Hacker News Reader

Comment by tuna-piano | Hacker News Reader

Amazon's rationale for its lack of profits is well known... Why does this rationale not apply to its cloud business?

Is the cloud business so good that it generates profits regardless of the investment it makes... or is the retail business so bad that it can't generate profits regardless of the investment AMZN makes in it?

View on HN

The switching cost for retail is really low; one Google and you know where you can get the same product for a potential lower price. This means that to win in the retail business you have to accept incredible low margins or even periodically a loss. Online retail is almost a 'perfect' market.

While with hosting, it isn't as easy to switch. When you have chosen AWS, of course you can switch to another vendor, but that is not without an investment in time and money. The higher the switching cost, the more you can raise the price above cost without customers (immediately) running away.

I would argue the exact opposite, but it's because it's not only just the product, but also the customer service and experience.

Switching from Amazon to a similar service is a huge risk to me. Will I get the same customer service when I order something and it comes shattered at my door? Can I rely on two day shipping being really 2 days? (I know some people here are whining about it being 3, the majority of us get it within 2, sometimes 1) Can I trust the reviews on the other sites? Will they know my preferences and suggest me other items or books as well as Amazon or are they coercing me? Do I get the same value added from the "premium" subscription such as Prime? Who knows.

Switching cloud providers? A snap, provided you have architected your systems according and not treated AWS offerings as black boxes and understand "OK I am using RDS, I can just switch to another MySQL provider easy enough" and prepared accordingly. "Ok I am using Elasticache, time for Redis." I imagine it's even easier if you stuck to containers. And unlike Amazon retail where there really isn't an online store that can match it in terms of customer service and experience, many would argue that a DigitalOcean or similar have better customer service, so there is competition there.

The impression I've gotten from friends working there is the reverse: The cloud business is making so much money it can't spend it fast enough. The retail side, meanwhile, is making money but spending it faster enough to remain "not profitable"

That's not really how accounting works

It's how it works at Amazon.

"How much gross profit is the retail business going to generate that I can spend on drones and the Amazon Echo v2"

For many years Amazon was proud that they did not make a profit.

The retail business is competing with a very old, very established industry. The cloud business pretty much created a new industry.

I think its also true that their online retail store created a new industry. They might not be the first online retail store, but they definitely spearheaded the online retail industry. I think the some of the reason why AWS is more profitable or has been profitable longer because.

- Amazon reinvested its revenue on its growth.

- AWS was a by product of Amazon retail store infrastructure (at least initially).

- AWS customer base are mostly businesses with big pocket and less likely to worry about spending money.

- Amazon retail store is built on the business of selling products cheaper.

- AWS profit margin tends to be higher than retail store.

Just thinking out loud.

"spearheaded the online retail industry"

"Online retail" is not an interesting category even if they did spearhead it. Amazon is nothing more than a catalog retailer, something Sears Roebuck and Montgomery Ward pioneered eons ago. Shipping goods to consumers out of a warehouse by common carrier is not a new industry.

Neither is leasing out compute resources.

The devil's in the details. Scale, convenience, ease of use...

How long has leasing out compute resources been an industry? 1-2 decades? I'm sure there were leases on timesharing systems in the 50s and 60s, but that's still just a half century, and hardly the same industry as AWS.

It's not the same industry as AWS in the same way as Amazon isn't really the same industry as catalog/mail order businesses of yore (around 1850s I believe that began).

You could get limited amounts of compute time on a mainframe back in the day, just like you could order flannel and get it delivered in days/weeks.

Now you get massive scale compute resources on demand and fresh fruit, veg and almost anything else delivered within an hour.

Neither of them are dissimilar from present offerings, but neither are the same industry as their predecessors.

> Why does this rationale not apply to its cloud business

Radically different margins. AWS is running on a 23.5% operating income margin. Retail doesn't come close to that. Operating margins on their traditional North America business are 3.5%, and 1.3% internationally. Retail has to do almost seven times the sales of AWS to match the operating income.

The cash cow they have in AWS, is supporting the outsized valuation in their stock. Once investors realized what they had in AWS, by revealing how it was doing, the stock soared to a new level of high valuation. You can go back to December 2014 as see when this happened, the stock had been stagnant for a while, then doubled in one year following the reveal of how AWS was performing. AWS will pay for a lot of patience with shareholders.

I think you're right, that comparing operating margin is the right way to look at this. But do Amazon and its investors really believe that investing in video streaming, fire phones, etc will improve that margin?

Considering Amazon is able to constantly expand into new retail fields (like groceries), launch a cloud services business, launch own its internal logistics business (that is the 2nd largest provider in the UK now) says to me the Retail business is probably more profitable than the media would have us believe...

From speaking with internal Amazon employees, their total profit is negative because they invest more money into expansion than they are taking in. They make an incredible amount of money and likely would be turning a profit if not for the extensive investments. But as it stands, expenses (including expansion costs) outstrip revenue, thus profit is negative.

Which is fine, so long as growth exceeds your debt-load. As it stands, it's one of two ways corporations should work. Either reinvest, or give out dividends... holding on to more money than your operational costs and/or shifting them to overseas subsidiaries to avoid taxes doesn't help the economy, or the communities the corporation services.

I'm okay with either disbursement or reinvestment/expansion/research. Those both grow communities/economies.

The retail and video streaming are really just a massive distributed integration test for their cloud services. We're paying for the privilege of doing their QA for them.

It's the same exact thing they used Toys R Us and Target for, when launching their partner shops originally. With Amazon's extremely frugal nature and thin margins, they'll always look for ways they can get someone else to pay for their learning mistakes.

This is an interesting question.

The e-commerce business seems to provide a near limitless opportunity for fixed investment towards getting your items to you faster. At what point will Amazon no longer have opportunities to invest in better logistics? 1 hour delivery for everything? There are hundreds of billions of dollars worth of investable opportunities there.

On the AWS side, I struggle to see a comparison. Yes, they need to build more infrastructure to scale the service, but there seems to be a saturation point. Where is the equivalent "last mile"?

Current global IT market is trillions of dollars, AWS is ~10B... lots of room there.

Of course there's room, but there aren't obvious ways to deploy billions of dollars. Products need to be developed for AWS, whereas for the e-commerce business there is a much stronger link between investment and ROI it seems.

Amazon's cloud products apparently don't have to be at the low end of the price spectrum. And yet they continue to dominate. That's much harder to achieve in retail.

Another difference is that Amazon already has the dominant position in cloud computing. Amazon's share of all retail (not just e-commerce), is not dominant.

They're upwards of around 1/5 the size of walmart, which is pretty significant... larger than most retailers that aren't Walmart, I would guess.

http://www.latimes.com/business/la-fi-amazon-walmart-2015072...

Reply on news.ycombinator.com