The article discusses "cloud revenue" of $26.7b but the more relevant Azure revenue is less than a third of that.

"Microsoft hasn’t disclosed how much revenue Azure brings in, but the Evercore analysts estimated that it generated more $7.74 billion in the 2018 fiscal year."

From: https://www.cnbc.com/2018/10/24/microsoft-earnings-q1-2019.h...

Even then, Azure services are consumed in the context of Office 365. (Azure AD)

On the flip side, I believe Amazon counts their Office and Drop/Box competitors under AWS as well so it may not be as apples to oranges as first perceived.

I suspect there are a lot fewer customers for Amazon WorkDocs / WorkMail / WorkSpaces than for Microsoft's equivalents, though. I've certainly never heard of anyone using them.

This is, in fact, the first time I've heard of them, and we use AWS to host our production services!

WorkSpaces are popular for big companies that use VDI to manage their workstations.

In that scenario, the Microsoft numbers are even more mysterious as OneDrive for Business just another non-billable component of SharePoint that's consuming that storage, which AFAIK isn't Azure storage, but in dedicated Office 365 nodes. The Amazon Drive service AFAIK is just a frontend to S3.

The interesting thing about O365 is that it combines a very valuable licensing cash stream with a variety of services.

The story that Microsoft SEs used to sort of tell a few years back was that Microsoft needed so many spinning disks to support Exchange workload IOPS that they had ridiculous amounts of "free" storage available for low-IO workloads like SharePoint/OneDrive. No idea if that was true years ago or is true today.

Don’t think Amazon is as much as a competitor here. However, Google Cloud does roll in all GApps revenue into their Cloud reporting

Office is one of Microsoft's most lucrative lines, whilst I doubt that most AWS users pay much attention to WorkMail/WorkDocs. Many big companies play games with reporting numbers, and this is just a rather transparent example. Unfortunately, the fact that it got the Forbes story rewards that this kind of trick.

I think Dropbox uses their own infrastructure.

He means Amazon lists Dropbox storage as a competitor to AWS storage services.

I think GP comment actually meant that AWS has their own products that compete with similar products like Drop/Box and Office and they include revenue from those products as part of their AWS revenue.

Yes, that's what I meant. Haven't figured out a clever way to say Dropbox + Box, Drop/Box seems OK. I'm also led to believe they compete with email as well.

This is nearly a "no true Scotsman" definition of cloud revenue, where if Amazon doesn't have a product in a market segment it shouldn't count. Google and Microsoft both have offerings in the cloud SaaS office suite market segment, and it's a real liability for Amazon that they do not.

Not really. When people buy Office 365, they're primarily purchasing a the MS Office product, and the cloud just happens to be the mechanism of delivery. That doesn't really make this cloud revenue.

Imagine if MS was forced to "unbundle" Office 365. Consumers could purchase a license to run MS Office on any platform. On of which might be hosted Azure. What percent of the Office 365 revenue would accrue to the products division compared to the cloud division. Sincere guess: 90%+.

Another hypothetical scenario. Imagine if Apple offered a deal on iPhones. Get an iPhone for free as long as you sign up for a grossly overpriced $1000 iCloud contract. All of a sudden Apple would have hundreds of billion in "cloud revenue". But no reasonable person would all of a sudden say that they're now the biggest cloud provider in the world.

I disagree, totally. Most companies who buy Office 365 want Active Directory just as much as the office products. Azure AD is an official Azure product; it's got Azure right in the name and lives on the Azure console. AD, hosted email inboxes, OneDrive storage; that's what companies are increasingly looking to O365 for. The Office suite is just a checkbox next to those things.

You can't punish Microsoft's categorization of these things just because Amazon doesn't have directly competing products. Well, actually they do; they have Workdocs and hosted AD. But no one uses them. Both of these are official AWS products; are you going to suggest we should start sharding out Amazon's product suite by revenue per product just to get a perfectly fair comparison of revenue?

G-Suite is categorized as a Google Cloud product. I guess we should just completely cut G-Suite out of GCloud's revenue total? Even though it does directory-like functionality and competes directly with Office, Outlook, etc?

Where do you draw the line? Ultimately, you draw the line at mechanism of delivery and payment, which is what Cloud has always been about. Office 365 and AzureAD are pay-per-seat, and thus by extension correlates with pay-as-you-go, just like any other cloud product.

>Not really. When people buy Office 365, they're primarily purchasing a the MS Office product, and the cloud just happens to be the mechanism of delivery.

And when people buy Amazon RDS, they're primarily purchasing a database product, and the cloud just happens to be the mechanism of delivery.

Imagine if Amazon was forced to "unbundle" Amazon RDS. Consumers could purchase a license to run the database on any platform. On of which might be hosted AWS. What percent of the Amazon RDS revenue would accrue to the products division compared to the cloud division...

--

Yes, this is the literal definition of a "no true scotsman" fallacy.

Amazon RDS is literally just a hosted offering of an off-the-shelf database server (MySQL, PostgreSQL, SQL Server). There isn't really anything there to "unbundle".

Amazon Aurora does appear to be a somewhat unique offering, but it appears to be tightly coupled to their infrastructure; I doubt that it could be unbundled without major rearchitecture.

I don't think this is that fallacy. It's also fallacious to dismiss something as wrong if an argument in its favour uses a fallacy.

If something is described as 'cloud' but uses a definition most people don't recognise or describe as 'cloud', that's at least worth noting. It's arguably prescriptive linguistics.

But there are unbundled DBs that people run. PostgreSQL, mongo, MySQL and the majority of those are run on the AWS cloud.

There is clearly a distinction here. For example, I don’t know if the word Azure is even mentioned when I go to buy an O365 license.

> But there are unbundled DBs that people run. PostgreSQL, mongo, MySQL

And people can still buy standalone Office installations.

> For example, I don’t know if the word Azure is even mentioned when I go to buy an O365 license.

How does this matter? Microsoft includes this as Cloud revenue and not Azure revenue.

>Imagine if MS was forced to "unbundle" Office 365. Consumers could purchase a license to run MS Office on any platform. On of which might be hosted Azure.

You can literally buy office as a standalone product. That's the ONLY way you could buy it a decade ago. So the current situation is exactly what you're describing.

As a counterpoint: I purchased office 365 purely for Onedrive. I was totally happy with my ancient office products (2013?) before, but I wanted to get some cloud storage with good synchronization and sharing tools as well as a good web interface. The Onedrive/Office365 offering provided that.

However it might be true that I'm the exception on this, and that most of the Office365 licenses are sold to corporations that are more interested in the basic office tools.

I also subscribed only for OneDrive. I was happily using LibreOffice.

With Dropbox dropping support for ZFS on Linux, I had no reason not to switch. It's cheaper than Dropbox and you get Office.

Office 365 also has the cloud sync and collaboration features that are a pre-req these days for a cloud service. Whether you use them or not is another story, but they didn't just change how they price it.

I'm not really sure how my copy of Excel and Word running on my work laptop (and subscribed to monthly) should qualify as "cloud revenue".. We do host our email on their cloud servers, but according to their own sales team, that is worth $5/month per subscriber for email only. Yet we pay $20/month per user for the full office suite installed on their systems. (and more for people with other tools like Microsoft Visio and Project) Its not a "no true scotsman" its that many of their products have no cloud component at all.

Microsoft gloms together the pay-per-use/subscription part that totally doesn't need "the cloud" (Office apps) along with the other bits that do (email, Skype/Teams, OneDrive, etc.). It makes sense that they do from a financial reporting perspective, because they sell them and you buy them as one thing.

How exactly should they split out the email or OneDrive component from Word or Excel of your $10-20/month? I don't think there's really a sensible way to do it without overcomplicating things.

The copy of Word and Excel on your laptop have cloud integration. You can do collaborative document editing a la Google Docs through the desktop copies or the web app versions.

Amazon does have a cloud SaaS office suite, actually -- Amazon WorkDocs / WorkMail / WorkSpaces. I've hardly ever heard of businesses using them, though; Office 365 and Google Apps have the market pretty well locked down.

I feel like there is a fundamental distinction to be made between infrastructure services (like AWS S3, EC2, RDB, etc) and end-user services (like Office 365). Combining the two doesn't make a lot of sense; procurement for the two typically happens separately, and businesses frequently mix-and-match vendors for infrastructure and end-user services.

Do Amazon employees use these tools themselves?

I was an engineer at Amazon for a couple years (left about a year ago) and I've never heard of anybody using these services. It's possible that some of the services that we used internally were simply frontends for those AWS services. Having an internal frontend for an AWS service was common.

Yes, WorkDocs is used extensively for document sharing. Chime is the go-to chat/meeting/VC app. People use WorkSpaces when they need a Windows or Linux test machine.

I would disagree, it's not a "no true Scotsman" argument.

Microsoft is actively competing against AWS IaaS and PaaS services, and leveraging embedded product features and license terms to force the issue with enterprise customers. Microsoft's spin is that those embedded relationships will compel customers to just join Azure. (Example: Windows 10 and Azure AD dramatically simplify PC management. How receptive is the market to that proposition? (I don't know, btw))

If they aren't getting heavy traction with their millions of existing enterprise customers adopting cloud services, they are doing something wrong that Amazon is doing right.

Additionally, the fact the Microsoft can sell Office and Exchange is a known, irrelevant fact -- they've been extracting a toll from every business since like 1995. What isn't clear is if they have the ability to compete toe to toe with AWS (the market leader in IaaS/PaaS).

I think there is another perspective here. Consumer vs Developers. When we speak of the ongoing cloud vendor wars between AWS, Google, Azure and the likes of Digital Ocean, we mean cloud platforms for others to build and host things on, not consumer apps delivered through the cloud. If that was the case, we would have to count something like Dropbox or Google Photos as well, right?

To me cloud provider in this context means a vendor that provides cloud platform for others to develop on top of. RDS counts, Office doesn't.

That's how I look at it and I'm sure that's how MS looks at it. Certainly they wouldn't be blinded by Office revenue and think Azure is close to beating AWS. Office competes with Google Work Suite which isn't even part of GCP (even though it's linked).

> To me cloud provider in this context means a vendor that provides cloud platform for others to develop on top of. RDS counts, Office doesn't.

I completely agree with you on this one.

However all the cloud providers include their SaaS subscription services revenue in cloud revenue. So when comparing market, revenue and stocks, that's what we should compare them on. But here we are in this thread looking at their "cloud revenue" to figure out which is a bigger "cloud vendor". That's like comparing Apples and Oranges.

But if that's the case why not consider Google search under cloud reveneue and call it a day?

Because you don't pay a fee to use Google search.

Advertisers definitely pay.

exactamundo

When someone is directly comparing Amazon and Microsoft "clouds", I absolutely assumed they meant Azure vs. AWS. To the exact generic definition I shouldn't have done that, but when it's specifically up against Amazon, well, things can be true but also misleading.

Its not so much "no true Scotsman" as, "what is a cloud and what is useful to track?".

For comparing infrastructure offerings, it would be really misleading to include things like Amazon Music or MS Office (both of which are cloud offerings).

they do.

https://aws.amazon.com/workmail/

That's a slice of the stack. It's not a replacement for Word/Excel -- Google has Docs for that.

That said, I think that's evidence that this fits in the cloud revenue stream.

A more relevant AWS service is WorkDocs: https://aws.amazon.com/workdocs/

On premis SQL server is clearly not cloud.

Frankly, Microsoft are coming off desperate.

But they have transitioned office to a Saas model to do this. I used to buy office every 5 years, now I pay per month. It is also so much more than just the office apps, there is Onedrive and Sharepoint and email too. I would say that Microsoft have done a pretty good job of making it's products cloud based.

Or you could look at it like Amazon is way behind Microsoft in Saas.

Well parts of AWS are cloud desktops, workmail etc... Those are counted on the AWS side. Why shouldn't their Microsoft equivalent be counted?

Does Google include G Suite in their cloud umbrella? Microsoft does a really good job of using products like O365 or other IT kinda services as a lead-in to Azure infrastructure. Which is ironic because O365 kinda sucks.

They do. On O365 I’ve said the same thing for a while, but the front end rewrites that are rolling out are really quite good. (Admin is still a tire fire)

I'm not so sure.

In the world of large enterprise which is just now waking up and jumping all-in with "digital transformation" projects and "cloud strategies" Microsoft appears to be winning the battle. I work for a company that services large enterprises and anecdotally I'm seeing the majority of them choosing Azure over AWS. I believe this is in large part because Microsoft knows how to sell to enterprises, they have great tie-ins with existing on-prem core infrastructure like Active Directory which is a big deal to large companies. When large enterprises start jumping on board they will move the needle pretty significantly as they ramp-up. I anticipate Azure closing the gap on AWS pretty significantly over the next 2 years.

The numbers are NOT misleading precisely because they compare the cloud revenue. Google includes GSuite revenue in their cloud revenue. So does Amazon which includes Work Docs.

Using their "cloud" revenue to figure out which is a bigger "cloud vendor" is a fallacy.

But if we're talking about cloud computing and microsoft offers these these services in the cloud, they should not be penalized because their customers once started on the desktop. They're demonstrating that they are making the transition.

Yeah cloud and ML numbers are becoming a running joke in financial circles, at this point I'm pretty sure Oracle counts JRE downloads as cloud revenue somehow.

I wish these large corporations didn't demand so much creativity from their accountants to bluster an image.

Not to mention that Microsoft VARs have been doing things like offering discounts on Enterprise Agreement licensing that correspond 1:1 against pre-committed spend on Azure.

Wonder if AWS reports the revenue from running retail on top of AWS? Retail does need to pay for AWS services..

They might but I don't think the retail infrastructure is big enough to compare to say the office 365 use.

I may be wrong but I just don't think it's costing them billions in AWS infrastructure to run.

Given that Azure's market share is so much smaller than AWS wouldn't this just be "Azure has higher markups that AWS"

In many instances Azure is actually cheaper than AWS. Their egress is a lot cheaper than AWS for instance.