Predicting Google Product Shutdowns

http://www.gwern.net/Google%20shutdowns

by kiba • 13 years ago

261 139 13 years ago

orbitur2• 13 years ago

Interesting post, but oh my god, the styling on this site is hurting my brain.

gwern15• 13 years ago

I know I designed the styling mostly to appeal to myself, but serious question: what's wrong with black text on white background? That's basically how books have been printed for hundreds of years.

orbitur2• 13 years ago

Made an imgur album: http://imgur.com/a/8XUDP

First is with AdBlock on, the second is AdBlock off.

It's better once you've scrolled down, but everything is poorly mashed together at the top. Multiple words in the left column are badly wrapped.

xnxn• 13 years ago

Just checking: is it supposed to look like this? http://i.imgur.com/PWbZHW5.png

The color is totally fine, it's really just that sidebar that's odd.

slater• 13 years ago

The issue is not black text on white background, but the layout choices.

This is what it looks like when I load the page: http://i.imgur.com/kmMsyEg.png

And here's how it looks after some fiddlin': http://i.imgur.com/JhwwjfY.png

molecule2• 13 years ago

Great article and analysis.

One answer to your question, "what's wrong with black text on white background?":

http://ianstormtaylor.com/design-tip-never-use-black/

dnr• 13 years ago

Monitors are not paper. They're much brighter.

chops2• 13 years ago

The spacing (or lack thereof) on the site is very distracting. The left menu has zero pixel spacing from the content, and so the side menu basically runs right into the content. The table of contents on the right isn't obvious that it's a table of contents for the article, it just looks like another menu of some sort on the right.

I did some quick restyling just to satisfy my curiosity with some notes. I know it kind of makes it looks like wikipedia, but wikipedia is easy to read.

http://imgur.com/G5QfOua

itafroma• 13 years ago

It's not really that bad, just a few quirks:

- Left sidebar has no right padding

- The right sidebar loses its left padding in narrow viewports

If you want to go hardcore, there are a couple of other things, based off of James Felici's Complete Manual of Typography :

- The measure (line length) at its max width is about double what it should be for the font size (which is inheriting the browser default, usually 16px). Measure in picas (1/6 inch) should be 2-2.5 times the font size in points (1/72 inch, but a straight conversion of 16px = 12pt is usually fine).

- Currently, the line height is set to 130%. If using your current measure, the leading should be about 5 points, for a line height of about 141%. If you halve the measure, the leading would be about 1.5, for a line height of 109%. However, I'd still set the line height to be ~140% because you have a lot of superscript and your links have underlines that go below the descenders. Leading in points should be Measure in picas / font size in points, rounded to the nearest half point. Increase if you have things that go beyond the descenders (like you do) or decrease if you don't have any descenders (e.g., all caps).

antoinec3• 13 years ago

Nothing's wrong for black on white for books. The thing is I don't read your site on a piece of paper but on a screen (and I'm probably not the only one).

mratzloff• 13 years ago

Well, there are several problems.

First, to address your point. Paper is not pure white, and ink is not pure black. Books are also not backlit. If you want to recreate the experience of reading a book, use a slightly off-white color and a dark colored ink (for example, a very dark blue or brown).

Second, most everything on the page is way too cramped.

Third, you need to constrain text width. It's a lot of work to scan 1000 pixels for each line.

I made some quick changes in CSS to address the basic readability problems, and uploaded a screenshot:

http://new.tinygrab.com/d5ac58cc253b078fa42a97b3c35417d28563...

(I was playing around with the links before I decided I was getting too far into the weeds, but the links in #content should have the underline. The links on the side shouldn't.)

CSS changes:

    html {
      background-color: #FCFCFC;
    }
    
    html, a {
      color: #303C3C;
    }
    
    blockquote {
      font-style: italic;
    }
    
    #header {
      margin-top: 50px;
      margin-bottom: 30px;
    }
    
    #sidebar {
      width: 150px;
      margin-right: 50px;
    }
    
    #content {
      width: 550px;
      margin-left: 200px;
      padding-left: 0;
      line-height: 150%;
    }
    
    #TOC {
      width: 200px;
      margin-right: -235px;
    }
    
    #TOC li {
      margin-top: 0;
    }

yew• 13 years ago

I actually like it quite a bit . . . The lack of triple-nested faux windows is nice and I usually set things to display compactly anyway, so there's that.

It looks like some people are having font issues, though. I'm not seeing any overlap between the columns, but there's a lot visible in some of the linked images.

bifrost4• 13 years ago

Its definately a bit hard to read as it is, its a bit more academic doc vs blog.

For lulz I fired it up in Lynx and wow does it look bad there too...

veb• 13 years ago

I usually try and use Readability to read long and interesting articles... but it did not like Gwern's site.

http://www.readability.com/articles/5jqxbits

gwern15• 13 years ago

Hah, what's going on there? I don't understand how it can take nice regular Pandoc markdown-generated HTML and crop it down to nothing but the source code. Maybe you should file a bug on Readability.

The_Sponge• 13 years ago

I don't think that Google Voice will get shutdown. Google needs a unified communication product, tying Google Talk, Voice, and Hangouts together.

w1ntermute5• 13 years ago

This is said to be coming in "Babel."

jonemo• 13 years ago

This is the first time I heard about changes to Voice and I really really hope they just leave Voice alone. This is the one Google product that is deeply integrated into my life and I rely on it continuing to work (it's also the only Google product I pay for). Sure it could become better, but lately when Google makes things better they usually get worse (at first, or forever).

bifrost4• 13 years ago

I think this actually fits with what I think about Google overall. Search is great, everything else is just a mediocre tie in to search dollars.

Google is attracting more "nesters" and less "innovators" from what I see as an outsider, which is likely good for long term products but bad for everything else.

Glass is already a bad word (RE: glassholes), Plus is a wasteland, Voice is only popular due to branding, as is Calendar and Docs.

I've seen announcements of products that appear to be more functional than Voice/Cal/Docs on HN with better security and performance, but I think the likelihood of those taking over is probably low until Google gets out of those markets. The products are better, but the mindshare isn't there yet.

turing2• 13 years ago

Glass is piquing more interest than any consumer electronics device since the iPhone, Plus is thriving, and Docs is by far the best cloud office suite I've seen. I am by no means saying that Google is the best in every market they enter--some of their offerings are pathetic compared to the competition-- but your claims of across the board mediocrity in their non-search products is severely overblown.

zone411• 13 years ago

Google+ doesn't appear to "thriving" to me. Here is one indicator: http://www.google.com/trends/explore#q=google%20%2B%2C%20goo... (flat line). For comparison, add "imgur" or "hotmail" to the input. Here is another where it looks a bit better: http://siteanalytics.compete.com/plus.google.com/. However, considering the amount of integration with their other products, I can't see it as anything more than a disappointment for Google.

rubinelli• 13 years ago

The concept behind Glass is fantastic. For all the talk of "cyborgs" and "glassholes", it's clear that it struck a nerve. What we don't have enough information to say yet is if Glass is the new iPad or the new Newton.

samuellevy• 13 years ago

I don't know why people keep insisting that G+ is a waste land. If you personally don't use it, or see a use for it, that's fine, but there are many people in the world who are not you.

Personally, I get far more value from G+ than I do/ever have done from twitter, Facebook, MySpace, or any of the many social networks over the years. I have a use for it, which it fulfills better than any of the other services out there, and so for me (and the thousand or so people who I'm connected to at various levels through G+), it's far from a waste land.

G+ isn't for everyone, and that's fine. It is a product which is now highly polished, and provides exactly what I and many other people have wanted from a social network. The exact numbers of users who actually actively use it may not be clear, but if it was a waste land, Google wouldn't be integrating all their services so tightly into it.

Google is not a stupid company, and they're not about to "break" all of their services to drive adoption of a service which can't stand on its own. G+ is a good service, and is providing massive amounts of value to Google. Reader was not creating new value for them; wave wasn't getting adopted; Google had reasons why continuing to support those products was a losing bet.

They are integrating so many of their services into G+, that you can pretty much guarantee that it's a winner (for them, at least).

</rant>

bernardlunn2• 13 years ago

Social is inevitably a winner take all market, so in this case the ignoramuses who never use G+ who declare it dead are probably right. Its not about great features, just ask FriendFeed.

anoncow• 13 years ago

They are integrating search so tightly with G+ that websites/authors have to get a G+ profile to make authorship data visible in search.

A fairer way would have been to use microdata alone to achieve the same thing. This would have been acceptable to other search engines too.

chrisbolt• 13 years ago

> I don't know why people keep insisting that G+ is a waste land. If you personally don't use it, or see a use for it, that's fine, but there are many people in the world who are not you.

It's more that nobody I know uses it. All of my friends are on Facebook, and everyone is talking about Facebook. If I try finding my friends on Google+, it looks like a wasteland.

> They are integrating so many of their services into G+, that you can pretty much guarantee that it's a winner (for them, at least).

Why should I care if it's a winner for them? Google has a huge reach, and they are leveraging that reach to push Google+, but people are still uploading photos and updating their status on Facebook. Google+ may have many unique users, but their use is far more shallow than their use of Facebook.

bifrost4• 13 years ago

>If you personally don't use it, or see a use for it, that's fine, but there are many people in the world who are not you.

I wholly agree, but G+ is barely better than tribe.net (for me and my peer group) when you compare it to what Facebook has in terms of content and users. There's also the whole "Nym" issue. I also can't understand why people are excited by giving Google MORE data, the "don't be evil" days are clearly over with.

DigitalTurk3• 13 years ago

> G+ isn't for everyone, and that's fine. It is a product which is now highly polished, and provides exactly what I and many other people have wanted from a social network.

Is it possible yet to change the visibility of old posts? The absence of small privacy features like this—features that FB does have—always was very off-putting for me. When Google+ was launched they made a huge deal out of Circles, but it strikes me that since then Google has pushed Google+ more as an alternative to Twitter than as an alternative to Facebook (and I don't like Twitter much).

Anyway, I used Google+ so little that I deleted my account in a petty rage when they shut down Google Reader. It's actually nice to have one less social network to worry about.

antoinec3• 13 years ago

Google Maps/Earth, Android, I wouldn't call that mediocre and not innovative...

bifrost4• 13 years ago

Maps/Earth are cool, but they've stagnated, Barely changed in the last few years.

Android is neat, but its not innovative either. The only reason I use one is because of the wide variety of HW available.

microtonal• 13 years ago

And yet, I have to resort to Garmin software and maps to do real trip planning (cycling, hiking, and even driving).

snogglethorpe3• 13 years ago

You're nuts.

Google certainly has some mediocre "products," but they have many great products as well besides search.

Gmail, of course, essentially sets the bar for online mail, and is still leaps and bounds better than any of its competitors.

Google maps is very, very, good, vastly better than most of the competition (a point made starkly clear by Apple's recent maps C-F).

Google+ is, technically, absolutely excellent, much better than FB. If all your old high-school buddies are only on FB, it's a shame, but there's not much Google can do about that.

Android is vastly better than iOS (I use both, and using iOS always makes it feel like my hands are tied behind my back).

...etc.

Google has its downsides (e.g. their online support sucks), and you may not like them for other reasons, but they're very, very, good at writing software...

shurcooL3• 13 years ago

> Android is vastly better than iOS (I use both, and using iOS always makes it feel like my hands are tied behind my back).

For me, I feel like I have 2 hands when using iOS but 17 hands when using Android. It asked me to pick one of 3 power-savings modes for Wi-Fi when in sleep mode. I don't want to have to care about that (nor do I want to pick one of the three options without thinking, because then I'll not have an optimal configuration). I have enough things to tweak on my computer.

philwelch5• 13 years ago

> Gmail, of course, essentially sets the bar for online mail, and is still leaps and bounds better than any of its competitors.

If Gmail keeps lowering the bar instead of raising it, that's not going to be true for much longer.

threeseed• 13 years ago

The days of Google Maps being vastly better than Apple Maps are fast disappearing. It is about 95% parity for Australia and UK at least and any differences are being eliminated rapidly. And the touted advantages of Street View aren't there since the data is at least 4 years out of date in some countries.

The fact is that Maps aren't something you can dramatically keep improving. You reach the "everything is mapped" point and then it becomes a keeping everything up to date scenario.

new2994• 13 years ago

> Android is vastly better than iOS (I use both, and using iOS always makes it feel like my hands are tied behind my back).

It's really in no sense "vastly better" the market pretty much bears this out.

From a developer/technical standpoint it can be a pain that you have less access to device features on iOS and the AppStore review process can be a nightmare. But developing on XCode isn't bad, as is the UIKit architecture. It's unfortunate that iOS isn't open source like Android, but I can see no way from a user or technical perspective that it's "vastly better".

wyclif• 13 years ago

As soon as I read "Plus is a wasteland", I knew you had no idea what you were talking about, and stopped reading.

bergie• 13 years ago

Indeed. Communities has made G+ something I read multiple times a day, at the expense of Twitter

jere• 13 years ago

I skipped ahead to the "10 most risky and likely to be shutdown products", read it, then did a double take "wasn't I supposed to be reading about Google?" Never heard of any of those.

Interesting poems. I was certain Ozymandias would show up though.

gwern15• 13 years ago

> Never heard of any of those.

Yes, and that'll be reflected in the hits data and be one reason they'll be rated so risky...

I love 'Ozymandias', but unfortunately it's become so overused it's cliche. So I picked poems that I was reasonably sure almost no reader would ever have seen (the exceptions: the _Julius Caesar_ quote, which I parodied; and the Basho haiku, which I gave my own rendering of) but were hopefully still good.

tomkarlo3• 13 years ago

That's because he's using search hits as a primary measure of how much services are being used, which (maybe) kind of works for consumer services but doesn't work at well at all for services used by developers (at the least, they should be judged on their own scale). Which is why you see Cloud Messaging popping up in that list - it's just not something many people talk about, but that doesn't mean it's not heavily used.

nijk5• 13 years ago

http://www.google.com/ig/directory

Ozumandius was never shut down, it was left up in ruins.

rachelbythebay2• 13 years ago

I have some thoughts about the spikes on the death dates.

September: all of the interns go back to school. These people who exist on the fringes of the system manage to get a lot of work done, possibly because they are free of most of the overhead facing real employees. Once they leave, it's up to the FTEs to own whatever was created, and that doesn't always work. I wish I could have kept some of them and swapped them for some of the full-timers.

March/April: Annual bonus time? That's what it used to be, at least, and I say this as someone who quit in May, and that was no accident. Same thing: people leave, and that dooms whatever they left.

obviouslygreen• 13 years ago

March/April: Annual bonus time?

I'm not sure if it's good or bad that I somehow never managed to work for an employer that offered any kind of bonus, ever. It granted a perspective of "this is what you get, and no more, unless you fight for it," which... again... can be good or bad depending on how you look at it.

I wonder how many 20- and 30-somethings are now or have recently been in a position where bonuses are reasonably common, and how that correlates with the size/type of company they're working for.

philwelch5• 13 years ago

At some companies, signing bonuses are starting to become standard.

gresrun• 13 years ago

Bonuses are the norm in the finance industry. I worked at a financial research company for five years and got a bonus of roughly 10-20% of my salary each year, depending on that year's profits.

consz• 13 years ago

Funny, I have the opposite opinion. Without a bonus, it just seems like "this is the work you need to do, and no reason to make any effort to do more". With bonuses that scale with your contribution, I have far more incentive to put in effort above and beyond.

Permit• 13 years ago

>September: all of the interns go back to school.

This isn't true. Internships aren't only during the summer for many schools. I'll be starting a four month internship in September as will thousands of my peers.

turing2• 13 years ago

It's true that Google has interns year-round, but I'm willing to bet a sizable majority of them are over the summer.

gojomo• 13 years ago

September: Googlers return from Burning Man with added appreciation of impermanence and the value of cleansing fire to clear the way for new creation.

gwern15• 13 years ago

The September one is a cute idea, but how many shutdown things immediately had a bunch of interns? For example, with Knol, there weren't any site updates or functionality changes for something at least a year before the final shutdown announcement went out, so how could that have had anything to do with interns? The spike in September is so huge I'm not sure interns explains it.

Maybe people are going on vacations (not necessarily Burning Man) and returning with renewed perspective and distance from things and are more willing to close down things. Not sure how one could test this...

rachelbythebay2• 13 years ago

It could also line up with Perf season.

I'm happy to report that I have successfully forgotten exactly when perf happens. There was one supposedly optional perf and one supposedly non-optional perf per year. I'm not sure exactly how that would affect cancellations of projects, though.

One thing I do remember seeing was a LOT of half-baked crap being foisted upon people (especially internal customers) because the end of quarter was approaching. Someone had to make their OKRs, and they didn't care how...

nijk5• 13 years ago

Or it is twice a year company wide review and financial planning, not major product decisions driven by temporary employees.

I love how people say Google is awful because it is a corporate hell where engineers have no autonomy, and then claim that most products are launched and run by rogue indivuals.

molecule2• 13 years ago

For anyone else that is wondering which Google product lived for one day ("Days: Min: 1"):

http://en.wikipedia.org/wiki/Google_X

bernardlunn2• 13 years ago

Great find. It must have been total chaos - good chaos - in those days.

nijk5• 13 years ago

Seems more like a Doodle (homage to OS X and its dock) then a product that lived for one day.

Although the March 15 2005 date was a bit off any OS X anniversary I knowm

timmm3• 13 years ago

You do realize that the 30% that Google takes from Android Developers from App sales doesn't go to Google right?

throwaway2048• 13 years ago

where does it go?

timmm3• 13 years ago

To the respective mobile carriers.

gwern15• 13 years ago

No, I don't.

But even if they don't, Google is still serving ads to Android users AFAIK. While browsing on exactly that topic (pondering whether to flip the profit bit on the Android entry), I ran across an estimate of $1 billion: http://www.forbes.com/sites/ericsavitz/2011/02/08/googles-an... This made the decision an easy one.

dudus2• 13 years ago

You do realize it does right?

chris_wot• 13 years ago

Ah, a study based on something Charles Arthur said?

I once had an email conversation with Charles Arthur, it didn't go well...

http://en.wikipedia.org/wiki/User:Ta_bu_shi_da_yu/Charles_Ar...

gwern15• 13 years ago

More like, 'a study prompted by noting serious flaws in an Arthur column and deciding to do better, while borrowing perhaps <1/10th of the final dataset from Arthur's data' (he had ~30, I wound up with 350).

mountaineer• 13 years ago

Definitely agree with Feedburner and Alerts being likely targets based on how well (or not well) they're supported currently.

jensenbox• 13 years ago

If anyone is interested in helping out, I have created a timeline for browsing such services: http://jensenbox.github.io/timeline/

gwern15• 13 years ago

That's a nice start. (It's a good idea - I was actually trying to do a similar 'stacked lines' plot to show the lifespans of all entries in my dataset, but I gave up at the time because it looked too hard and grotty to do in R.)

benaiah2• 13 years ago

> Gmail seems like a service for the ages; but once upon a time, so did Hotmail (even in 2012, 8 years after Gmail’s launch, still the #2 email service) - yet, RIP Hotmail (1996-2013).

That's silly. Hotmail didn't go away; it was rebranded. All the email and accounts were transparently migrated - you could even use both front-ends at the same time for a short period. The important part is, Microsoft still runs email and nobody lost data. Comparing it to the Reader shutdown (or a shutdown of Gmail in a similar fashion) is an apples-to-oranges comparison of the first degree.

gameshot911• 13 years ago

>The important part is, Microsoft still runs email and nobody lost data.

Unless you fail to log in at least once every 270 days, at which point your entire account and everything in it is wiped.

jacquesm2• 13 years ago

That's probably a safety precaution but if it is like you describe it is also a risk. It could be a safety precaution to make sure that accounts do not eventually become compromised in the case of deceased person.

But if the account is deleted entirely that could indicate the address can be re-used and in that case any password reset mails and other important email might land in the new owner of the address' inbox.

hga• 13 years ago

Wasn't that true before the Hotmail -> "Outlook" transformation? I.e. a Microsoft free email policy that as jacquesm notes could have good reasons behind it.

eli• 13 years ago

It's always been that way. In fact, I'm almost certain the time was shorter in the early days.

Proleps• 13 years ago

> That's silly.

It is. The rebranding and new interface are a huge improvement and a sign Microsoft is taking there email service more seriously. I don't really get how someone could interpret it as the loss of a service.

joering2• 13 years ago

In the near future someone may say: that's silly. Gmail didn't go away; it was rebranded. It is called Google+ now.

benaiah2• 13 years ago

Except Google+ and Gmail are very different things. Hotmail and Outlook.com are exactly the same thing with a different UI.

cremnob• 13 years ago

Google Finance is definitely not being maintained and I wouldn't be surprised if it's killed.

robk• 13 years ago

Picasa seems unnaturally safe in this model. It's also stopped iterating generally and Google's acquired a totally separate mobile app for photos (Snapseed). And Google really doesn't make any material amount of money off Picasa either.

buro9• 13 years ago

And Linux support (which was only through Wine) was officially dropped.

Even before the recent spate of services being shut-down, shortly after the Linux support was dropped, I moved Picasa to a kill list and started migrating what I was using it for to something else (Dropbox for storage and sharing).

I presume it's going to be killed, to me that's what the product state signals.

jcastro• 13 years ago

Dropbox also has photos now, when you log in on the bottom left there was an opt-in for the beta testing. It's really nice compared to Picasa/G+'s "2 different photo services but not really" design.

abraham• 13 years ago

Picasa is slowly getting replaced with Google+ photos. I don't expect it to be around all that much longer.

gwern15• 13 years ago

Yes, Picasa's rating seems wrong enough to me that I called it out in the discussion as probably wrong. (Not sure how one would improve the model... how do you objectively rate or quantify something like 'stopped iterating generally'?) And it make not make 'any material amount of money', but it is still selling advertising and whatnot on Picasa: https://support.google.com/picasa/answer/166073?hl=en So the binary variable for 'profit' gets flipped...

Snapseed is not doing any better, though. IIRC Google already shut down an app for Snapseed.

I think Picasa might wind up being merged totally into Google+, although by my established criteria that would not count as a shutdown.

Su-Shee• 13 years ago

You mentioned in passing that you have a feeling that Picasa isn't surviving for long due to it's "old-fashionedness".

Considering how important "usefulness" and "usability" are by now, maybe add a criteria along those lines - how large is the "your mom uses it and likes it" factor, how "contemporary" are the UIs, how much effort puts Google into polishing it (it does a lot with G+ and Gmail for example) - something like this.

Another criteria could be "level of integration" - how much can you use a Google product as a standalone project or how deeply connected is a product into another one (e.g. Picasa used outside of G+) - which might in the end indicate not a direct shutdown but a product's dissolution into another one.

Shooti• 13 years ago

"Picasa" the desktop app maybe, but I think the reason Picasa web albums/API lives on is that it provides the backend for Google+ Photos and they don't want to break existing apps using that API just to rename it, so continue to build stuff on it. Most recent example would be the Photosphere viewer API they just released: https://developers.google.com/panorama/web/

ComputerGuru• 13 years ago

We just launched our Google FeedBurner replacement in anticipation of its upcoming doom: http://feedsnap.com (release info http://neosmart.net/blog/2013/worried-google-will-kill-your-... )

We're planning on "launching it" on Monday.

newnewnew• 13 years ago

When will blogger be killed and rolled into Google plus?

gwern15• 13 years ago

If you believe the model, roughly 'when hell freezes over'. I find it a little hard, but I guess we'll know who to believe during the next 5 years if Blogger is indeed killed.

k3n• 13 years ago

Regardless of the actual subject, it looks like an excellent analysis. I haven't verified the maths, but I like that he includes every possible calculation, source code, and external reference that was used. A very academic and data-driven approach.

EdiX• 13 years ago

Am I the only one who sees Google Groups at risk? It could be seen as a competitor to Google Plus the update it's getting is awful and looks like something designed in 2009, the updated before that happened a decade ago, it's not very popular with the general public... it sounds a lot like Google Reader to me.

gwern15• 13 years ago

You'd think so, but the comparison to Google+ doesn't seem very apt - it shouldn't be a threat any more than Gmail is (but something to be leveraged).

klapinat0r• 13 years ago

It's been a few years since my 3 stats courses, so pardon if I'm wrong, but aren't his p-values for the logistic regression kind of... important?

Intercept and log(hits) sure, but the rest have rather high p values indicating that it's likely to be present even if it has no effect (lower p-value, more robust, < 0.05 => < 5% chance of incorrectly stating it has an effect (when it does not)).

Or am I missing something basic here?

Predicting Google Product Shutdowns

Related Stories