Hacker News new | past | comments | ask | show | jobs | submit login
Google Wave’s failure is a lesson for modern real-time collaboration tools (taskade.com)
211 points by feross on April 8, 2020 | hide | past | favorite | 162 comments

Wave's problem was even more fundamental than I think the article is getting at. Wave was, as an end-user facing product, far too amorphous. It could be used for email, for documents, for chat, and for just about anything else via bots.

The UI and features allowed for all of these things, but guided you on none of them. There were no first-class workflows for any of the possible use cases.

No, I don't want a conversation to spring up anywhere in the middle of a document. I want the workflows and UI for edits, suggestions, and comments to be distinct and guide me and all of my collaborators towards clear and efficient ways of accomplishing my goals. With Wave we had to invent our own conventions for each task. That's why everyone shrugged and left.

Wave as an enabling technology and platform for apps would have been great though. Make it easy to repackage the basic features in a tailored UI for chat, email, docs, etc... Maybe even leave in the general interface to allow for hybrid apps, universal inbox, etc. At least the specific apps would have avoided confusion on what Wave is actually for.

I think I sort of agree with you.

But I think that imposition of workflow is not taken kindly by users, users only like things in line of the status quo.

I think the greatest sin by Google was shutting down Inbox. I swore by it (the little travel cards, the ability to see and search email quickly -- e.g. the image previews, every day Gmail angers me because I cannot do things I could once do with incredible ease).

They should have had Gmail as a conservative conventional email tool, and slowly morphed Inbox to have Wave-like features. Slowly, iteratively, so that users could discover and organically make their own workflows.

Absolutely, our problems are not technical they are social and you overwhelm them with a lack of direction and too much choice, most folks, myself included will get paralyzed with what they should do next.

Checklists and linear flows are extremely important. Just look at how different bug systems or wikis are used by groups, same tools, different groups and their usage patterns are totally different.

Wave should have been a platform, and the workflows should have been curated as an organization specific behavior.

Somehow that didn’t seem to be a problem in Minecraft or git.

I think that there is a value in making software that is powerful and does not immediately make apparent precisely how best to use it. Examples also include NLE video and audio suites, and much CAD/3D tools.

Your tools do a thing. The operator makes the workflow. Make good tools, not a yellow brick road sold as one-size-fits-all.

The first few times I opened Live or Illustrator or After Effects or Sketchup I had no idea how to use them; I had to watch tutorials on YouTube to learn the mental models.

What webapps are like that? I offer that they have a corresponding loss of power.

Git is a problem for a lot of people. And A/V tools only become useable if you follow tutorials or courses - there are entire companies making their business on selling pro tools or after effects tutorial videos.

Maybe we should allow a market for education rather than assuming everyone should be able to use a starship with only 5 seconds of looking at the user interface.

I don't think this works in a market economy - some people are able to "use a starship with only 5 seconds of looking at the user interface". So they will always be incredibly more advantaged that people who aren't able to.

> Somehow that didn’t seem to be a problem in Minecraft or git.

Quite a few people are paralyzed by exact that issue in Minecraft. And quite a few people overcome that paralysis only because their friends or parents play and show them what to do or because they run into entertainment channel that is primary fun, but also guides them later. There is significant network effect in Minecraft.

The lack of initial guide combined with worlds being set on survival and with third person being default means that someone has to explain the game to you, else it is boring crap.

Git is perceived as more complicated to use then it should be by many people. The thing with git is that you dont have choice and have to use it.

Yet somehow both Minecraft and git are massively popular, despite these facts.

Git got massively popular, largely because Linus Torwarlds made it and started to use it for Linux. The mercurial was competitor, but was commercial and did not had such popular person and popular project behind. That does not mean problem did not happened, that means that when enough people have to use it (due to Linux) then you can start network.

Minecrafts ability to overcome the paralysis due to there being sufficient social pressure to overcome it so that you join friends and there being basically universal adult approval of the game does not mean the paralysis does not happen for many people. It is also sufficiently youtuber friendly, meaning you can do things even non-players find fun to watch on youtube. These things help to overcome paralysis which absolutely happens, especially when they are attempting to play alone and did not learned from youtube videos wtf they should do with it.

The wave could not possibly have "there will be fun with friends" nor "you wanna be like the other kids" attraction Minecraft had to overcome the same issue.

Mercurial was not the commercial competitor to Git. The commercial predecessor to both Git and Mercurial was Bitkeeper. The Mercurial and Git projects were both started as a response to Bitmover (the company behind Bitkeeper) announcing changes in how it would license Bitkeeper for open-source projects.

I used inbox regularly as a task/todo scheduler -- something I've never been able to get in the habit of despite many attempts with various techniques and softwares.

Get an email that means I have to do a thing? Swipe and schedule the email to come up when I have to do it. It comes up and I want delay, tell it to come up the next day. And use the exact same interface to create reminders for things that didn't originate in email. It made me more reliable and less stressed.

But I think the killer feature is it changed the way I interact with 90% of my email: I didn't. It was very good at showing me exactly the emails I care to interact with and no more.

I try to stay unsubscribed from all sorts of mailings but back on gmail I interact with way more mail I don't care about than I did with Inbox. Which has got to be bad for google, so when I say killer feature, I guess I mean it probably killed Inbox.

I never really understood inbox, and definitely didn't know that it was gone. I had some vague awareness that GMail was being rebranded as "Inbox" or something, when android pushed me a notification once. I guess I never figured out what it was.

Inbox was an experimental alternate UI for Gmail. Over time many of the Inbox features were added to Gmail, and eventually Inbox was turned down.

(Disclosure: I work for Google, speaking only for myself)

By “many”, Google meant exactly two. The snooze function, and an Archive icon that appears on hover. Of all the Inbox features to migrate, those were all we got.

I was so disappointed when they dumped me back into what was essentially a legacy product. It would be like if Apple randomly decided to forcibly downgrade every device to iPhone OS 1.0. “You really didn’t need any of those non-phone functions anyway!”

In addition to Snooze and Hover Actions, there's Undo Send, Smart Reply, and Nudges which were all built and tested on Inbox before being added to Gmail.

Undo Send was definitely a Gmail "labs" feature long before Inbox existed [1].

You're right about the other two though. Nudges are nice. Ridiculous, though, that a useless gimmick feature like smart reply made it in but not actually useful features, like travel bundles or even the "finance" category.

[1]: https://mashable.com/2010/08/22/how-to-undo-send-in-gmail/

Wave didn't really fail. It was abandoned.

Microsoft Teams does this mishmash of documents, chat, file-sharing, notes, discussions, etc and is having good adoption. But Teams had several iterations and several years for adoption and wasn't abandoned.

Yep, as far too many people have discovered to their pain, that's the Google Product Way: throw something at the wall; if it doesn't stick and gain viral-like adoption in a short time, kill it and move on.

and product after product, it just seems that this approach isn't ideal at all, in any scenario.

I'm currently starting to use Microsoft Teams and I really dislike this mess, that said I use what other use so..

True, but at the same time this exact amorphousness and ability to step away from any specific workflow is what made Wave great - in my own subjective opinion.

I agree with the first two paragraph, but when it comes to the personal opinions - I disagree with the third. I did wanted a conversation to spring up in the middle of a document that was originally an email. That was novel, unique and empowering.

However, at the same time, I agree with the objective part that everyone had to invent their unique conventions, and many had struggled with that. I was just lucky that we had it "just for fun" and figuring out the ways we can use it was a rewarding experiment, not a struggle.

I feel similarly that there was a benefit and opportunity to the amorphousness of it.

At the time I was experimenting with a roleplaying game bot, and felt there was a lot of really interesting colloborative storytelling possibilities in the flows you could have. Start an action as an "email" to your GM, add a bot roll (or equivalent; the RPG I was using at the time was diceless) in the middle of it. Flow that into a threaded conversation about the results/consequences of the roll with your GM (and possibly other interested players), then use that conversation to clean everything up into an edited final "document" of that turn/action.

There's probably a lot of missed opportunity from Google having canceled Wave so soon after opening it to the public. There was a feeling, right up until Wave shutdown that a lot of people were in the middle of interesting explorations of collaboration forms, many of which we are unlikely to again see such a large scale experiment in.

Wave was brilliant, but in classic Google fashion it was developer-driven instead of UX-driven so it ended up as some platonic ideal product instead of something that actual lead people to do things they needed to do.

>it was developer-driven instead of UX-driven...

This right here speaks so much truth to so many side projects that into products that then fail.

I wish we could just make people smarter instead of having to dumb everything down with handholding UX. Too bad that's impossible.

In general, people aren't dumb. They simply don't care to learn how to use your software and just want it to "do the damn thing". Also, your software probably isn't as easy to use as you think.

Well, 50% of people are dumber than average.

50% are dumber than the median. Confusing mean and median might be a telltale for which half of the population an individual belongs to :-)

I thought we had all agreed to pretend that intelligence is normally distributed? ;-)

”Average” can mean either.

UX isn't, or shouldn't, be about dumbing it down. It's about really understanding people's problems and making sure you are actually addressing them.

I don't see a contradiction between offering a task oriented "dumb" UI and surfacing a consistent model that allows people to get more powers if and when they need them. Good tools do both.

People are getting smarter as time goes on, so maybe we'll have good UX in a 100 years or something.

In a way Wave had been that enabling technology by becoming a compelling demo of Operational Transformation that was later incorporated into many other products with much more focused workflows (e.g. Google Docs).

OT is always "just fixing this one corner case" away from working. I can't blame Google they dropped Wave, extending, maintaining and fixing OT for new kinds of data/operations is a nightmare I wouldn't wish to my greatest enemy.

this is absolutely the case. first product that really demonstrated what plausibly-real-time sync / collaborative editing on documents could feel like (and this is back when AJAX was still (kinda) fancy)

There's also just a lot of inertia embedded in existing conventions and paradigms. When the spreadsheet was still relatively young, there were a number of products that were spreadsheet-like but different and none succeeded.

Was the spreadsheet as we know it today the very best possible approach? Maybe. But not necessarily.

Of course, the original Visicalc model has been massively extended in certain ways. Take Pivot Tables as one example. But core spreadsheets as most of us use them are still basically Visicalc with a nicer interface.

Agreed. Many people say it but it seems lots of people who make new products forget — you have to meet the consumer where they are today and hold their hand to the bright future that you envision. Jumping directly to that future often doesn’t work because people aren’t quite ready for it or don’t get it. The iPhone started off ridiculously simple if you compare to what we have now with iOS 13 or whatever. If google just basically made google chat actually good and last forever then add features to it or something they could have probably brought people to Google wave eventually. But they launched it like you started watching season 2 of west world half way thru. It’s like... uh what is going on here, huh?

Yeah that hit the nail on the head for me.

You’re app needs work if people’s reaction to trying it out is, “ok... now what?”

>No, I don't want a conversation to spring up anywhere in the middle of a document. I actually loved Wave specifically because I could do this.

We'd start off an IM conversation, where we'd discuss basic points. Then we'd start editing prior bubbles and turn them into tables of content as well as chapters of our design document. It was a seamless flow from basic brainstorming into full-blown design documentation. So much mental overhead which would normally be involved - stuff like needing to separate in order to write the actual documents, then come back to combine it all - was removed entirely thanks to Wave.

> With Wave we had to invent our own conventions for each task

The same is the case with git. But git ended up a big success, while wave did not.

Ironically, we are most capable and productive when we're constrained and have limitations.

I think it's more apt to say that we are more productive when we don't have to spend time making decisions (or building entire workflows)

Nah, it's just that choice paralysis is bad

I had no such confusion, neither I required any guidance. I simply used it for my needs which it fit perfectly well (while it lasted). I know many other people who felt the same.

Wave is like Internet at the beginning of it’s rave. From one side of things it’s only sending packets to different addresses, on the other side you only needed a market to get a global economy depending on it.

Right ! Everything was do-able... nothing was easy...

Google quit. Google always quits. Google is terrible at products. It's simple.

I enjoy reading a detailed analysis of a product's failure, but at the same time, lets remember that Google's leadership tends to fail the same way, over and over again.

I assume that Google is trying to follow a "fail fast" philosophy, so if a product does not do well in a set time frame, it is shut down. That is a great strategy for conserving capital. But it does mean that Google won't get into some markets, because other players are simply more patient, and are willing to stay in longer and get through the rough phase.

I was thinking yesterday that, in this aspect, Google is very much like General Motors during the decline years of the '70s and '80s. General Motors also had a similar habit where they'd bring out a new model (which would of course have issues). Then, they'd devote an enormous amount of time and money iterating on the model. And finally, just as the model started to show signs of market success they'd cancel it, and go back step 1. Meanwhile, their competition from Japan was steadily iterating and perfecting a stable lineup of vehicles

Sometimes I think that the only reason Google has maintained its level of success is that there is no Toyota or Honda equivalent for GM-Google.

From the outside, around 2000, I used to marvel at Nokia's ability to deliver. Then after starting work there and seeing "how the sausage is made" I switched to marvelling at the implication that all of their competitors must somehow do it even worse...

There are multiple. Fb is one, slack and zoom is eating into enterprise communications etc... Plenty of competition where google couldn't take off.

Disclaimer: former google employee

But Google still dominates Ad Space and Search (which is connected). And they still dominate search, growth rates for competitors don't really seem meaningful compared to Google's market volume. And unless governments step in, search market share will secure their dominance on the ads market.

So unless someone manages to offer a superior (not just comparable) search experience, Google will continue to thrive despite all other products.

True, search is google's moat.

However, most people now go to amazon.com to find out what they want to buy. I still use G for comparison purposes, but it is becoming less and less.

On the consumer side, user's time is now going into all sorts of apps from IG, FB, SNAP, Twitter, Games etc... The only thing that google has some handle on is Games through play store and display network. I guess play store is another big money maker for google.

Google has a really good product line up - don't get me wrong, i love photos, docs, gmail, search, android, hangouts. But we seemed to have been missing something there with widespread adoption and monetization. Disclaimer: Former Google, current FB employee.

> However, most people now go to amazon.com to find out what they want to buy.

Maybe by frequency of purchase, but probably not by value. When I was car shopping, I used Google quite a bit to look up dealerships, reviews, etc. Did I use Amazon? Not even once. When I was looking at vacation destinations (prior to COVID cancelling all travel plans) likewise, I was using Google, not Amazon. Same with researching my last laptop purchase (which I ended up buying directly from Lenovo).

I use Amazon a lot, but I would guess that measured by dollar value, my off Amazon purchases are much greater than my Amazon purchases. Measured by number of purchases, of course, I've made many more transactions on Amazon than I have off of it.

you are likely correct. i use google flights, i would search on google for a car.

but fb is coming up with marketplace, ig has shopping related stuff, so i am not sure who the winner will be. It may end up with two companies sharing the market.

Not true. Search doesn't matter if people aren't there. I search on Amazon for stuff. I search on reddit for questions I have. I search on FB marketplace if I wanna get something 2nd hand. I post on FB group specific to topic. I use slack or discord to ask questions from live experts.

So yes they dominate the search, but "search" isn't what it could have been. I always thought how would someone ever defeat current giants. And while I heard the phrase I didn't get it, the next Facebook won't look like Facebook. Giants simply become irrelevant they are not defeated just forgotten.

The reference to Japan also brought up memories of Sony's brilliant play in the (admittedly declining) Camera market - starting off as a joke, then acquiring Minolta (not exactly a prestigious brand) and now being bigger than Canon and Nikon. In contrast, Pentax, which once ruled the Camera market (not to speak of Kodak..), is not even acknowledged by Ricoh today.

Tony Northrup's vlog on this is very interesting.


My take is that they probably just don’t care enough. It’s like Intel and their forays into all kinds of adventures outside of processors. Eventually these things are so small compared to their dominance in the core market that as an organization they simply don’t care enough to make it work.

It probably started as an initiative of someone high enough to get the resources, but the org never doubled down on it or change one inch of its roadmap to make it work.

Google had the first mover in modern email and maps and had very easy onboarding for both from search. Google somehow made Android work (which seems like a miracle in retrospect) and also Chrome (some would say they had to make Chrome work), but that’s all I can think of (it’s still huge, obviously, but also narrow in scope for an organization with that magnitude of resources).

As long as Google is filled with connected, talented, bored people it will keep pushing out all sorts of product experiments.

Google also doesn't live up to its standards, at least regarding Android apps.

Take Google Photos, for example. I have a biking app which automatically makes photos when a geofence is entered, stores them into DCIM/biking. So per ride I get a set of about 20 photos.

The first time a photo was made, Google Photos asked me if it should include that those photos in auto-upload. Yes, I wanted that.

Now, after the bike ride, I have to start the Google Photos app, wait for some seconds until it detects the photos, then manually close it (swipe up), reopen it, wait a couple of seconds until it finds another new picture which it missed in the first pass, close it, reopen it, wait, close, reopen, wait.

I need to do this about 5 times, because for some reason the app doesn't find all the photos in the directory at once.

It's not only this biking app and not only with DCIM subfolders. Any added folder which receives photos, and most noticeable with apps which take a sequence of them automatically is affected.

I don't get it how something as trivial as globbing a directory gets it past Google's testing standards.

There might be a weird setting that's not working properly. This is not happening to me.

Yes your app's bug is a systematic problem at Google. Then again these are the kind of click baity comments that are upvoted against Google.

it might be that the biking app is not calling MediaScanner after taking the picture.

This does sound like the app doesn't register photos correctly in the MediaStore. Or perhaps the phone didn't start the maintenance window to start the uploads - Android (like iOS) doesn't continiously scan filesystems since that would destroy the battery.

The biking app isn't calling MediaScanner at all. The thing is that once Google Photos is opened, it scans all the directories which it is supposed to watch. Then it also does find photos, you can see it getting populated with the new images one after another, but it always misses some.

That's what I'm referring to. In this context, you can't blame any third party app.

I was working on a video recording app and without calling MediaScanner, recorded videos would not show up in file managers or media apps. Android has a cache layer for file system access that needs poking if you create new or change files.

I'll agree with this. Wave launched with an incredibly broad set of things built-in... and then saw next to no changes whatsoever until its shutdown.

Granted, they probably didn't see the usage they hoped for... but that's extremely common for a brand new thing. It takes a few iterations to get something that fits the market. It's part of what makes waterfall strategies so problematic: you have one shot, and it'll usually miss.

It had been invite-only for most of its life and only opened to the general public for something like a month or less before its shutdown. It would have been dumb for Google to shut it down on "lack of usage" if their own artificial barriers were the reason for such.

Lack of usage is getting a lot of blame in these threads, but let's not forget it was most likely a Google+ related strategic decision. Wave was built with an XMPP-like or email-like federation model in mind and was shutdown in basically the exact same cohort with XMPP-based Google Talk and RSS-based Google Reader. Google was strategically no longer interested in standards-based (or standards-aspiring perhaps in the case of Wave) social networks.

This makes me wonder how much different things would be if Google had doubled downed on an Open Social (tm) model instead of trying to me-too Facebook et. al. by building a walled garden (more like a walled dirt lot I guess)

Sometimes they do the opposite, like trying to shove an obviously subpar product [Google+] down everybody's throat. Both approaches are bad strategies.

Tech follows a power law. Putting more work into something doesn't necessarily make it better. It could just be a bad idea no matter what. So failing fast is a good thing.

Google made the right call with Wave, finally. The better call would have been to not release it.

> The better call would have been to not release it.

Why's that? Wave wasn't a humiliating mess. It turned out not to really fit anywhere, but that's the kind of thing you learn by letting people try the product, no?

Google dogfooded Wave before releasing it publicly.

My question stands.

> It turned out not to really fit anywhere, but that's the kind of thing you learn by letting people try the product, no?

Google let people try it before releasing it publicly.

I still don't see why Google shouldn't have released it. I already pointed out that Wave's problem was that it never found its place, despite a solid implementation, which is just the sort of issue that makes itself clear after being released to the public.

Yes, Gmail FAILED Android FAILED Chrome FAILED. Let's not forget the elephant in the room SEARCH also FAILED.

Also saving capital but a 10yr bet on self driving cars.

To be fair, Gmail and Chrome are 16 years old, and Google acquired Android 15 years ago. Their main search algorithm of course was developed over 21 years ago. That doesn't leave a good track record for the last 15 years or so...

There's nothing on par with the main search product, but here are some of their products that seem successful from the 2010s: Google Drive, Google Flights, Google Hangouts, Tensorflow, and Google Photos. This isn't an attempt to be an exhaustive list, but these are all products that a lot of people on this board probably use frequently and have stuck for years.

Hangouts is an example of a failure: Google had a huge hit with Talk/Hangouts and for a while it was THE leading chat app. ...And then they mostly abandoned it, told everyone to use one of a half dozen different replacements which all failed as people moved onto other things. Had they continued to modernize and prioritize Hangouts, all of the newer chat startups would never have even existed. It "won" the early chat app fight and basically heralded the death of AIM, YIM, and MSN messenger, only to be murdered in-house by Google's own product teams.

Google Photos I think is reasonably successful, but mostly as "where your Android's photos end up". I am not sure how many people have a strong relationship with Google Photos. And I still hear about people upset about Google killing Picasa in favor of it. It is worth nothing that Picasa, the foundation of Photos, was not something Google developed, but something Google acquired from outside.

I disagree, Google Hangouts was a failure from the start.

Former Google Talk supported XMMP, including federation, which meant you could connect with multiple clients and talk with people talk with people that don't have Google accounts.

With Hangouts politics took over. They eventually told people the line about XMMP not evolving fast enough for their needs. But I remember their leadership complaining about their competition not playing nice.

So when Hangouts happened I saw absolutely no reason to use it. There were plenty of alternatives on the market already, at work we were all using Skype. And at that time I was working with former Googlers that preferred Google's shit.

You might give me the line about the general population not caring, bla, bla, but Google is in the habit of betraying their power users and early adopters. They did it with Google Reader. They did it with Inbox.

And I was actually glad to see Hangouts fail. It might have been a good implementation of yet another chat app, but few people cared.

For my social group.. what killed MSN was a combination of Facebook and moving to mobile.

Picasa was damn good software. I still miss it.

It gets ever-more rickety but it's still usable, and I still use it: the combination of decent face recognition, useful tools, fast UI, locally-stored photos is a combination that I value and I've yet to find an acceptable substitute on Windows or Linux.

If it gets too rickety for you, you should try PhotoStructure. It's locally hosted, can be run on docker/headless or on a desktop, the libraries are read/writeable across platforms, and it has the most robust asset merging and tag inference that I'm aware of.

My beta users are using it for free in exchange for their feedback. I'm hoping to release the final beta within a week. After beta there will be free and paid tiers.

Read more: https://photostructure.com/about/introducing-photostructure/

Changelog: https://photostructure.com/about/release-notes/

Have you considered Digikam? It does a lot more, but you don't have to use that.

Edit: why is this being downvoted? My point was that even the things listed as recent successes are not especially recent.


Flights is from 2011, nine years ago. And they bought the logic at the core of it from ITA.

Google drive is a genuine success, but from 2012, eight years ago.

Even hangouts, which is a failure, is from 2013.

What became Tensorflow appears to have begun seeing internal use in 2009-11.

https://en.wikipedia.org/wiki/Google_Flights https://en.wikipedia.org/wiki/Google_Drive

Discounting tensorflow is such a ridiculous argument. All of these products were used internally in some way or form before released outside. You will find some reason or the other to not count successes.

I’m not discounting it. My point is that it doesn’t seem especially recent.

Even by official public release it’s almost five years old.

The OP was arguing that google had had several recent successes, but their list tends to be 5+ years old. The point is that such a list is not especially convincing.

I should note that my point in mentioning the early origins of tensorflow was that google had a reputation for being more innovative in those years. So it is also meaningful that Tensorflow’s origins are from that period.

Google’s modern reputation is introducing a new chat app each year and deprecating the old one.

That's the reputation HN wants to impose. Especially the people who couldn't clear the interviews.

Do you have any recent successful products you can mention? I’m open to there being some.

I think you’re adding stuff into my comment that’s not there. My point was that nothing on the list was from newer than 2015. Except maybe google photos, but that came out of Picasa/Google+

Google Flights is great but of course Google had nothing to do with it. ITA Matrix was an eerily-useful flight-finding tool for many years before Google bought it and renamed it.

Photos has some kind of weird resolution limitation that turned me off the product immediately. (Plus I'm not crazy about Google having all my personal photos to mine.)

Tensorflow seems pretty good; I'd prefer not to be controlling high-speed computational resources with a super-slow interpreted language that cannot even multitask, but everything runs on Python nowadays.

I don't use the rest enough to comment.

Python can multitask. There's a global lock, but most time consuming things don't hold it, so these things run in true parallel.

Photos allows you to upload in original quality. You can always pay for services.

I wonder if anyone else was surprised by how useful Google Keep is, and how it doesn't seem to get any publicity (like people don't seem to know about it, but when I've introduced it to family members they've generally loved it).

Also I really like my Google home, Pixel phone, and Nest cameras -- privacy concerns aside :grimace:

I tried Keep, it is OK but the refusal to add an API is scary. It looks like it will not be a product integrated or supported like the others are.

Sadly, Keep fits the pattern of a good but not wildly successful product that will likely receive an "Update On" notice someday.

Drive, Photos, Tensorflow. Are you going to choose another arbitrary time period ? I"ll still come up with products.

I have also noticed how few people use any of these products :D

I think the real problem is when google releases a free[1] product they have a track record of discontinuing them seemingly (from user's point of view) at random so that people are beginning to feel that it's not safe to build on any new products from them.

[1] "Free" in this case meaning actually free, e.g. doesn't actively support their core spying + advertising business

Did you miss this?

> so if a product does not do well in a set time frame, it is shut down

Gmail was a success from the beginning. Android is doing pretty well. Chrome did a mix of growth because it was just better at the time, and ads which made it popular. They all did very well.

> Gmail was a success from the beginning

Interestingly, I don't think that was obvious from inside the company.

Maybe some day this will be declassified by one of the early PMs or engineers, it's a great story.

Interesting. The amount of invite code sharing / trading was huge from the moment I learned about gmail. I'd love to hear the story of the issues!

Some thing driving the invites was the promise of some GB of free storage, which some people abused in creative ways via their APIs (I remember a fuse driver storing files as mail attachments ...)

I remember a lot of buzz around invites for Google+, too.

And early on, despite a lot of anti-marketing that I still suspect was simply bought by FB, it was really great and vibrant. In its most original implementation, even.

Later on, the campaign of "complaints" about how empty it was got them to change bits of it, pushing "follow those people" on you that was filled with brain-numbing celebs and the like, and the magic was gone.

Whoever orchestrated that campaign should get a raise (and shouldn't meet me, as G+ was last web-based social media that I could earnestly use)

I believe it was an internal product before it was ever a publicly available one. IIRC it was successful inside and was public dogfooding at that point. Maybe I am mis-remembering.

I mean if they don't try things - "Google just sits on cash". If they do try things and fail "Google just kills products". HN will never be happy with what Google does. I just wish people here are consistent with these attitudes across companies.

There are other options than not do anything or kill what you don't need. For example the news reader could be easily split off or sold. theoldreader seems to be doing well even though they had to reimplement the whole service.

Wave got partially opensourced, after a while. I'm sure someone would buy Google trips.

Sure, Google itself wouldn't benefit much from those, but if they're killing the service anyway, it would improve the image.

> For example the news reader could be easily split off or sold

For things on the Google stack selling isn't really easy.

Things they acquired , like picassa, could probably be sold again, but there they wanted to migrate to Google Photos.

>> For example the news reader could be easily split off or sold

> For things on the Google stack selling isn't really easy.

If Google would build their new products using only publicly available GCP features, that would be a really smart move.

It would signal that they're confident enough about GCP to build their own products on it, and they can spin off products that are small successes but not large enough to move the needle at Google scale. Or if they're not profitable, open source them and allow people to run their own instance of it on GCP if they want.

Interestingly one of the few things they spun off - Niantic (ingress/Pokemon go) runs on Google app engine (and/or other Google cloud offerings)

Please apply the same standards to other companies. How many big tech sell off killed products. There are numerous IP issues etc.. Not to mention how ingrained these products are to internal infra. Google is pretty good about open sourcing internal libraries as a counter point.

I think there's a good chance Wave could have been Slack, if the team had focused on businesses instead of consumers. But at the time Google wasn't culturally set up for that kind of business, and gave up on Wave instead.

(Disclosure: I work for Google)

Yeah, when Hipchat and Slack became things I remember thinking - wow, Wave was perfect for this.

Google's loss for killing products. They've slowly made a reputation for themselves where no one bother's trying anything new Google makes because its not going to stick around.

I know right ? Slack is horrible for archiving (and the bloat, OMG the bloat); Wave being more integrated could've had an advantage here.

My number one memory of Wave is the incredible, outstanding, astonishing amounts of bloat.

When you hit "oh god, this is actually bloated as hell" in Slack, it's way after you're hooked and can't switch.

This was interesting:

"When the Maps project was completed, Lars and Jens moved on to develop a product dubbed “Walkabout.” Their ultimate goal was to answer a number of questions about how people communicated online:

Why are there divisions between email, chat, and document-based communication?"

That very first question they asked, they didn't get a good answer? Maybe that's why it failed?

I feel like there are very good reasons there are divisions between all those things. Those are all different things, and I use them all in very different situations, and those divisions are a good thing, and there's no reason I'd want one thing to do that.

I remember when they launched this, the technology seemed amazing, but I couldn't think of any reason I'd use it.

Wave had a huge promise. Today we communicate through various broken, half backed and decades old unimproved systems. Emails have hardly got any new features, even the simple things like "upvoting" in discussion is not allowed. Chats give little control over archival and searching. Forums on various websites are managed by uninterested parties who have little interest to improve them and may dissapear at any time. Group discussions quickly gets out of hand and become unwieldy. Despite all the progress, its super hard to apply good filters, surface interesting discussions and contributing to them meaningfully as an universal capability. There is a dire need for good communication software that integrates these various channels while still providing common set of powerful features that we can keep improving beyond 1980s standards.

I think the PERCEPTION (not necessarily the fact) of the difference between sinchronicity and asynchronicity is key here. Chat: synchronous. Mail: asynchronous.

And then: real-time collaboration and edition is kind of orthogonal and very different (and difficult) to understand (very few times do you do something really collaborative, even on a blackboard).

Seconded. And there's a special ring of hell for people that want me to attend synchronously while they slooowly compose what amounts to a lengthy email...

My favorite post-mortem failure mode of Wave was that people were not actually leaving chat trails behind for future searching.

As people were typing, the other people could see the response in real time and would begin typing a response, the first person would see a response was being typed, delete their message and start typing a response ad nauseam.

This, plus the infinite reply indentation lead to some very interesting conversations with my friends, which turned out to be an unreadable mess only days later.

It was great for real-time brainstorming though.

It's important to note that Google Wave has failed as a PRODUCT, not as a CONCEPT. There are a number of incredibly successful post-wave platforms, like Bitrix24, which is way more popular than Slack/MSTeams in Eastern Europe or Vietnam or South Africa, for instance (make sure you try https://bitrix24.com if you've never heard about it).

Here's my analogy. Google Hangout has failed (as did G+), but Zoom is booming as are Facebook or VK. While I agree with much of the analysis, it's important to understand the difference between concepts and products. This is very helpful when trying to understand why Skype had been struggling.

The biggest problem we had trying to use it (in two different organisations) was that we couldn't get the entire team on it as it was invitation only, with limited invites. That completely killed any reason to use it.

I enjoyed google wave very, very much. I knew it was way ahead of its time, but I hope it inspired some features in today's products (OneNote? obviously Google Docs)

Google Docs already existed. They overhauled the real-time collaboration system to be based on Wave's, but that's about it.

But somehow it was also at the right time. Collaboration tools were not so available and interactive at that point so everything new was accepted. If if was available today, I cannot imagine any development team using it to collaborate. It's just so colourful and dense (and fun ;)), I think most people would refrain to use it.

Google Wave was an inspiration to the product I am building right now[1], which is a personal wiki / note-taking app based around cards. The RTC isn't quite there yet (only a decade behind Google, haha), but there are a variety of other features (like multi-parent nested hierarchies) that I think are quite cool!

[1] https://supernotes.app

OneNote is older than Google Docs. Writely was released in 2005, Google acquired it and released GDocs in 2006. The first version of OneNote to be released was in Office 2003.

It was cancelled so fast after its hype, I recall thinking there must have been some kind of bust-up among different layers of management. Google cancel things, but usually not that quickly.

Particularly, I wondered if Google in the US thought too many engineers at Google in Australia were leaping onto the hype and they wanted them back on core business.

That is, however, purely uninformed speculation and guesswork.

All online communication methods are trending towards being Google Wave. I've believed this for a while and haven't heard a good counter argument (open to them!).

If you believe this then I think there are a couple of conclusions one could reach:

1. Google are smart. They saw what the logical conclusion of trends was and designed a product that satisfied those, yet users didn't want it.

2. Moves by communication tools in that direction, adding scope or "bloat" to their featureset is therefore doomed to failure, which fits with the cycle we see of productivity tool disruption -> bloat -> reduced productivity -> newer simpler disruptors.

I think it's a bit more nuanced, and is not quite either.

Google Wave had two relevant components: (a) its network infrastructure and (b) its user interface.

- by (a) I mean the extension of xmpp

- by (b) I mean not only the visual design and ux, but also the accompanying development ecosystem around integrations/media renderers/etc.

The innovations of Wave were, I believe, primarily in (a). Anything cool in (b) was merely a feature built directly because of/on top of something novel in (a)

I'm not sure why exactly Wave failed, but I suspect it was primarily because of the execution of (b).

And I think your point about online communication methods trending toward Wave are largely concerned with Google's innovations with xmpp

B) is Google’s Achilles heel. Their corporate culture just doesn’t take UX seriously even as it kills products — I’ve been seeing GCP fall into disfavor because the UI is second-tier, and that’s something where a tiny team could make big improvements very quickly because there’s so much low-ganging fruit.

It didn't even seem like (b) failed so much as Google decided to leave the (a) game entirely because standards-based or standards-proposing social networks were out and the Google+ proprietary walled garden was in.

Google uses many if not all of the features from Wave in Google Drive. Thus I do not consider it a pure failure, although it doesn't look like it panned out for them the way they thought it would.

Some of the features from Wave also made it into Google Docs, albeit somewhat evolved. (Specifically here, I'm talking about the algorithm for handling collaborative edits)

FWIW, i was part of the team that was working on a similar tool back in 2009-10. Google Wave blindsided us and we thought we would be killed before launch. We launched anyway and failed to get much traction: You can see our intro video here: https://www.youtube.com/watch?v=RjezWHeNOas

Another issue at the time was scaling wave. I remember there being issues as threads grew in size.

I think it was pretty clear that Wave wasn't going to take off as a replacement for email (or, much else) once the fifth or sixth person asked how you could "make a wave send you a note in gmail when it updated".

I remember tinkering with it at that time. It was an exceptionally good tool, a Swiss army knife, but at the company we could not find a real use case for it. Alas it was a small team. Most businesses had their more or less working solutions, and habits hardly change. I think it was shut down too early, not giving enough time to get in. Providing a more business-oriented solution instead of the near-bare "API" and being a tool, it could have been the best collaboration tool of the current situation...

Still so, the Wave announcement was for me at least quite a memorable moment in the history of the web, showcasing what's possible when embracing the new possibilities of the internet and pushing the boundaries of what people think about.

A lot of inventions, innovators or hackers don't achieve big scale but still have impact with the ideas they have put forward, no need to label it a failure right away :)

I think the suggestion that the failure was a question of timing, "positioning" and a poor launch is naive. It wasn't a very good product. It had some nifty features but they were all jumbled together into a mess that didn't solve real problems. Was Google even using it internally?

Things like this is in my experience not so much a technical problem as a people problem. No inside info, but if the PM manager doesn't look eye to eye with top management (and if they are unable to solve it for any reason) then the project is going to shut down.

And it will be blamed on the tech, or whatever.

Google Wave was ahead of its time. It has more to do with timing than technology or product.

Wave limited new users to the point of shooting itself in the foot. If you had a small group/team of six, and only five had Wave access you couldn't use Wave because of the sixth.

You don't grow a team-centric tool by not allowing teams to use it.

I was so excited when it was announced.

I got my invite early November 2009, then... Well then I was the only one of my friends using it, so my interest quickly decreased. It's sad it never took off, the idea was really cool.

The preview email: https://postimg.cc/dkGMG0sZ

I wonder how much of Wave's death was from resistance from inside Google(GMail team, or even other execs). Fear for the destruction of GMail even from another Google product.

Google Wave failed as a product, but as a technology it was successfully absorbed into the Google Docs suite with Docs, Sheets and Slides having real time editing capabilities.

Quip is the modern day Google wave. They copied Google wave to a T and for some strange reason the market has accepted their product as useful.

In light of recent events, would Google Wave still have failed if they launched in 2020? It’s a curious question for me.

Teams and Slack have covered the collaboration market quite well (and there are others as well). What really took off are video conferencing solutions (e.g. Zoom).

Google had products here but were never able to offer a single, easy solution. If they had combined Duo and Hangouts from the start, maybe they'd have the success that Zoom now has.

Its success aside, I thought Wave was really cool. It worked pretty well for school group projects.

I was a fan of Wave, and then Apache Wave. I liked Apache Wave’s simpler UI, but I found the code base difficult to understand. At least for my skills, it was not hackable. It has been several years since I have spun up Apache Wave, not since Apache Foundation put the project on ice.

I loved Google Wave. I even wrote a Gadget for it

Quip is the modern version of Google wave.

If Google Wave launches in today's climate, will it survive?

I assume that was meant to be a conditional, and if so, the answer would be an unprovable counterfactual.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact