Apple, Microsoft & Facebook Bring More Traffic To In-House CDNs, Impacting Akamai’s Media Business

Yesterday, Akamai reported Q3 earnings and announced that revenue from their media delivery business would be flat or down for Q4, year-over-year. For Q3 their media revenue was up only 5% year-over-year and the company said, “traffic and revenue growth slowed considerably in some of our largest media accounts.” Following guidance that was well below expectations, Akamai stock dropped $9.76 a share in after hours trading. It was an odd earnings call as Akamai suggested the reason they expect media growth rates to continue to moderate in the “near term” was due to customers having “less traffic growth overall”. Except that’s not what’s happening.

The cause of what Akamai is seeing is a result of Apple, Microsoft and Facebook moving a larger percentage of their traffic to their in-house delivery networks. This is a trend that all three companies, especially Apple, have been doing for some time, but in the past 45-60 days, Microsoft specifically has taken a lot of their traffic in-house, which is lost business for the CDNs. Akamai said the impact of what they are seeing is “expected to be magnified by their do-it-yourself efforts” but then said “most of the impact is from less traffic growth overall”. Apple, Facebook and Microsoft aren’t seeing “less traffic” and certainly aren’t seeing lower traffic growth rates overall, compared to the past few quarters. So Akamai’s explanation really makes no sense.

All you have to do is trace where their content is being delivered from to see less of it is coming from third-party CDNs. Akamai said they,” believe kind of the bigger slowdown in traffic here is that traffic overall is slowing.” And why do they “believe” that? What data do they have to show that Apple and Microsoft’s overall traffic is slowing? I haven’t seen any such data in the market, from third-party companies that track traffic growth amongst consumer services. And when Akamai was asked why customers traffic is slowing, like Akamai claims, they responded to the question by saying, “why it’s slowing for our customers is difficult for us to access.” Confused yet?

In the Q&A potion of the call, Akamai said, “if the overall traffic is less growing at a slower rate and less than expected well, there’s a tendency to fill up the do-it-yourself effort first, and then we would get the remainder which leaves us with even slower traffic growth.” Earlier Akamai said the overall traffic was slowing, but then later they say “if” it was slowing, and when they say “less than expected”, less than who expected, Akamai or the customer? Because if Akamai simply overestimated the percentage of traffic a customer who owns their own CDN is going to give Akamai, that’s not a “slower rate” of traffic growth, that’s just bad estimating on Akamai’s part.

Heading into 2016, Akamai expects media growth rates to continue to moderate in the “near term”, which isn’t good news for them, especially considering how much they continue to talk about the impact OTT services have on their business. One thing many may not realize is that while OTT services are growing, the ones that are seeing the most growth, don’t have their content being delivered via third-party CDNs. Akamai also said it was “worth noting that media pricing overall has continued to decline at normal historical levels”, but they didn’t say what those levels are or give out any numbers. For most customers, that is true, and the decline in pricing is stable, down about 20% this year. But for the largest handful of customers, who push the most traffic, pricing is down more this year than last, in the 40%-50% range.

Akamai also said that, “competition in the media business remains constant but is not expected to be a significant factor in our traffic and revenue estimates.” This also is suspect as the term “significant factor” could mean a lot of things and doing some traceroutes on those who have lots of traffic growth, shows competition impacting Akamai. To highlight one example, looking at Sony’s traffic for the PS4 now shows that Level 3 has been added as another delivery network, in addition to what Akamai and Limelight Networks are already delivering. So while Level 3’s share of traffic may not be taking away traffic Akamai and Limelight already have, it is less traffic they would have gotten, with Level 3 now being included.

Note that when Akamai says “media” they don’t necessarily mean “video”. Media includes video but also non-video content and software downloads. So there could be other media customers involved, that aren’t video related, that is also impacting the growth of revenue from Akamai’s media customers. The company also said they have, “purposefully slowed down the rate and pace of head count additions and discretionary spending to align with our near-term top line growth expectations,” which isn’t surprising.

One thing Akamai is still doing well is using vague and high-level statements, without actually saying anything. The company mentions their new deal with Microsoft Azure saying, “Microsoft sales force is also planning to sell Akamai’s market-leading acceleration and security solutions.” While that is true, what they don’t say is “when” Microsoft plans to start selling it. Because when I asked Microsoft, they said they would only be re-selling Akamai’s CDN services and would not be re-selling value add services any time soon. Akamai also said their, “overall media traffic is still projected to grow at a substantial pace” but again, substantial is a useless word without definition.

As usual Akamai also went on to mention how many “edge servers” they have, which is a useless stat since we all know that capacity and performance is measured by more than just the number of servers a CDN has. No customer buys CDN services based on a server number from a vendor. Some parts of Akamai’s business are still being marketed like it’s 2002, when customers focused on server count and the simultaneous number of streams a CDN could support. Akamai also called out their, “superior communication and video transport protocols which are designed to deliver the kind of higher-quality picture that is expected by users and broadcasters alike”, except that Akamai had problems delivering the 60fps stream for Yahoo’s NFL stream this past Sunday.

Akamai needs to focus more on data, numbers and giving the market real insight into their business, as opposed to lots of generic, vague and high-level words and phrases. At the end of the call Akamai said, “media growth rates are going to be effectively flat to up very significantly, or down very significantly.” Flat, up, or down. There aren’t any other options.

  • iwod

    That is no surprise. I could never understand why something as ( relatively ) simple as CDN could not be replicated. All the major cloud services, be it Alibaba Cloud, AWS, Google, Microsoft Azure will have their own CDN, and they are selling it as a package with their compute services. Netflix, or other On Demand TV are all using their own CDN.
    Hence CDN company by nature should and could only scales up to small to medium level.

    Apple, with its iOS, OSX, App Store downloads and update, Apple event streaming, Apple Music. Apple should have started its own CDN long ago, and not the only rolling out this year. ( Again it is their conservative nature, and doing it right and good, although my opinion is that they were late and not so good with Cloud / CDN delivery.)
    I specifically left out Apple TV because I think Apple are moving to a App / TV Category where the content provider source their own delivery. I hope Apple actually provide a delivery platform for these content, scale to a level that would benefits them.