Thread by @QuinnyPig

I've been critical of the @FinOpsFdn for a while but apparently not so critical that they were too ashamed to release "The 2021 State of FinOps Report."

You may follow along with this thread via data.finops.org if it's still online by the time I get done.

First we--wait. *math math math* Did... did they just add together all of their respondents' numbers and not de-duplicate "folks who work at the same company?"

Aren't these people supposed to be good at a thing that requires exactly this?

I mean... get six respondents from a giant AWS customer like Netflix, CapitalOne, etc. and suddenly you can change the entire shape of the survey!

I'm always clear to articulate that the Duckbill Group's work (duckbillgroup.com) trends towards born-in-the-cloud workloads and that changes our perspective. That's only ~15% of the respondents to this thing.

It turns out that 90% of the respondents to a survey labeled "FinOps" and run by the "FinOps Foundation" consider "FinOps" to be a likely career path.

Tell me more, Professor Obvious!

"Well okay, Corey! Did you know that this problem becomes more urgent the more your cloud spend resembles a telephone number?"

Cloud spend proportion brings us some numbers. Their "crawl / walk / run" model leaves out "stagger," "flee," "curl up in the fetal position," and "nope the hell out of this mess" maturity stages.

It's worth pointing out that companies above 500 employees represent 1% of the businesses in the United States, but 70% of the respondents to this survey. If you're an early stage startup or know all of your coworkers it is very clearly Not Speaking To You.

From this we learn two things:

1. The graphic designer is conversant with many shades of "urine yellow."
2. The Finops Report from the FinOps Foundation polling FinOps practitioners isn't fully bought in on the nonsense term "FinOps."

"The higher the spend, the greater the need" (accurate) and "$5 million spenders look a lot like $500 million spenders" (laughable) are contradictory points that sound good until you put them next to one another.

"By 2040, my math shows that 80 billion people will be employed in FinOps roles."

In most larger organizations we speak to (more than a few) both 4 and 7 are wildly inaccurate numbers.

Gartner tells me over half of cloud spend globally is in the US, so I imagine they'll be calling the FinOps Foundation shortly to apologize for their mistake.

"Disagree and commit to calling it FinOps" is certainly a choice they have made.

They launched the foundation four years ago. If respondents to YOUR OWN REPORT “say “it doesn’t have a name yet” how confident are you in a report titled "State of FinOps?"

"8% vendors" and "we don't distinguish between IT and Engineering personas" are the only two bits of actually relevant data here.

A good question to ask would have focused less around job titles and more around "what's your skillset?" This could more entertainingly be framed as "choose between working with an API, Excel, and PowerPoint."

They FinOps'd the page design so you can only see one answer at a time. Look at these glorious incentives to get folks to participate, such as dashboards, feedback, weekly meetings, publishing policy documents, and having no earthly idea what the word "incentive" means.

These are mostly accurate. You too can master them all if you give the effort a full 192%.

Their graphics person would encourage you to Apple Watch your spend.

Yes, your company is spending basically the GDP of a mid-sized country every month on cloud, so it's very important to get a discount on training opportunities for your staff.

NO, you recognize that money is fungible, ask for $100K in service credits, and then pay for training!

I've been dunking on the @FinOpsFdn basically since they started, so calling out @LastWeekinAWS by name was almost certainly physically painful for them.

Additional notes on that last image:

a) Last Week in AWS is (most weeks) focused on the @awscloud ecosystem, not cloud economics.
b) Over 20% of the FinOps Foundation's respondents don't find the FinOps Foundation to be useful.

Ouch.

Originally the FinOps Foundation was started by Cloudability, and seeded by their staff and customers (other vendors weren't allowed in). Showing them neck-and-neck with CloudHealth has gotta be painful.

Also: well done to the @awscloud Cost Explorer team (it's really improved a LOT over the past few years!) and to the folks using the excellent-but-not-a-cost-management-tool @datadoghq...

The ordering of the tools is fascinating to me. Cost Explorer dominates, THEN two pay-for vendor tools, then the much-better-adopted Azure Cost Management. Hmm. That smells funny.

You can learn a lot from this report. For example, this image teaches us that for a group focused on improving visibility of cloud spend, they're profoundly awful at visualizing data.

Here they admit that they're getting their collective asses kicked by that titan of Cloud Economics, @msexcel.

This image is to reassure anyone who sees "2021" in the title and thinks we're living in the future, to a throwback era where "number of servers" was a meaningful metric.

It ignores autoscaling, Lambda, container stories, big vs. small servers, etc.

Of course they throw in "multi-cloud" to taunt me, but don't mention what the actual spread is. Are some big shops all on one provider? Does Oracle correlate to also being an IBM customer?

"It's a massively multi-cloud world, which is why we're leading with the terms native to a single cloud provider."

Completely wrong direction. In a purist-serverless world, the spend would be a function of user numbers / traffic levels; highly predictable via unit economics, a random crapshoot in terms of actual spend prediction. This requires... complex discussions.

Old and busted: sending teams email they won't read.
New hotness: sending teams Slack messages they won't read.

So now what? Step 1, share the report on Twitter. CHECK!

Step 3 is interesting, because I applied to join the @FinOpsFdn and was rejected.

At launch I was considered "a vendor" and thus ineligible for membership. I know because I was told this by the president of the FinOps Foundation who was also the co-founder of Cloudability.

So initially I wrote them off as a hamfisted marketing play. Then they became part of the Linux Foundation / CNCF / whatever, and ahh! It's a brand new dawn. NOPE. My colleague @jesse_derose cares about this stuff, but doesn't sell anything.

https://twitter.com/jesse_derose/status/1349892742411821058

So apparently if you're into cloud economics you should join--but only if you're not quite into it enough to start consulting in the space. Basically they'd like a roster full of prospects that they can market to.

I will now take questions!

A clear across-the-board winning recommendation in the Cloud Economics world is to drink more water.

https://twitter.com/justintstanley/status/1357132657641943041?s=20

Honestly I have a hard time deciding. It's free for practitioners, sliding scale with broken CSS for vendors:

https://twitter.com/emilyst/status/1357132916317294593?s=20

A team that does this effectively isn't just engineers; it's cross functional.

I'd also like to point out that Cloudability and the rest charge a percentage of your cloud spend, so the tool might cost more than the team.

https://twitter.com/poiThePoi/status/1357133158513270790?s=20

A good question. I think that post-Apptio acquisition Cloudability didn't care as much, and Foundations multiply like rabbits?

https://twitter.com/darryl_ramm/status/1357133446837981185?s=20

Ever create the headline, then write the article, find it doesn't fit the headline, but refuse to go back and fix the headline because now you're committed?

It's that.

https://twitter.com/JDonavan/status/1357134573256212481?s=20

I'm increasingly of the opinion that it *is not possible* to create a cloud cost tool that doesn't suck / solves the problem just due to the incentives required.

https://twitter.com/colinbjohnson/status/1357138419734904834?s=20

Building your own is an inevitability past a certain point of scale.

For the rest of the world Cost Explorer continues to iterate.

https://twitter.com/levi_mccormick/status/1357141240731623425

Target rate for most of the third party cost services is closer to 3% of your cloud spend.

Suddenly my offering @awscloud bill consulting for a fixed fee makes a lot of sense, doesn't it...

https://twitter.com/changeinside/status/1357144442898178053

There are one or two things they all do that customers appreciate (plus a boatload of things they don't) but the ability to run one or two specific queries generally isn't worth "a few million bucks" to most companies in this context.

https://twitter.com/changeinside/status/1357152685942845441

Agreed. Perspectives are valuable, but only to a point. I hate tools you outgrow due to budget.

"You get that's a $20K contracting project to build, right?"
"Seems expensive."
"You're paying $4 million a year for a vendor to build the shinier version."

https://twitter.com/CajunItalian/status/1357210363033317378?s=20