NNSquad - Network Neutrality Squad
[ NNSquad ] The Google Buzz Launch -- and the Limits of Downing Dogfood
The Google Buzz Launch -- and the Limits of Downing Dogfood
http://lauren.vortex.com/archive/000683.html
Greetings. There's an old Hollywood adage suggesting that most of the
time, "any publicity is good publicity." When it comes to the launch
of Google Buzz, there's definitely some truth to that saying -- the
widely discussed privacy issues associated with the launch have
yielded the product a significant global awareness far outside the
world of current Gmail users. And reports are that usage of Buzz is
(sorry, I can't resist) buzzin' along at a very significant clip.
Still, the very public privacy controversies regarding Buzz over the
week since its debut (hard to believe it's only been a week) are both
fascinating and instructive.
In "'Google Buzz' -- and the Risks of 'Automatic Friends'" I noted my
own concerns about specific features of the original Buzz start-up
experience defaults, and expressed the hope that Google would
reconsider those defaults
( http://lauren.vortex.com/archive/000680.html )
I wrote that piece on launch day after my own initial experiments with
the product. Between then and now Google has announced two
sets of significant changes to Buzz that do a good job of
addressing the issues that I noted ( http://bit.ly/c3my85 [ Gmail Blog ] ).
But as seems to be the case with anything involving Google these days,
one comments publicly at one's own risk. After I was widely
quoted as praising the first round of Google's Buzz changes and
noting that, "The thing hasn't been out a week, it's going to take
some period to hash out." -- the volume of vitriolic "hate" e-mail
that I received on the topic was both large and in some cases rather
bizarre ( http://bit.ly/coq36c [AP] ).
These missives fell into several categories. The "Google Conspiracy"
set are always fun reading. In the case of Buzz, the theory seems to
be that the initial default settings were part of a "secret plot" by
Google to abuse users' e-mail contact lists and associated data. A
glaring problem with that supposition is that there was nothing at all
secret about the default followers policy that Buzz established.
While many users may have not initially understood the full
implications of the defaults, or alternatively (as in my case) may
have felt that the defaults had some inherently risky characteristics
or were problematic in other ways, the settings certainly weren't
secret. It was clear from the onset what the model was for the
"initial populating" of Buzz followers.
Another group of these correspondents complained that I shouldn't have
praised Google for the changes they were making to Buzz, even though
the changes were pretty much exactly what I had suggested would be
useful. The implication of such "damned if you do and damned if you
don't" logic is that unless a product is 100% correct right out of the
starting gate, it deserves to be condemned to an inner circle of hell
forever.
Frankly, I look at this from pretty much the opposite point of view.
If you always play it totally safe in product design, for fear of
making any mistakes, true innovation is slowed or in many cases even
impossible. That Google erred in their initial design of the Buzz
defaults is significant, but far more important to me is the extreme
rapidity with which they publicly acknowledged these problems and have
moved to fix them -- and word is that even more changes addressing
various Buzz issues will be forthcoming very shortly.
But caustic communications within my inbox aside, one might still
reasonably ask how Google apparently so significantly misread the
likely reaction to the original Buzz defaults in the first place.
I don't have any inside information on this score, so like anyone else
on the outside of Google I can only speculate. But it seems certain
that Buzz was extensively tested within Google itself for a
significant period before it was released to the public a week ago.
This sort of very wide (but still internal) testing of a product
through actual use is commonly called "dogfooding" -- that is, "eating
one's own dog food."
It's an excellent way to discover and hammer out technical
deficiencies in a product, but can have significant limitations if the
reaction of users within the "dogfooding" community leads to a less
than fully accurate extrapolation to how the user population outside
the confines of the firm itself will react.
The Google corporate culture is remarkably open on the inside, with a
tremendous amount of information sharing among individuals and
projects. It's easy to imagine how many enthusiastic,
pre-public-launch Google users of Buzz might have inadvertently had
something of a blind spot to the more "compartmented" nature of e-mail
and "social messaging" communications that is much more the norm in
the "outside world."
This highlights a key limitation of dogfooding, or even of testing
involving non-corporate early adopters. If sample sets are not
sufficiently large and especially broad in terms of different sorts of
users in different kinds of situations, it's possible for internal
enthusiasm to lead any engineering team to assumptions that may not
necessarily be optimal for a released product facing a global user
base.
Whether my speculation above does or doesn't resemble what actually
occurred internally at Google related to Buzz, it is demonstrably true
that to the extent we can formulate a product's design to anticipate
and encompass the widest practicable range of user concerns and
sensibilities, the lower the probability of launch missteps.
But even when such missteps do occur, the ability to react quickly,
openly, decisively, and effectively to address resulting concerns is
paramount, and Google's responses to the Buzz privacy controversies
have been an excellent example of doing so in very much an exemplary
fashion.
--Lauren--
Lauren Weinstein
lauren@vortex.com
Tel: +1 (818) 225-2800
http://www.pfir.org/lauren
Co-Founder, PFIR
- People For Internet Responsibility - http://www.pfir.org
Co-Founder, NNSquad
- Network Neutrality Squad - http://www.nnsquad.org
Founder, GCTIP - Global Coalition
for Transparent Internet Performance - http://www.gctip.org
Founder, PRIVACY Forum - http://www.vortex.com
Member, ACM Committee on Computers and Public Policy
Lauren's Blog: http://lauren.vortex.com
Twitter: https://twitter.com/laurenweinstein