Our visit the other day to the GooglePlex was extraordinary on a number of levels. At its simplest, the Gmail group opened its kimono to bloggers and what’s left of the mainstream media - full stop. We were asked not to live video the announcements of Gmail Labs, and to clear photographs with the team in case we accidentally revealed some strategic details. But the tour of the Gmail team cubicles, the Google Reader unit, the spam guys, and the user testing facility was fascinating, particularly as it became clear how much was being done with a very small group of code warriors.
By contrast, when Jeff Raikes‘ replacement as president of the Microsoft Business Division, Stephen Elop, took over the unit encompassing Office, Dynamics CRM, and Unified Communications groups, ComputerWorld estimated he controlled at least 26,000 workers responsible for generating fully a third of Microsoft’s 2007 revenue. Certainly Gmail doesn’t represent all of Google’s Apps (Office) investment, but enough of it to make clear how devastating this nimble strike force is and how catastrophically it can undermine Office.
The “features” rolled out in Gmail Labs range from trivial to obvious, but the power is not in what Google engineers have produced on their 20% time initiative. Rather, it’s the feedback loop that results when users can recompile Gmail with a personalized addition of such features. First, they vote with their feet, sending signals to the team not just of what they pick but how long they use it, when they discard it, and what they pick next. It’s a million-plus user testing facility for free, an easy way of extending Google’s original strategy of scaling up as usage grows, and most importantly, a wedge for the viral Greasemonkey development community that the approach aims to stabilize.
Take the first 13 or so apps and throw them out; I like adding photos to chat and bookmarks to email yet probably don’t care if they go away. Now look at the functionality exposed in these sample tools: extending chat for example, which boasts in addition to Photos tools to hide status messages and add keyboard shortcuts to things like Focus chat contact search and the fascinating Focus last chat mole. Moles, by the way, are those little Gchat windows that open up and array intelligently at the bottom of the Gmail window or can be popped out to stand alone.
Let your mind wander a little and you can see how significant this granular control of the Gmail console can become. I asked product manager Keith Coleman whether a Labs feature could remember the position of chat moles once resized and placed outside the Gmail container. Add a focus pull to your favorite mole, say the Twitter XMPP gateway, and then use a version of status removal to filter Track messages. Well, you get the idea, and so did Keith. For free. From highly motivated Plan B users.
Of course, Google is not the only cloud that can take advantage of this iterative feedback loop. On Friday’s Gillmor Gang, I asked Google API lead (and former Hailstorm architect) Mark Lucovsky whether he agreed with many that Microsoft Live Mesh was just about replication of data, or like me, that it was the tip of a Titanic-sized iceberg. Factor in Coleman’s insistence that any third party API was fair game for inclusion in a Google engineer’s Gmail Labs feature, and that would by definition include Microsoft Mesh.
Coleman owns Google Reader and Gchat as part of Gmail Plus, so it seems likely that future services will tie in Apps, aid in constructing enterprise versions filtered around information aggregation, automated push services out along the XMPP bridge, and even developer tools that use the combined services to fashion bug-tracking, code generation, and collaborative project archives. Today, it’s a personalized recompiled Gmail; tomorrow, it will be extended to affinity groups around gesture-mandated dynamic builds that adjust based on behavior and proliferation of open standards.
Part of Google’s impetus to do this was to route around the instability of Greasemonkey scripts which stressed out the Javascript architecture with ad hoc strategies. But Greasemonkey will return with a vengence as soon as developers outside the company realize they can take off from these approved experiments and wire up external API’s. Quick Links, for example, could be extended with a TinyURL-like bridge to encapsulate FriendFeed conversations and export them to the Twitter or Mesh cloud, or be written into Google Reader’s new Shared Item Notes and broadcast to Gmail contacts under user control instead of the current “Friend” contact mining that damages GReader’s privacy integrity.
Inevitably, the combination of user feedback and behavior and external pirate Greasemonkey innovation will reach a boil. Ray Ozzie and Stephen Elop are on notice, and have their work cut out for them. As Lucovsky notes, Redmond is not just sitting idly by: check out Scott Guthrie’s description of Silverlight networking improvements such as Cross Domain Sockets and Background Thread Networking. But while Gmail may not have a third of Google’s revenue, they now have a hell of a lot more developers working for them than last week.
See all
Post
Zzzzzzzzzzzzzzzzzzzz…
Zzzzzzzzzzzzzzzzzz…
Zzzzzzzzzzzzzz…
WTF Does this mean:
Today, it’s a personalized recompiled Gmail; tomorrow, it will be extended to affinity groups around gesture-mandated dynamic builds that adjust based on behavior and proliferation of open standards.
I have heard people who cannot log into their Gmail account for days. I cant believe this could happen to Google
Steve, sorry, you sound a bit ignorant to what it takes to produce a s/w product.
You can’t compare Office or Dynamics CRM with something like GMAIL.
They are totally different when it comes to the guts of it.
Not sure if 26,000 people are justified, though.
I find it easier to build solutions for the Web than for desktops. Less effort. More frameworks, more reuse.
Go ask someone who knows.
Both Dynamics and Office have an online version
How much of a change even small developer teams can make is also very visible in the open source software sector and on the web. Even small start-ups can become a serious thread for well-established giant companies.
On the web, e.g. Google and Facebook started small and quickly became multi-billion business.
In the open source arena: MySQL also turned into a one billion business. And others will follow as the business world more and more embraces open source software. The core developer teams of open source projects are often small, but with the help of the community, their software develops and spreads much faster than proprietary closed source software.
I expect a lot of movement in the booming business intelligence sector as open source companies like JasperSoft ( http://www.JasperSoft.com/ ), Pentaho ( http://www.Penatho.com/ ), and Rapid-I / RapidMiner ( http://www.rapid-i.com/ ) gain traction. Open source business intelligence and data mining solutions are keaving their niches and increase their market chairs. These comparatively small open source companies can save their customer companies many millions and may be even billions in costs and to increase their efficiency and sales significantly. So, while they are small companies, they have a bigger and bigger impact on the economy.
Just my 5 cents. ;^)
Cheers,
Frank Xavier
These posts demand a lot of concentration and time to read and don’t really reward the reader for the time and effort.
This post pretty much boils down to - GMail labs will allow Google to leverage off the developers to grow Gmail/Apps, which is more efficient than Microsoft’s model of inhouse development.
@4: Exactly. WTF is he talking about? I used to think maybe his thoughts were moving faster than his fingers, and the English language just wasn’t expressive enough to communicate the profundity behind his diatribes. Now I’m becoming convinced that the explanation is much simpler. He has absolutely no clue what he’s saying or trying to say.
This guy’s pieces are like Gahan Wilson’s “Science Fiction Horror Movie Pocket Computer.” I’m convinced I could create a story generator for Gillmor that works exactly the same way. Hey Mike, if this is the kind of content you want, a simple randomized Perl script would suffice. There’s no need to pay this yahoo a dime.
Why isn’t there a Gillmor sock puppet, Loren? The dialog writes itself. You could string any series of plausible clauses together and create a Gillmor diatribe that makes just as much sense as this one (and all the others). Just have the puppet read the output of the Perl script.
The idea of an individually-recompiled Gmail is really interesting… and I believe in it not at all. Unless there’s something fundamental that Steve isn’t mentioning, users are going to get stuck with whichever subset of Google-written and Google-blessed extensions that the Gmail guys decide to include as options. Thery may be trivial, they may be obvious, but they’re not yours. They’re Google’s.
This is one of the few non-privacy drawbacks to SaaS. You can’t own the server that the app runs on, and so you *don’t* have the flexibility to “recompile” the app to include the extensions that you need. Actually, you can’t own the *servers* — I’d bet Scoble’s left nut that Gmail is fundamentally spread out over hundreds of servers and that you simply couldn’t run it on a single box or two. Other options like Yahoo’s Zimbra actually do allow for this kind of on-prem open-API flexibility, but Zimbra has sort of disappeared since getting bought out (and if Yahoo runs it as SaaS, it has the same constraints as Gmail).
The problem I have with the employee # comparison is that I don’t get what it proves. MS Office is huge and as you noted, brings in a 1/3 of MS’ revenue. They have support for a variety of versions on different OSs. Gmail’s a web app that brings in what for revenue? Zero purely on it’s own. The way it does bring in revenue is serving up text ads. So you’ve got the huge text ad group that makes it possible for Gmail to be worth the bother of having Gmail. Oh, and don’t forget all the network and server people that make Gmail possible.
But really, if we’re talking about numbers, what the heck is Google doing that they need 20,000 employees?
I agree with Brad and Judo. This was a painful read with no gain. Where’s the news???
Steve,
Sure, the bloat in any software company with more than 1k employees is horrendous. MSFT like IBM, Oracle, SAP has lots of non-developer/non-PM overhead (my guess: at least 70%) due to their (enterprise) sales model.
Another data point: Google has only 1 to 2% market share in Office. 2% of 26k (at MSFT) = 520 employees. Combining GMail and Docs, Google has significantly more in this area (I heard: around 1k, if you include international channel sales and support in this area – as the MSFT number does).
> It’s a million-plus user testing facility for free, an easy way of extending Google’s original strategy of scaling up as usage grows…
You are aware that MSFT has this in Office as well (since many years)? Office users have to opt-out of this monitoring at installation, if I remember correctly.
My point: I will not help unless Google gives up the traditional software industry perspective that one product (for an application area such as E-mail) is enough. Office apps today are like spaghetti sauce: 1 product cannot make more than 10% of the users *really* happy anymore; some want chunky, some want spicy … [I guess you know Malcom Gladwell’s talk about this (http://www.ted.com/index.php/talks/view/id/20)].
In math terms: the million users will break into (at least) 10 different clusters with very different expectations from their online E-mail app. MSFT is fighting with this since 10 years, but afraid to have “Word for Letters”, “Word for Books”, “Word for Technical Docs” (It does have “Word for Blogs” aka “Live Writer” now). Google is also just looking at 1 product. That’s the chance for Zoho and 37Signals etc., who have made the quantum leap into the post-industrial software market – understanding that “1 size fits all” was given up by the consumer product companies in the seventies already.
This is not to say that the developments of the MSFT and Google *platforms* are not interesting, but there will be space for many apps between MSFT’s “4XL” and Google’s “S” approach.
P.S. First Steve Gillmor weekend article on TechCrunch I could pretty much understand (although the “compile” reference, of course, sparked appropriate response).
Still way to long to get to the point, but perhaps your most coherent post yet Steve
I think you’re trying too hard my friend, just get to the point. Just like your recent long post, a load of gash, get on with it mate
Who is this gillmor fellow why does he write like this? it doesn’t come off as ’smart’ , it’s annoying.
Like a fighter who only throws big flailing upper cuts that rarely connect.
Steve, start with some short jabs make sure they all connect, then at the end throw that uppercut knowing that you are going to connect on it.
TC should go to ny they will teach you how to write.
For foreigners reading this article is a real nightmare and as i see for people in states too.
My advice: read after you write, then publish.
gillmor just loves gestures and affinity groups. how long till a Steve Gillmor randomized article generator appears?
6/1 - “increasingly the roundtrip between Gchat and Google Reader is producing the high value signals (gestures) that fuel affinity group formation and targeted feedback loops”
5/18 - “Live Mesh can fit into this like a glove, feeding downstream vertical versions of affinity groups to skinned Silverlight containers”
5/11- “The software switch is an affinity-based construct that manages the signal-to-noise ratio of the information flow based on the contouring signals (gestures) of the members of the group…”
“We’re at the doorway of gesture farming, where individual gesturers go beyond implicit behavior harvesting and aggregation and overtly share not just what they like but what they ignore”
4/27 - “Can Mesh support Twitter streams orchestrated by identity mapping via affinities and abstracted to devices across OS, mobile, and corporate divides via Silverlight?”
i just vomited in my mouth a little.
Steve, maybe text-based blog posts just aren’t for you.
How about video posts instead? I can’t imagine that you talk as bad as you write. Nothing more to say: @4, @9, @10 already cover the problems.
You seem to have interesting stuff to say, just trouble communicating it.
Steve,
You pretty clearly have no clue what you’re talking about here.
Gillmor should apply for a job to CIA. He will “gillmorate” their secret documents with buzzwords and 50 word sentences. Poor foreign spies.. they’ll never crack the “Gillmor code”.
pot smoker?
caldwell -
when you save your changes to Labs, the app is recompiled to your specs. Expanding that is certainly possible. No reason why virtualized servers can’t serve personalized versions. In fact they are already doing it. As to owning the hardware, why is that determinant of anything? Security indemnification insurance will solve that one.
Duncan -
It’s not that I am more coherent, but that Google’s strategy is.
Individually-recompiled Gmail: Steve, I didn’t think about it this way. Let’s see what the future will bring. I wish someone would create Gmail-like POP3 client.
Steve “Grammar” Gillmor strikes again.
Steve, please go take a remedial writing class at De Anza Community College. Just think of it as an “affinity group for enhancing the low-value suckstream” that is your writing.
Steve, mind sharing what you’re smoking with the rest of us?
I love a good high.
Agreed, this strategy analysis is easier to understand because it is closer to be realized than the other mentioned in previous posts.
Personalized VMs for each user is the next computing frontier and to tune these systems you need user feedback (hundreds of thousands) instead of just several thousand employees
Why you have so ‘Affinity’ with ‘Affinity Groups’?
Um, if the gmail team is that lean, what are the n thousand other engineers that Google has hired actually doing?
Also, after several years, I believe I read that gmail has 10M users which seems awfully low compared to legacy 1.0 services.
GMail will get interesting to me when I get a terabyte of storage with microsecond response time. Until then I am reading about of alot of window dressing. I assume the strategy is that the eye candy will keep the kids from repeatedly asking are we there yet.
One reason why running an app in your own datacenter is significant is that it’s necessary for integration with systems behind your corporate firewall.
A second reason is worry about Google’s privacy policies, which is a common bit of net paranoia that I don’t happen to share.
A third is that owning the server allows you to author and install a homebrew extension whenever you want. You seem to believe that the Gmail guys will be willing to let anyone upload and integrate arbirary code. I disagree. I think there’s almost no chance that’ll happen short-term and I’m skeptical long-term as well. Can you imagine the headaches that would have resulted if, say, Facebook had recompiled their application for each possible subset of their ten gazillion sheep-throwing apps?
If I read it right, your points were:
-I’m feeling good about myself since I was invited to the googleplex
-microsoft is bloated, google isn’t
-letting customers prioritize features with their clicks is innovative
-gmail with a lot of bells & whistles could do many things beyond email
-therefore microsoft is in trouble
By adding some supporting detail & context you might have been able to stretch that to 4-5 paragraphs but 9??? Are you a management consultant in your spare time? Would this have been easier for you in powerpoint?
To quote James Brown: “You’re like a dull old knife that just ain’t cuttin’
You’re just talkin’ a lot and sayin’ nothing”
I can’t work out if you’re a brilliant satirist, Steve, or in need of fourth grade.
this is basicly a transcript (or plug) of the Gillmore Gang podcast which is great to listen to but hard to put into words
Interesting stuff. I wonder if Google will open up Labs to non-Googlers. Like the widgets for iGoogle
Steve- Please listen to advice and work on good writing for the web. This may work on the back page of a print magazine, but your writing stinks for this format.
Trevor
Not only are you wrong about this but I”m getting tired of these types of comments and may start deleting them. Stop wasting your time admonishing me for what I am doing and skip these articles.
Another great article, Steve. You are the man! I can’t believe people don’t like your writing. It’s so cogent and concise! And your mastery of technical protocols like XMPP is revelatory. You must be a software engineer in your spare time, because you really seem to have a deep understanding of all the great technical issues of our time. Also, the way you listen to and accept your readers’ feedback with such humility is truly inspiring. Finally, I’m not ashamed to say that you are just a really good-looking man. Rawr!
P.S. Steve, is this the kind of comment you’re looking for? I’m just trying to get something — anything — past your comment-censorship.
I second that JP.
Sorry to say Steve that folks on TC aren’t shy to say when they don’t like something, and its tough to dismiss them as dumb.
Mike made a very insightful comment a while back that he originally fought against reader comments and only later realised that he should listen more.
My background is finance and over time I realised that truly insightful folks don’t use hackneyed pharses and are able to express opinions concisely.
From an outsider’s point of view, the Gmail Labs method of testing software features sounds like a great way to try out alpha ideas on the alpha geek crowd. Fast, cheap and without much risk should a less-than-perfect one escape from the zoo into the wild. Thanks for your coverage and I’m interested to see which of the features make it into the beta product. Also, I wonder when the beta will become a “real” software release, since it’s been in beta for years…
A few points;
If you actually read the comments, you can see that there are those who understand and others who don’t. Some participate in the conversation, have valid points (both aligned and contrary) and others just throw rocks at the window.
Some understand and love Coltrane while others understand and enjoy Kenny G. Believe me, I understand this concept far too well.
It doesn’t matter if it’s writers, musicians or other communicators, they each have their own voice and style and it’s the audience who decides if they will listen or not. It’s not the writer’s job to shapeshift to the liking of the reader (after all, that would make them a politician) but to continue to find their voice, honestly express their thoughts and see if it adds to the mix, so to speak.
This is a case where we’re seeing new angles and connections in so many businesses. Instead of the locked silos of paid software and content we see things given away until the addiction sets in, the dependence upon a service or connection being the reason that someone is locked in. In software, services, music, video, messaging and just about anything else, you can get it for free now. Want it reliable, quieter or bigger? Remember in every alley or playground, the first one is free.
Being able to see why businessmen sink money into these services and how one little message service or a free email account can lead to a way that can connect it all together for users, developers, advertisers and everyone else involved doesn’t always lend itself to a soundbite or a shiny piece of tinsel hanging from a branch. The rivers have found a new path to follow (still following the path of least resistance) and some people can actually use their brains to figure out where to build their house so they still have a great view but don’t live in the flood plain… or at least so they don’t have to spend all their money on flood insurance.
The thing I enjoy about this kind of communication is by asking me to think, I actually begin to have my own take on a subject, continue to refine and extrapolate the idea and by sharing it, gather more information that can help focus my own thoughts and creativity. I know plenty of people who don’t want to waste their precious time thinking about anything and you know, that’s ok. The world has room for essay, multiple choice and yes or no questions.
So spin the wheel on your iPod and listen to something that reminds you of the good old days, spin it until you hear that song that burrows into your brain like a tick for a week or two or don’t listen at all… I’m glad that not all blog posts are cut and paste PR releases with the word “PASS” or FAIL” added afterward…