showing the difference in performance, so it is giving up a share of the market to Limelight and Level 3.

At the Content Delivery Summit, I was asked many questions about content owners using dual vendors and if that is a shift taking place with customers. While some content owners have used dual CDN vendors for years, many others don’t. It really does come down to the preferences and needs of the customer. Some want two or three CDNs; some tell me they will never use more than one. So I don’t see a trend taking place either way, but I do see more of the largest content owners such as Apple, Netflix, MLB.com, NFL.com, Comcast, and others using two major CDNs today.

As I have written before, I don’t expect to see a big decline in pricing this year. The CDNs all know that they can’t give this stuff away; they have to have some profit margin. And if pricing declined 35% for the average customer last year, I think we’ll see only about a 15% decline this year, on average. The content delivery business is all about the economies of scale. CDNs have to multiply the volume of traffic on their network many times over before the next round of major pricing discounts can take place. I think it will be the first or second quarter of 2010 before we once again see a big drop in pricing across the board.