Why I Just Bought Amazon (AMZN)
Back when I set up my Covestor portfolio, I thought long and hard about tech/web stocks and couldn’t come up with any that I wanted to own. The two large cap Internet companies that I really like are Amazon and Google. I subsequently bought Google (GOOG) in late October at $672 and it’s up a bit since then.
Google trades at 33x next year’s earnings and I suspect, as usual, that the street is underestimating Google’s earnings power. So in my mind, Google is not that expensive as many other Internet stocks are.
Amazon (AMZN) is harder for me to get comfortable with. It’s trading at 100x this year’s earnings and 55x next year’s estimated earnings. And with so much of its earning power tied to its core retailing business, that seems expensive to me.
But there are two reasons I love Amazon. The first is that our family buys everything that we can on Amazon. If we want to purchase something, we first visit Amazon. If they have it, we buy it there. It’s because Amazon already has all of our payment and shipping information (all the various incarnations) and because they have never let us down. Amazon has nailed the online shopping experience. It’s the Wal-Mart of the web (without all the negativity that surrounds Wal-Mart).
But being the Wal-Mart of the web has never gotten me that excited about Amazon the stock. There is something else going on that does get me excited about Amazon the stock. Amazon is slowly but surely building a platform for web applications and services that is superior to anything out there. Almost every new company we see these days is using Amazon’s platform (S3 and EC2) for some part of their service.
We had a conversation at our portfolio CEO summit last month about Amazon. There were some questions about how reliable Amazon would be. One of our CEOs, a very capable software engineer in his own right, pointed out that if Amazon runs on it, why isn’t it good enough for a startup? That was a good point that got more than a few nodding heads.
For those who don’t know, S3 is a storage system for the web. If you have images, audio, video, or just files you need to store, you can open an S3 account and store them there instead of on your own servers. It’s highly scalable and very cost efficient. EC2 is a virtual server. If you’d rather provision a server in Amazon’s data center than get your own server and put it into a colocation facility, it’s a relatively simple thing to do with EC2.
This week, we got to see the next card in Amazon’s hand. They introduced a web database service that works with the rest of their platform. Just because you can store files in S3, doesn’t mean you can make easy queries to them. For that, you need a database layer. A database for the web. It’s pretty interesting that with all that the web has brought us over the years, we don’t really have a native database for the web.
My friend Dave Winer describes this new database that Amazon is delivering to the web in a post yesterday. Here’s what got my attention:
Today, when a company raises VC, it’s probably because their app has
achieved a certain amount of success and to get to the next level of
users they need to spend serious money on infrastruture. There’s a
serious economic and human wall here. You need to buy hardware and find
the people who know how to make a database scale. The latter is the
hard problem, the people are scarce and the big companies are bidding
up the price for their time. Now Amazon is willing to sell you that, to
turn this scarce thing into a commodity, at what likely is a very
reasonable price. (Haven’t had time to analyze this yet, but the other
services are.) Key point, the wall is gone, replaced with a ramp. If
you coded your database in Amazon to begin with you will never see the
wall. As you need more capacity you have to do nothing, other than pay your bill.
Further, the design of Amazon’s database is
remarkably like the internal data structures of modern programming
languages. Very much like a hash or a dictionary (what Perl and Python
call these structures) or Frontier’s tables, but unlike them, you can
have multiple values with the same name. In this way it’s like XML. I
imagine all languages have had to accomodate this feature of XML (we
did in Frontier), so they should all map pretty well on Amazon’s
structure. This was gutsy, and I think smart.
Now let’s think about what happens when thousands of web startups start using Amazon’s database system to support the services they build. Thousands of web applications using a common and shared web native database. That’s going to allow some very interesting things to develop. It may become necessary at some point to build on top of Amazon’s database if you want to get a level of interactivity with other similar services. Is Amazon building a data API? I don’t know because I am not technical enough to truly understand all of this. But it sure smells like a big deal to me.
The funny thing about Amazon’s platform play is that it feels like something Google would have done. But they didn’t (or at least they haven’t yet). Amazon is out googling Google.
That makes me want to own this stock. And so I bought some shares today. Amazon and Google now make up about half my Covestor portfolio. The rest is oil and commodity plays and Toyota which is a hybrid engine play.
I have always loved Amazon and now I have a good reason to own it. And maybe you do too.
My friend Vanessa had a great post about Amazon a while back….worth a readhttp://www.fridgebuzz.com/2…
Great link/postThanks for sending this
Hi Fred, when I read winer’s post yesterday, I immediately thought if the implications for VCs. If Amazon can deliver on this promise and companies can bootstrap longer, how will that affect VC investment, peformance, etc? Could this push back the need for VC funding further in the companies life cycles, possibly necessitating larger VC investment and larger fund size?
I think its one more step toward capital efficiency in web services startupsBut I still think this is good news for VCs and entrepreneurs alikehttp://avc.blogs.com/a_vc/2…
I completely agree that what Amazon is doing for entrepreneurs in the web space is of great value. I’ve taken advantage of EC2 and S3. However, I believe that they’re going to need to improve SimpleDB internal structure before it can become truly useful for actual work.The current implementation has some wonky requirements for working with integers, such as they all need to be the same length. For example, if you need to store numbers in the 1,000,000 range, you’ll need to zero-pad everything so that 1 winds up being 0000001. Additionally, there’s no way to store negative numbers, so you need to store an offset somewhere and do the subtraction (and then zero-pad that). Those are just two issues among others.Granted, technically these aren’t tremendous hurdles to overcome. However, it does add a hurdle to one’s agility. I’m sure someone will come up with an adaptor to handle the annoyances.Thanks for musing openly.
this is a common misreading the simpleDB docs: you only need to resort to zero-padding and offsetting negative numbers tricks if you intend on using the “<” or “>” operators on numeric attributes, and have that comparison be numeric. SimpleDB treats all of its values as strings, so < and > use lexicographic ordering– not numeric ordering. While I agree this is a little annoying, it’s not technically a requirement.
Not only it enables web startups, Fred, but also TV channels, rock bands and any content or application that needs to scale massively.Amazon web services can compete orthogonally with Yahoo and iTunes too, when every artist out there wants to manage their output themselves.It’s only a matter of time before the next Radioheads are able to rent a “rock-band-application”, upload some mp3 and sell their stuff massively.Not for the big-boys-only anymore.
Hi Fred,You mentioned near the end of your post that you had a position in Toyota as part of a hybrid engine play. I don’t recall any posts (after searching a bit) that discuss the reasons behind this and would be very interested in what the motivations are. I’m in the market for a small/mid-sized car here in the Bay Area but every cost model I’ve done for the hybrids ends up more expensive overall.
Honestly I haven’t researched the whole hybrid engine thing enough to be able to post intelligently about itToyota is one of my smaller holdings and it hasn’t done much since I bought itI need to read up on this whole topic a bit moreFred
Thanks for the response. I’ve been digging into it a bit and will be posting what I’ve found and what assumptions I’ve made. So far the best deal for overall cost (not just to me but including production and eventual disposal costs that may not be completely borne by me individually) looks to be any reasonably efficient used car from the ’90’s. It’s been very interesting looking into the various subsidies (both government and company internal), tax breaks, and other costs that go into the production of a vehicle.
The Amazon DB has numerous shortcomings, but may evolve to include all the functions of a fully functional database that most web services need.
That seems to be the prevailing wisdom. My bet is they do what it takes to get it rightFred
To quote the database super Server industry foremost expert, Kingsley Idehen of OpenLink Software:”For every database function of the old model, increased complexity, increased administration. Stored procedures, replication, unified data model, DB to DB connectivity, etc.” I might add, that the new languages used to spin out new web based applications are all trying to make everything a nail, because all of the databases that they have to work with (mySql, Postgress) are hammers. This has created a specialist industry where we have something that did not exist before: Programmers that have an intimate knowledge of the serving and request structure of the client’s remote calls, to a degree not known in previous era’s.In other words, you can be a Ruby or PHP programmer, but to be really in demand, you have to know A LOT about the back end. You can get started with basic knowledge, and get a small demo working, but once you get to even mid sized apps, you find that there are all these issues in the database closet that are very, very arcane.I am going through this right now, where my poor little bootstrap venture has such a starting mezzanine complexity requirement, that is a useless exercise to start with under provisioned complexity. If you are going to have thousands of small user objects that need to updated and tracked on real time, you have to solve an number of vexing problems that even heavy databases have problems with. Sure, an off the street Rails guru will get you so far, but that data back end……http://www.squidoo.com/Thru…
i agree 100% fred. not to sound like an egghead (and i probably don’t understand the book completely), but i think their moves around web services are a textbook way to respond and react to innovator’s dilemna. compare them versus ebay. amazon took a much more audacious path around an emerging trend while ebay focused on just squeezing every incremental penny off their auctions.great stuff, fred, although i feel like it’s still a little pricey 🙂
It is pricey but I like to be invested in innovators. If it goes down I can always buy moreFred
As a Web X.0 entrepreneur and someone who has been in the software industry for 20 years, I have kept a close eye on AWS for aforementioned reasons. Unfortunately I couldn’t use it for my current startup because it was still in private beta at the time. Recently at a AWS event co-sponsored by KPCB, I asked a KPCB VC about what AWS means for VCs. It has become so affordable (AWS, VPS, shared hosting, etc) to build a Web X.0 site that a lot of startups don’t need VC money, at least not for a good long time if ever. Companies like SmugMug, hotornot, plentoffish, etc will never need VC money because their burn rate is way below their revenue stream. Fred, I would love to hear your thoughts on this subject.
I left a comment on this thread with a link to a post I wrote almost a year ago called ‘web 2.0 is a gift, not a threat, to vcs’Go take a look and let me know what you thinkFred
Good post. I especially agree with this sentiment “All that area between the red and blue lines is risk that has been taken out of the equation for VCs and equity that should largely accrue to entrepreneurs.”My question is what about companies like smugmug, hotornot, plentyoffish? They never took VC money and are profitable and have no plans to take VC money. Are they the exception or soon to be the norm for web 2.0 startups? Even in a case like facebook, it seems like they are in command of the terms rather than the investors putting the investors at a disadvantage.
I am not sure that investors ever had the advantage when it comes to a highquality venture deal run by top notch entrepreneursfred
Until there is a fairly decent competitive offering, I have to think that it is too risky to put your whole company on AWS.
You are long Google yet you link to Yahoo Finance? Seems odd to me.
I wanted to generate smart links on the tickers and I know that a link toyahoo finance will do that. Not sure if a google link will do that.fred
Fred – Obviously Amazon’s service offerings are getting traction with customers. But what do you think of it as a business? For example Nelson Minar analyzed this when S3 and EC2 first came out and concluded: “[T]his best case scenario would be a 2% earnings improvement. Nice, but worth the distraction to their primary business?” Have you done your own analysis?
No I have not, but I suspect that the economics of this business willchange.If a significant percentage of web services adopt Amazon’s serviceofferings, I think they will be in a position to make money in more ways.fred
Fred, there’s a company that’s already added a DB layer to the web, so you can treat the entire web as a large data store. Josh Knauer is the founder, and its’ a mature business… here’s one of thei projects (out of CMU) http://www.maya.com/infocom…
Thanks CharlieI’ll check them outI love blog comments. I learn more from you all than anyone else!fred
i could disagree with the statement ” If you codedyour database in Amazon to begin with you will never see the wall. Asyou need more capacity you have to do nothing, other than pay yourbill”. The DB service _may_ be that simple but the EC2 and S3 arenot. Its not a true elastic service without a bunch of programming done on your part. There are even products out there that you might need to buy if you aren’t ableto do so. (weoceo comes to mind)…As far as their future aspirations to build a “web data API” thatmakes commingling services easier down the line…probably not sotrue. everybody has an API, i dont see why 2 startups having theirdata in the same database service would make it so much easier that itbecomes a no brainer. some people need secure data (wesabe, flickr), some people dont (weather, traffic, maps) and on and on.
It’s not all about what can be built.It’s also about what can be imaginedand what happens when the entire company is focused on something other thanan engineer in the middle of a huge company.And network effects are important tooGoogle might build a wikipedia clone or an etsy clone or a twitter clone butthat doesn’t mean anyone would use itfred
I think the VC business will be fine. I don’t think Google can monetize most web services better than the owner of the service can. Google ads are not the most native monetization system for most web services.Ultimately the owners of these services will need to build organizations to best capitalize on the opportunities in front of themAs I said in my ‘web 2.0 is a gift’ post last year, this is all good for entrepreneurs and VCs as it reduces risk for everyoneFred
i agree that cheap commodity like technology infrastructure is great for entrepreneurs. and maybe great for vc funds which are focused like USV or First Round (that is, First Round’s upcoming new fund which looks a lot like USV). But how does this help big or huge VC funds?And what do you make of the increasing size of VCs funds? the need for capital seems to be going down but the supply of capital seems to be going up?