Saturday, Oct 22 2011

Amazon Revenue: Its All a Little Cloudy for Me

Written by
amazon jeff bezos1-225x300Sometimes Friday nights create the best blog fodder. I think a few “after work” beers and the tweets are flying thick and fast. This blog is a quick response to a tweet exchange (one hereanother here) between Brian Flannery (Eucalyptus), Randy Bias (Cloud Scaling) and myself last week. The origins of my thinking were actually primed by an article on CloudAve by Randy Bias titled ‘What is Amazon’s Secret for Success and Why is EC2 a Runaway Train?’.


In a previous blog post, “Competing with AWS”, I indicated that the future leader in the cloud space would have to consider more the sales, delivery and support life cycle of enterprise customers. The thought being that cloud was ending the period of marketing hype, grass roots adoption and the business model of the past needed to transform.

Randy’s post  was promoting the idea that AWS would be a US$1BN by 2011 and US$16BN by 2016 with seemingly no barriers to growth. So when the same comments appeared on twitter I could not help but give my $0.02.

Below is a summary of sound bites from some of the tweets with some expansion of my thoughts;

“Amazons revenue is growing 100% year on year

Although it seems AWS should be the biggest amount in the “other” revenue category of its filings, there are other businesses in that category. As an “other” category it is hard to determine how much is attributed to what. Co-branded credits cards could be a very high profit, reasonable revenue business, but you don’t know. We assume AWS is huge, because of the barrage of usage stats we have available. Its inferred, but never implicit. Making direct connections is always avoided by Amazon.

“I know people who work for Amazon and its growing gangbusters.”

No questioning this.. The infrastructure usage numbers are impressive. If you work for Amazon, I am sure you see lots of internal communication about AWS growth (boy did I us to see a lot of Exadata growth press when at Oracle) and see huge investments in infrastructure. Look at the public filings and it’s hard to believe the capital spending is all about online stores and kindles. I doubt there is much internal communication, NDA or not about actual revenue numbers, there are some serious public company issues. How do you know if all this infrastructure activity is directly attributable to revenue? The bigger the apparent infrastructure build, the more likely that profitability of the business model is low or maybe even negative. If they are making all this investment, why aren’t they explaining it to the shareholders (of which I am one).

“They have huge growth numbers in users.”

Absolutely. It’s fundamental to the business model. I have four accounts myself, and I pay Amazon $0.00/year. All trial accounts for different purposes. Subscriber growth is an old game played by Telco’s and OSS software companies and many other new economy businesses to establish dominance. Telcos keep dead accounts forever to show growth in subscribers even though churn is huge, OSS companies give away free and promote huge downloads. So subscriber number validates nothing. Show me the revenue per sub :)

“Why doesn’t Amazon report AWS revenue.”

I can only think of two reasons
  1. If the numbers are good, then big money competitors would step up the investments to buy the market. It will happen because time will validate the concept, but Amazon is trying to delay that time to give them a larger advantage; or,
  2. The margins and revenue are not good or at least not as good as the marketing and usage numbers.
Maybe Amazon has a second act. With all the hype and marketing, we don’t need to show everyone how great our business is. We are getting to share kick and the market acceptance without it. When the hype curve subsides, we will launch the figures and jump up again. Bezos is smart enough to play that game.

“I talk to large customers who are getting great advantage from AWS.”

No doubt, but are you talking to them because you want to talk to AWS customers.. Is it representative of demand?  Quantify ‘great’ advantage? It’s just a very skewed data collection and cannot validate anything. Prius is the bestselling car in America right? I drive around Silicon Valley and it seems every second car is one. Unfortunately, the bestselling car is a Ford F-150. Data really needs volume and context to have true impact.

“Current AWS business is a loss leader. The big payoff is enterprise users.”

The primary customer acquisition strategy is for free or credit card users. There are plenty of bigger users (read a Netflix), but the business model was not based around this. The enterprise sales arm is growing, but definitely not competitive with most infrastructure vendors or service providers. It does seem from the outside that enterprise customers are big part of the goal. Many of the new features are aimed squarely at this customer segment.

“Amazon is the market leader, the 1000 lb gorilla.”

In the context of public IaaS cloud providers they are definitely the market leader. In the hearts and minds of cloud professionals, they are the though leader in the same space. But their leadership depends on how you define the market. When at Sun, I used to hear repetitively the strong % marketshare of shipped UNIX server numbers. The problem was DELL and other were shipping a truckload of servers that ended up having Linux on it. AWS is competing with traditional outsourcing, traditional hosting companies and a raft of other VERY large companies. If they start impacting those revenue numbers and marketshare, then we will find out who is the 1000lb gorilla and who is just a very fat monkey.

“I have non-public information (or NDA) that proves the revenue for AWS”

OK. That’s great. It does not help me or thousands of enterprise customers make up their mind on the longevity of the business model. Until the cloud market gets some transparency in the revenue numbers, then the stability of the service is always under a cloud. There is a cost to business in adopting AWS and they cannot afford to not get that back. If you look at Rogers Five Factors for the diffusion of innovations, you will see observability. I believe that references customers and revenue numbers are the biggest factors impacting this part of the innovation curve.

I will add one comment about NDA’s… I worked in a vendor for 10 years and the use of the NDA ‘excuse’ in the sales process was a calculated tool. Some customers accepted the story and some didn’t. It goes like; “we have a customers with 000′s of XXX deployed, but because of an NDA we cannot share the companies name”/ If it’s not a blatant lie, then it is a big indication that the product has such limited adoption that they have not come across a company willing to communicate the information. Not that many companies have problems with being reference customer in my experience. The biggest barrier to this is usually the number of legal people etc.. to sign it off make it too hard for the average sales guy to get it done. Most sales guys are not comp’d on reference customers. Reference customers are very important and if a vendor could get them, they would.

Industry analysts and media (including professional bloggers) should spend more time trying to uncover real data and help create more accountability and transparency. This will be one of the greatest value creating strategies for the cloud market as a whole. Sound bites (like this blog), un-substantiated rhetoric and marketing hype just contribute to the belief that cloud is all a bunch of fluff.

Now don’t get me wrong, I believe in Amazon’s success. I put my personal money where my mouth is and have a meager investment in $AMZN shares. I am in awe of Bezos and the whole Amazon management team. Gizmodo might be right in anointing him the new Steve Jobs. Time again they have proven they can compete. I love the online stores. I think the tablet/e-reader strategy they have is spot on in the face of a seemingly unbeatable phenomena.  There is no intent of negativity in any of my comments, no right or wrong. All points of few on this topic have some validity. To me it is equally true to say that;

‘AWS is following a loss leader strategy initially to gain market leadership before monetizing a high margin enterprise business’

as it is to say

‘The current business model is a “runaway train” they don’t need to worry about gaining marketshare’

I just want the alternative point of view to get some air time.

Contributed by: Brad Vaughan