Hacker News Reader

Comment by scarmig | Hacker News Reader

It's a bit of a conundrum: every company is going to shut down products. What's the best way to launch new, risky things? Startups can simply go all in on the product, and if it fails, the company dissolves, though people rightly are hesitant to go with a fly by night startup for exactly that reason.

I'm not even sure that Google has shut down more products than an equivalently sized company. But it's certainly shut those products down in such a way that it's generated far more backlash and ill will than anyone else.

Just brainstorming, but perhaps a large company, when launching a new product, could establish some kind of dedicated trust to provide credible assurances that e.g. the product would be supported for at least 10 years.

View on HN

> I'm not even sure that Google has shut down more products than an equivalently sized company.

AWS has shut down I believe only one service in its entire existence (SDB). And only when they had a viable alternative (DynamoDB) and helped their biggest users make the move.

I can't recall any Apple service that has been shut down without an alternative. They've certainly cancelled hardware programs, but that doesn't break your existing hardware. And they give plenty of warning for iOs phase outs. They pissed some people off by dropping support for old apps, but only after most of the big ones had been converted.

And Microsoft is the king of long term support. How long did they keep supporting DOS in Windows? Or 32 bit programs. Or Windows 2000!

> I can't recall any Apple service that has been shut down without an alternative.

I can think of an esoteric one that no one misses: Ping (their music-based social media network)

There were a decent number of features in Apple's iTools / MobileMe package which were discontinued with no replacement. One big one that comes to mind was iWeb (static web hosting).

Also their "Print a photo book" from Photos (at the time it was call iPhoto I think?). You could design a photo album and they would send you a really nice bound copy. They discontinued it but had 3rd-party plugins that filled the gap for a while before even those died (I think).

Ironically enough, Google Photos still has this capability.

Amazon shutdown Drive (their dropbox competitor), there are plenty of consumer services Amazon has shutdown. Google's problem has been their tendency to shutdown perfectly good services because they aren't successful as ads. Reader is the posterboy for this; a service probably used happily by millions (or tens of millions) that had the plug pulled.

The fear with Stadia wasn't that Google may just shut it down, it's that it could have been very successful, with millions of happy users, and Google still would have shut it down. That's what separates Google from other companies.

jedberg said "AWS has shut down I believe only one service in its entire existence (SDB)" - emphasis on AWS, not Amazon as a whole. Big difference when you're talking about developer services, as opposed to end-user services.

Now, it's a perfectly valid point that Stadia is a consumer service, like Drive - but in a thread where there are discussions about developer services (GCP, for instance), making the distinction is important.

>(GCP, for instance), making the distinction is important.

The distinction is meaningless in that case because GCP hasn't sunset any services either.

GCP did shut down IoT Core. Only one I know of, though.

Google has shutdown many developer-facing services, no?

> it could have been very successful, with millions of happy users, and Google still would have shut it down

Weeps in Google Play Music and Hangouts.

> Reader is the posterboy for this; a service probably used happily by millions (or tens of millions) that had the plug pulled.

I feel like there's a whole generation of people who are now in senior decision-making roles who are still bitter about Reader. Killing that product, specifically, is likely costing billions per year in GCP revenue.

Windows phone was the biggest shutdown of all time. Off the top of my head, there's also Edge and Silverlight.

Edge and Silverlight both had replacements. Windows phone was shut down because almost no one cared.

The difference is Google shuts down services people actually like and use simply because they aren't profitable enough

10% of Europeans cared, but it wasn't enough.

And windows phone still got updates for over 2 years even after the shutdown, despite the fact it had been semi dead for years before then.

In terms of market share, Windows phone was a few percent, but that's still millions of phones.

Silverlight may have a replacement, but that doesn't help the devs who sunk their time developing with it. Same for people writing extensions for and debugging compatibility with Edge.

So is your argument that people liked and cared about Stadia? Because right now it feels like a really odd double standard. Go read through that canonical "Killed by Google" list. How many of those products were more popular than Windows phone / had users be more invested in it / etc? Be honest, it's a handful. Most of them you hadn't even heard of.

It feels like you have a double standard at play here, and are giving both Microsoft and Amazon a free pass on their abattoirs of dead products by various excuses, and totally ignoring that those same excuses would apply to Google's.

What's the replacement for Silverlight? Or do you mean it had replacement in a sense that nothing stops devs from going back to JS?

Bias up front: I work at Google but this is just my personal opinion.

> AWS has shut down I believe only one service in its entire existence (SDB).

Sure, but that's comparing a different branch of the company. Stadia is a consumer product, not a paid developer product. On the consumer side, Amazon has discontinued plenty of things (as every large corporation has):

According to this article[1], Amazon has canceled Haven, Amazon Spark, Amazon Restaurants, Amazon Storywriter, Amazon popup stores, Dash buttons, Amazon Tap, Instant Pickup, Amazon Tickets, Whole Foods 365, Amazon Fresh's Local Market Seller, Quidsi, Endless.com, MyHabit.com, Amazon Webstore, Amazon Destinations, Amazon Local, Amazon Wallet, Amazon Local, Fire Phone, Amazon WebPay, Amazon Askville, Amazon PayPhrase, and Amazon Auction.

Relevant to this thread, Amazon Games is technically still around, but they canceled Nova, Intensity, Breakaway, Crucible, and the Lord of the Rings MMO. Many top executives have left.

[1]: https://www.businessinsider.com/amazon-products-services-fai...

Microsoft shutdown stores, resulting in loss of purchases for customers. I lost a few ebooks when they shut down their ebook store, their Zune Marketplace and Play For Sure was a confusing fiasco with the final result of authorizing purchases for one final resting place before the DRM servers were killed.

So they don't exactly have a spotless record.

Steve Jobs's Apple liked to deprecate the whole API underpinning their OS from version to version. That and "upgrade" products with increasingly inferior versions (Final Cut, iTunes, XServe).

And then there's any era of Apple and their habit of removing consumer choice and forcing customers more into their closed ecosystem.

> Steve Jobs's Apple liked to deprecate the whole API underpinning their OS from version to version.

Can you provide some examples? The years of supports for things like Carbon are hard to reconcile with that claim.

> I can't recall any Apple service that has been shut down without an alternative. They've certainly cancelled hardware programs, but that doesn't break your existing hardware. And they give plenty of warning for iOs phase outs. They pissed some people off by dropping support for old apps, but only after most of the big ones had been converted.

Not sure if it counts, but dropping 32-bit support on macOS may be counted.

Did they ever actually shut down SimpleDB? Last I checked, it was still online, although maybe they stop most accounts from creating new tables now.

>What's the best way to launch new, risky things?

Slow-roll invite-only launch to establish a core user base, work out the bugs, show staying power, and build from there. Exactly what Google did in 2004 with Gmail.

The same strategy that failed with Google+. It's not as foolproof as you're making it out to be.

That's not really apples to apples though. Gmail worked with external SMTP services to send and receive mail. Google+ required that anyone you wanted to connect to also have been offered and accepted an invite (or received one of your limited invites).

It seems plausible that social networks have more of a network effect than things like a game streaming platform. Obviously there's still some network effect and a certain marketing aspect to it, but I don't think they're directly comparable.

You have to convince game developers to port their games. If the reward is a few thousand users, it's going to be a hard sell. Also the R&D and infrastructure required to even start up the project is too high. Building a Gmail (especially back then) is small potatoes in comparison.

Well for a thing like Stadia it's not doable, you have very complex hardware and r&d to do first.

That works great for a service where costs scale with users so you can run lean in beta.

Unfortunately I don’t think Stadia fits that description well; you need to build hardware, network PoPs, license games, etc.

Maybe there is a private-beta approach that really iterates, initially uses off-the-shelf hardware, only launches in one state, has limited games, etc. but it’s hard to make a splash like that.

I think if Stadia had just been better (so everyone using it was raving about it) the Google reputation might not have mattered.

It just ended up not being a game-changer economically, people still want to buy consoles etc.

The model of thin-client gaming might win long-term but it’s just not a clear winner yet.

> Unfortunately I don’t think Stadia fits that description well; you need to build hardware, network PoPs, license games, etc.

Hardware is literally scaled by customer demand, and was one of the reasons gmail went with the invite only launch.

Network PoPs are less an issue when you are already piggybacking off of Goog / GCP infrastructure, and you can mitigate the remaining costs by per country launches. And they did exactly that. It's still not supported in Hawaii. https://support.google.com/stadia/answer/9338852?hl=en

> I think if Stadia had just been better (so everyone using it was raving about it) the Google reputation might not have mattered.

Stadia was too late to market, in a field where content is rare enough relative to the number of competitors to have bargaining power. And they have no vertical integration to lean on. Nvidia has GPUs in house, MS/Sony/Nintendo have game devs in house for exclusives. Amazon _might_ be able to parley Twitch into a profitable Luna, but its a long shot.

> Hardware is literally scaled by customer demand

Hardware manufacturing is, but design is most certainly not. It takes as much work to design a controller that you build one of as it does a controller you build a million of.

Okay, fair, design costs have to be amortized across the userbase, and probably the correct solution to this is to build hardware people want to use with or without Stadia, so your design isn't anchored down. I.e. I can (and do) pair PS5 and Switch Pro controllers with other OS's. Obviously those other platforms don't offer the force feedback or HD rumble, but at the very least it's a signal to the your designers and the market that you expect your designs to be competitive in the open PC market.

Obviously less viable for the set-top boxes, but still a valid strategy if you can do it.

Hardware is well understood to be OpEx intensive, and have high up-front costs. Look at the pricing for injection mold dies for example. The opposite of a good fit for an iterative approach.

In my defense I was thinking of the GPUs and servers, not the client devices (does stadia have clients?). In that scenario, might HW be capex -- a durable good you buy and use for many years, and can sell if you no longer require it?

Ah I see where you were coming from. Yes on the server-side that applies. The compute costs scale elastically with user count.

They do have a client-side console/controller, and to be fair I had to look up why they did their own (https://en.wikipedia.org/wiki/Google_Stadia#Hardware).

> While Stadia can use any HID-class USB controller, Google developed its own controller which connects via Wi-Fi directly to the Google data center in which the game is running, to reduce input latency.[8] Google is also exploring further ways to reduce latency, using an idea called "negative latency" which involves prediction of user input through various means so that any apparent network lag between controller and game response is minimized.[12] During its GDC 2019 keynote reveal, Google confirmed that the controller would also feature Google Assistant, which will automatically search YouTube for relevant, helpful videos related to the game they are currently playing at the touch of a key.[13]

I don't think any of Google's manufacturing capex is recoupable here; it's specific to making that thing, and if you don't want to make any more of that thing it's trash. (I'm not a hardware guru though, so that's not a statement made with high certainty. There might be bits that are reusable.)

> Maybe there is a private-beta approach that really iterates, initially uses off-the-shelf hardware, only launches in one state, has limited games, etc. but it’s hard to make a splash like that.

That's… kind of the problem? They have a reputation for abandoning products after making a huge splash. The only way around it is to stop looking for big splashes and start building products slowly instead. Pixel phones were notoriously only available in selected countries. Google Fibre is even more limited.

I never had stadia so I don't know, but was it games only? Or was it a YouTube TV box/Android TV and a game box? You could build a compelling user base to develop the scale and then make it more about the games.

A coworker and I were chatting about it, what happens if/when Apple drops and Mx SoC in the AppleTV? There are the obvious apps, you create a camera add on and add FaceTime to the living room. Things like that but you also have a very serious machine that can go head to head with PS5 and Xbox, legitimately. I wouldn't be shocked if something like that were to happen.

Conversely, they tried the same approach with Google+ and the slow rollout absolutely killed the momentum of a desired service.

Microsoft, Nvidia, and Sony have all launched game streaming products without this kind of coverage. It's a Google problem.

Importantly all of their products are still alive, supported, and even if they're doing superbly do not appear to be on the verge of shutdown.

There are a few problems to Google's way of doing things, having witnessed it from the inside. In no particular order:

1) Google tends to be over-optimistic and under-skeptical when it comes to new products. This is largely driven by organizational dynamics: Google's corporate structure encourages fiefdoms that come up with the Next Big Thing(tm) - everyone involved is encouraged to be wildly over-optimistic about their products, and there is not a countering skepticism from upper management to impose the right amount of discipline re: these wild-eyed claims of TAM, growth, etc. The net effect is that Google launches products that aren't sufficiently baked, with vastly overestimated initial growth. This creates disappointment as the products bounce off the market and do not get anywhere near the (completely fictional) projections.

2) Google's go-to-market strategy tends to be under-baked as well. This is related to point #1 - heavily over-optimistic projections causes Google to accept woefully substandard GTM plans. Stadia launched with an incredibly poor lineup and burned a lot of the initial goodwill and press which stalled any kind of momentum they could've gotten.

3) Google organizationally isn't set up to reward individuals that turn around troubled products. Promotions heavily favor new product, not fixing existing broken product, especially once the product has lost executive favor. This causes team death spirals - failing products experience intense team attrition that further hampers any kind of turnaround plan.

4) Google has comparatively high executive turnover vs. similar companies. This results in rapidly shifting high-level strategy. Products and projects fall in/out of favor so quickly it causes whiplash. Other companies (see: Nvidia, Sony, MS, Apple) seem to be able to identify product areas of strategic importance to the company, executing against it, and having the executive support to continue resourcing these projects even if they initially fail/disappoint (see: Apple Maps, PSVR). Google constitutionally does not have this ability - they talk a lot about multi-year investments in strategic areas but in reality their commitments are fickle.

Though do any of those do the Stadia model of having you "buy" the game specific to them to be able to stream it?

Other systems I'm aware of mostly piggyback on some other platform so your "ownership" extends to local usage also (like how Nvidia's system works with your Steam library), or are just Netflix-esque subscriptions that give you access to the available library as long as you're subscribed (like PlayStation Now, well, whatever it's now called under Plus, and Game Pass streaming).

Neither of those models has the same type of concern over losing your purchases. Google's track record is obviously a factor too, but the business model is as well.

It does sound like their distribution power is a double-edged sword when it comes to launching new products.

On one hand, the distribution power makes it extremely easy for new products to get lots and lots of users really quickly. On the other hand, it can give a false sense of security when it comes to product-market fit.

The only real solution I can think of is deliberately launching new products without the Google branding and without relying on the built-in distribution channels, working towards product market fit the hard way, and only after that should they consider taking advantage of Google's distribution power to accelerate growth.

They should have split Stadia into two conceptual halves: one with the technology to split the processing from the input and output, and one to do the hosting of the processing. Then when they decided to kill Stadia, they could have let other people run the latter, so that the former would still be useful to people who already owned it.

> every company is going to shut down products

There's typically a fight, though.

Shutting down most product lines doesn’t mean said product is rendered unusable.

It does mean any further support is non-existent. With PaaS and SaaS, this implicit contract between buyer and seller no longer holds.

> What's the best way to launch new, risky things?

To show great confidence in it and address the elephants in the room as directly and clearly as possible.

I think the main issue here is that the perception of Google being fickle and uncommitted means it's harder for third parties to want to commit resources to. Strong signaling from Google on long term commitments has to be made, but I think that Stadia is in a bit of a pickle because of its nature.

With a console, I assume there are some general timelines developers get on how long the console is going to be around, so it's a lot easier to develop a strategy for working with it because you know off the bat you likely have at least N years, your projects will take Y years, thus you understand how many projects you can put onto it before the console obsoletes.

With Stadia though, since it was just PC games and Android games being streamed, there are two ways you can try to understand it:

- It never obsoletes as Google just upgrades the hardware and OS to keep new fresh games coming in

- It obsoletes as soon as it's too costly for Google to refresh the hardware and they decide to cut their losses

My guess is a lot of people thought it would be the latter and just didn't want to invest time into it. I'm not sure how the process for getting a game on Stadia was, but based on a quick look at some articles, seems that Google was struggling with this aspect even as late as 2022 [0] with trying to help make the process more convenient and faster. That's 3 years into the platform already and they were still teaching developers how to get their games onto Stadia efficiently, and I have to imagine Google was already looking at the numbers for the datacenter costs and going "welp".

So how could Google have really changed it? My take is have this convenience and strategy for the porting from Day 1. I did not use Stadia or really follow it (just not interested in Cloud gaming in general), but looking at this article and the history of articles on porting games to Stadia, seems that it wasn't an attractive process from the beginning, for an already iffy platform for developers, with the looming fear that Stadia would not make the numbers to keep Google's interest.

Combine that with Players already unhappy with not actually owning a lot of their games and distrusting Stadia, I guess it seems like Google just couldn't quite sweeten the pot enough to convince them to pay full physical game price for a game they didn't really own and ran the risk of being removed due to obsolescence (a perception on players part perhaps, but this is again a communication issue for Google)

[0] - https://www.forbes.com/sites/krisholt/2022/03/15/google-stad...

Reply on news.ycombinator.com