Google Cloud offers a model for fixing Google’s product-killing reputation

jdale

Ars Legatus Legionis
18,401
Subscriptor
Enterprise APIs basically get a roadmap that promises stability for certain APIs.

Great.

Google Cloud Platform is actually facing down a 2023 deadline to beat AWS and Microsoft, or it will risk losing funding.

But if it doesn't work, we'll cancel the whole thing and the API problems will become moot since you have to migrate to a new platform anyway.

That's reassuring, right?????
 
Upvote
255 (258 / -3)

stormcrash

Ars Legatus Legionis
11,140
It's a good promise but the question is will they be able to keep it? This will require a big shift in the way the division plans and delivers projects and features that to put it bluntly doesn't fit with Google's engineering approach overall at the company.

And if G Cloud really is staring down a 2023 deadline to "beat AWS and Azure" (whatever that really means) the future could be looking grim as I don't think this pledge without any current action to give it teeth will be enough to change the momentum of the market or platform
 
Upvote
58 (61 / -3)

Scandinavian Film

Ars Tribunus Militum
1,535
Subscriptor++
Rumor has it (according to a report from The Information) Google Cloud Platform is actually facing down a 2023 deadline to beat AWS and Microsoft, or it will risk losing funding.
I know it’s normally best practice not to comment on rumors, but maybe just this once, Google should respond to this one.
 
Upvote
111 (113 / -2)

xoe

Ars Scholae Palatinae
7,496
I think GCP has other problems too. I recently had a project that required running up to 40 game servers for as little as 5 minutes at a time. With vanishingly little experience with paid compute services, I was up and running in under an hour or two, on AWS (EC2). After several weeks of experience with AWS someone requested help replicating my work on GCP, it took us a couple hours together to figure out how to create an image, apply it to a template and launch multiple servers.

So now, myself and 4 others in our group are familiar with AWS and avoid GCP. Among us, I would reasonably expect that at least one of us within a few years we will be called upon to offer advice on which service to use for a project that will be lucrative for a cloud provider. Guess who we will recommend?
 
Upvote
76 (84 / -8)

nancy-drew

Ars Centurion
357
Subscriptor++
The best part of this whole thing is that Google doesn't consider fixing its ridiculous pace of changes of API and product abandonment, because it's terrible for users.

NO -- instead, try to create a new, "stable" line of changes, as if customers will now trust THAT. lol.
When Chrome OS has three dev channels before its stable branch, and the stable branch bricks devices, I don't think anyone can trust Google. At a high level, across the organization, they do a terrible job of prioritizing user and customer experience.

That's double-stupid when you consider they need as much public goodwill as possible banked up to avoid being anti-trusted into smithereens.
 
Upvote
90 (93 / -3)

kylector

Ars Scholae Palatinae
1,302
Subscriptor
Who knew that when customers have a choice -- unlike with Google Search -- they will choose to not be stabbed in the back by changes and abandoned products and APIs.

I guess Google didn't?

I switched to Duckduckgo almost a year ago and don't miss Google search, to be honest.
Not to derail this thread, but a few months ago I switched from Chrome to Firefox, and Google to DuckDuckGo. I notice a difference on both fronts. FF simply does not render fonts as nicely or animate as smoothly as Chrome. And while DuckDuckGo does provide good results and sometimes gets me all I need, Google does in fact show me better top 10 results. I don't know if it's pure algorithm or because Google has additional contextual data about me to sort slightly better, but it does get me better results.

However, I think having one dominate search engine is bad, and far worse is one dominate browser. (As history has shown us!) I'm trying to get better FF support by skewing the usage metric up ever-so-slightly.
 
Upvote
64 (72 / -8)
Lets come back in five years and see if this is still around.

My previous dealings with Google have been terrible. Despite being part of a high profile government agency, Google were disorganised and every meeting had different people. Obviously nothing ever got worked out.

Compare that to Microsoft, who have been simply great to work with and got us out of a couple of self-made disasters. Once, even Steve Ballmer gave an approval for something we needed. I had no idea things could get escalated like that (it did take a year though).

I have learned as a Business Analyst that consistency and stability are the most important things.
 
Upvote
119 (120 / -1)

UserIDAlreadyInUse

Ars Tribunus Angusticlavius
7,912
Subscriptor
"No, no, the API you're using is still supported. It's just that it's running through a compatibility layer in order to be compatible with the new backend. Oh, and that compatibility layer is also running through a compatibility layer, As is that one. And that one....and that one...and also that one. Actually, as it stands, that API - and yes, I know it's only two years old - is running through so many layers that I'm pretty sure it's metallic hydrogen at the core. But it is still supported."
 
Upvote
63 (65 / -2)

trucmat

Ars Centurion
238
Subscriptor
A well-thought out and well-tested API would be written so that the front-end API calls can survive change in the back-end code by which they are implemented. A spaghetti code API on the other hand is fragile and tolerates change poorly because cases weren't anticipated, code wasn't properly tested, or needless calls are made that reduces each modules independence. Hearing that Google constantly deprecates and breaks prior code makes me strongly suspect Google is following the spaghetti model of software development.
 
Upvote
66 (68 / -2)
Post content hidden for low score. Show…
This is odd, it reads like a marketing article sponsored by Google because they know they have a brand issue, and are #3 in cloud adoption.
I ran, there are several statements that are highly critical of Google in this article. I highly doubt any company hired to promote GCP would pay for and approve the release of this article.

That said, this is Google we are talking about, and shooting themselves in the foot seems to be as high a priority for them as creating messaging services. So who knows.
 
Upvote
33 (33 / 0)

InIgnem

Wise, Aged Ars Veteran
142
Subscriptor++
This definitely falls into the category of when I see it I'll believe it. My company (which I own) uses Google Workspace - aka Calendars, Gmail and Docs, but there's no other functionality I'm willing to give Google as a core program for running my business, in part because those are about the only things at Google I deem safe. And even then - Docs are only used for things that are internal - spreadsheets or meeting notes, that could be easily migrated if needed as raw txt files.

We'll see if they can earn trust, but no way I'm going to be the guinea pig for this announcement.
 
Upvote
27 (29 / -2)
One year is too little time for large orgs.

For a successful example look at Java, it has remained backward compatible since 1996. You take a 1996 binary and just runs on the latest JVM. Or the posix API or Linux's userland.

You don't break your customer's investment. Period.

Once an API is deemed stable and supported, it should be supported essentially forever (or 5 years at least).
 
Upvote
101 (108 / -7)

entropy_wins

Ars Tribunus Militum
1,701
Subscriptor++
I think GCP has other problems too. I recently had a project that required running up to 40 game servers for as little as 5 minutes at a time. With vanishingly little experience with paid compute services, I was up and running in under an hour or two, on AWS (EC2). After several weeks of experience with AWS someone requested help replicating my work on GCP, it took us a couple hours together to figure out how to create an image, apply it to a template and launch multiple servers.

So now, myself and 4 others in our group are familiar with AWS and avoid GCP. Among us, I would reasonably expect that at least one of us within a few years we will be called upon to offer advice on which service to use for a project that will be lucrative for a cloud provider. Guess who we will recommend?

I got in GCP with this startup after meeting the manager at SC19 - and we talked "cluster". The GCP slurm product has recently lurched into quite useful as they have moved the configuration to terraform, so should you wish to jump to AWS or Azure or.... you'll have no problems.

They have also picked up some debian/rockylinux images...

That said, the GCP web interface is *dreadful* ,slow and clunky.

And that's using chrome....

S
 
Upvote
6 (10 / -4)

krimhorn

Ars Legatus Legionis
39,865
Many Google products suffer from the specter of unceremonious shutdowns, and that's enough to force customers to seek out alternatives. The primary fix to the problem is simply mitigation—i.e., stop shutting so many things down all the time
The primary fix is to stop creating so many redundant products and enhance and fix the ones you already have. Instead for every thing they halfassedly create they axe two other more fully featured products. If they were also doing this in the cloud services space it's no wonder they didn't manage to get much marketshare. Especially from the golden goose that is Enterprise.
 
Upvote
64 (65 / -1)

sarusa

Ars Praefectus
3,280
Subscriptor++
You know, there was another big company which had the same problem in the 2000's... That company killed compatibility right and left, up and down, and killed their Linux distribution each time they released a new one.
That company is called: Red Hat. Look at them now.
Red Hat also let you stay on whatever LTS release you want. Google does not give you that option (except on Android, where it's a misfeature due to mfrs not updating their devices, but you can't really choose which release you're stuck on). But when GCP breaks its API, you are broken.
 
Upvote
25 (27 / -2)

PETEM19

Wise, Aged Ars Veteran
135
MAN.. if this article isn't ON THE MONEY!! this is exactly why i stay away from any google "hardware" and think long and hard about any new services they might offer. yes.. i do use their email and voice service because i started to use them in their early days.. but thats it.. if there is an alternative to anything they offer it considered before their service.
 
Upvote
14 (15 / -1)

unconcerned

Ars Scholae Palatinae
1,058
Many Google products suffer from the specter of unceremonious shutdowns, and that's enough to force customers to seek out alternatives. The primary fix to the problem is simply mitigation—i.e., stop shutting so many things down all the time
The primary fix is to stop creating so many redundant products and enhance and fix the ones you already have. Instead for every thing they halfassedly create they axe two other more fully featured products. If they were also doing this in the cloud services space it's no wonder they didn't manage to get much marketshare. Especially from the golden goose that is Enterprise.

Enterprise hates changes in upstream APIs. Imagine if it takes 5 years to fully develop a product just to discover the host platform doesn't exist as it used to be..
 
Upvote
35 (36 / -1)

DAG42

Ars Scholae Palatinae
610
Who knew that when customers have a choice -- unlike with Google Search -- they will choose to not be stabbed in the back by changes and abandoned products and APIs.

I guess Google didn't?

I switched to Duckduckgo almost a year ago and don't miss Google search, to be honest.
I try DuckDuckGo every once in a while. I also wanted to like Bing since for a while they were offering serious rewards just for using search (points toward xbox game purchases if I recall).

... to me nobody can get close to Google's search performance.

Super ironic example: I was in the middle of giving up on DuckDuckGo again and was perusing their website, like their job listings (to work there). I was reading through a job description and came across an acronym I was unfamiliar with, so I DuckDuckGo'd the term... first whole page of results was useless. I Google'd it and had to click literally nothing for Google to find and isolate the definition, showing it without me even clicking further.

I wouldn't mind spending more time in non-google search engines if it meant the same outcome as googling, but others aren't even that close behind.
 
Upvote
23 (29 / -6)
This definitely falls into the category of when I see it I'll believe it. My company (which I own) uses Google Workspace - aka Calendars, Gmail and Docs, but there's no other functionality I'm willing to give Google as a core program for running my business, in part because those are about the only things at Google I deem safe. And even then - Docs are only used for things that are internal - spreadsheets or meeting notes, that could be easily migrated if needed as raw txt files.

We'll see if they can earn trust, but no way I'm going to be the guinea pig for this announcement.


Where I work just switched to Google Workspace and Jesus Christ is this shit awful. I have yet to open an Excel spreadsheet that works correctly in Sheets (Shits?).
 
Upvote
3 (9 / -6)

panton41

Ars Legatus Legionis
11,115
Subscriptor
aaaand-its-gone-iqqx8x.jpg
 
Upvote
29 (31 / -2)

Blind Badger

Ars Praetorian
596
Subscriptor
Despite being one of the world's largest Internet companies and basically defining what modern cloud infrastructure looks like, Google isn't doing very well in the cloud infrastructure market. Analyst firm Canalys puts Google in a distant third, with 7 percent market share, behind Microsoft Azure (19 percent) and market leader Amazon Web Services (32 percent). Rumor has it (according to a report from The Information) that Google Cloud Platform is facing a 2023 deadline to beat AWS and Microsoft, or it will risk losing funding.

I don't really get this. My impression is that cloud infrastructure is a profitable market where you don't need to dominate to make money. So while I understand that Google would like to improve its position, I don't get why they might want to give up if they can't make progress.
 
Upvote
14 (15 / -1)

effgee

Ars Praefectus
4,545
Subscriptor
”To make sure we follow these tenets, any change we introduce … is reviewed by a centralized board … and follows a rigorous product lifecycle evaluation.”
Translates to:

”Dear customers/hens – we are aware that you view us with a certain amount of trepidation, but we promise that this time will be different. We pinky-swear to safeguard the keys to your house and to protect you from harm. Any change to this policy will be rigorously reviewed by an independent group of members of our species.
Sincerely, Foxes”


Yeah, no. Definitely not falling for that one again. Not from Google, at any rate.
 
Upvote
23 (25 / -2)

DAG42

Ars Scholae Palatinae
610
The best part of this whole thing is that Google doesn't consider fixing its ridiculous pace of changes of API and product abandonment, because it's terrible for users.

NO -- instead, try to create a new, "stable" line of changes, as if customers will now trust THAT. lol.
When Chrome OS has three dev channels before its stable branch, and the stable branch bricks devices, I don't think anyone can trust Google. At a high level, across the organization, they do a terrible job of prioritizing user and customer experience.

That's double-stupid when you consider they need as much public goodwill as possible banked up to avoid being anti-trusted into smithereens.
My Pixel phone experience has been the same.

About 1 year into ownership they updated it such that my primary use case for the phone (a media player) was no longer viable. My mother in law had the same situation with the same model phone. All media was unable to stream consistently, no matter which app you were using (audio-only and video streams).

It was absurd - they've burnt the bridge for me. Never a google phone again.
 
Upvote
4 (8 / -4)

jamesb2147

Ars Tribunus Militum
1,640
Despite being one of the world's largest Internet companies and basically defining what modern cloud infrastructure looks like, Google isn't doing very well in the cloud infrastructure market. Analyst firm Canalys puts Google in a distant third, with 7 percent market share, behind Microsoft Azure (19 percent) and market leader Amazon Web Services (32 percent). Rumor has it (according to a report from The Information) that Google Cloud Platform is facing a 2023 deadline to beat AWS and Microsoft, or it will risk losing funding.

I don't really get this. My impression is that cloud infrastructure is a profitable market where you don't need to dominate to make money. So while I understand that Google would like to improve its position, I don't get why they might want to give up if they can't make progress.
Because they were originally VC funded and know the *real* profits are in monopoly power. I'm not joking.
 
Upvote
18 (20 / -2)