[Sugar-devel] Debrief of Sugar on a Stick v1 Strawberry launch for all teams

[Sean DALY]

We have had a successful media launch of the Strawberry release of
SoaS; coverage is ongoing a week after the launch.

I feel very strongly that a successful launch like this can only work
if everyone is on board together, from developers to marketers, from
packagers to designers, so I have preferred starting this integrated
thread rather than continuing David's separate threads; I also feel
that the longer-term SoaS-distro issue should be discussed separately.
Although we did manage to avoid confusion from the last-minute
timetable change through some hard work, we may not be so lucky next
time; communication between teams is vital, especially as we grow.
Routine work should of course stay compartmentalized, but I am
convinced the key to a launch's success (aside from great software :-)
is that we all pull together and make an extra effort at launch time,
pulling back after launch.

Coverage began with an article in MIT Technology Review a few hours
before the press release went out; we were Slashdotted several hours
later. This was followed by a BBC News report the day of the release,
and we have been picked up around the world every day since by tech
media, bloggers, and even some Spanish language print newspapers.

I want to share some observations, and mention several techniques we
used this time which multiplied coverage, as well as some missed
opportunities. Comments are encouraged pleased.


* Press release editing.
We got the PR done 30 minutes before the Friday evening deadline and I
thank Walter, Fred, David, and Caroline for their very helpful
co-editing with me directly on the Google Docs document and IRC
discussion. I had been concerned about an Activities positioning issue
and we made a good choice through consensus. We were able to trim 150
words in the final minutes yet the final release had enough
information to interest editors worldwide.


* Prelaunch journalist briefings.
Some journalists were briefed with the releases beforehand, under
embargo. This common practice gives them time to decide if they want
to work up a story or not and provides an opportunity for direct
discussion with us for background and quotes. It also provides
precious lead time for us to provide visuals (journalists won't waste
time fishing, and without visuals will just google and snatch the
first thing they find, including bad logos and dated screenshots).


* The last-minute timetable change.
We successfully spun the move of v1 from the "Q3 in the fall" to June
as "part of the plan" and diverted some attention from the numbering
with the Strawberrry code name which was universally liked. Only one
news site noticed we had changed our story, and their coverage arrived
late; journalists who have been following us kindly didn't bring it
up. That said I can't stress enough that our very wide coverage was a
direct result of our simplification of the numbering system to
"beta-1" and "v1"; most news sites judged this release as our first
major milestone since the creation of Sugar Labs. I agree with David
and Caroline that our next major media push should stress content over
technical info to generate teacher interest. As part of avoiding
last-minute crises in the future, to avoid surprises I sent the press
release to all the lists before it went out on the wires. The
marketing team work is of course available to all.


* Launch datelined LinuxTag Berlin.
Do a Google News search in English on "LinuxTag"... you will notice
that our launch is the only widely aggregated news. In other words, no
one else did a media launch in English from LinuxTag with coverage as
wide as ours. This was effective because it showed that we are mobile,
e.g. with roots in but not anchored to the Cambridge base. We also
managed to create a "halo effect" for GNU/Linux; several articles
including the BBC piece mentioned distributions. Although we make the
choice (correct in my view) to not place priority on our FOSS nature,
preferring to stress education, we raised the profile of Fedora with
half a dozen articles which reported that information. The coverage
gave us a boost at the booth, and the design sprint starring Gary and
Mike helped us to have focused marketing materials for both LinuxTag
and NECC.


* Targeted mailing list bigger with each launch.
We have a simple rule: anyone who writes about Sugar and/or OLPC gets
onto the list. We are at 500 names and to date have received only two
requests for removal (one from a small news outlet where other
journalists remain on-list, one stating the address used is not the
best one for PR submissions). We have very complete coverage of tech
journalists, but spotty coverage of education news outlets including
blogs, so I am turning my attention to that. Some journalists and
bloggers who treated us unfairly in the past are not on the list or
are sent the PR afterwards; of course, writers of tough but fair
coverage get the PR like everybody else. Note that as most journalists
receive dozens of such press releases every day, occasionally a
journalist will learn of us through another news article, then search
her mail to find our unread release before trying to find our press
page.


* Multilanguage PR in targeted press mailings.
The press release went out in 5 languages: English, Spanish, French,
Italian, and German. Bravo to the translators who provided me the text
before the deadline! This was extremely effective in obtaining wide
coverage, many sites digested the press release in the local language.
Coverage in Brazil was late, I am convinced we would have had more
coverage earlier if we had had a Portugese translation available (we
still don't). Unfortunately, Christian was unable to post the
translations to the press page, see next point.


* Bottleneck problem for SL homepage and press page.
Christian is very booked up with the little one and his work and has
not been able to post the press release translations or a homepage
callout for SoaS, as well as other agreed-upon requests such as a
homepage link to the gallery page instead of to the QuickTime movie,
the sitewide navbar, and links to visuals on the press page. We would
have had even wider coverage with translated press releases up, and we
need to examine how to avoid this bottleneck in the future. I myself
did not find the time to prepare PDFs which include our logo and the
PR visual and ideally our back catalogue of PR should be translated
too, I may put these in the "one-hour contribution" section.


* Availability of visuals.
The fabulous SoaS beauty shot by my friend the talented Philippe
Cantiniau and topped by Gary and Christian's last-minute work on the
Strawberry "sugar on a stick" logo text was very widely carried by
news outlets including the BBC because of its shininess, and
apparently the confusion we feared between SoaS the ISO and the
branded USB stick was not widespread - every article I saw mentioned
it was freely available for download. It's sometimes debated that
nonprofit libre software should not glorify "products", especially at
the expense of software, but any observer should see that the OLPC
XO-1 beauty shots had an enormous impact in raising awareness. Also, a
great many websites carried screenshots of Buddy View with
collaboration; the large colorful icons in that screenshot kept their
visual code when thumbnailed, better than the Neighborhood View. I
opened a Flickr stream for my Sugar-themed photos just before the
launch and the BBC photo of the XO and Intel Classmate running Sugar
came from there; I will add some screenshots too. I would like the
press page to link to that photostream and Mike Lee's as well,
journalists really like finding CC content they can use.


* PR image hosted on SL site and not distributed with release.
We fell down on getting our visuals well referenced. Fortunately many
(but by no means most) journalists and bloggers followed the hyperlink
in the press release, but Google Image searches are disappointing; a
major Chinese news site published an old draft logo. The clear
solution is 1) distribute an image itself as well as the link with the
eReleases/PR Newswire release, as we did with the March release logo
and screenshot, 2) as mentioned above make visuals easily findable on
the press page and in the website gallery.


* USB boot issue on older computers.
Slashdot readers who did not RTFA complained of SoaS's uselessness
with old PCs unable to boot from USB (due to BIOS limitation, I have a
PC like that myself). Fortunately, clueful commenters informed readers
about the boot helper CD. We reacted quickly by updating the
Strawberry page and other pages of the wiki, with callouts about the
boot helper CD; we were unable to change Christian's homepage as
described above. In the future, we will want to have a Boot Helper CD
visual available. The BBC article was updated three times during the
day of the launch, and one of the running updates added info about the
boot helper CD.


* Wiki pages not quite ready at launch.
The Strawberry page was posted as the release went out, but was poorly
referenced from the wiki homepage, the Sugar on a Stick page, and the
press release itself. Updates were quickly added, but it will be much
better in future to get this done at least 2-3 days before media
launch; remember, some briefed journalists have the PR under embargo
and will evaluate us based on the status of what is online.


* Video of SoaS booting.
One site did a home video of loading SoaS to a stick and booting it;
this was very quickly widely referenced, which shows that either we
don't have a good vid like that ourselves, or a publication has more
credibility on the subject than we do, or both :-)


* Mac compatibility.
We were late in getting the updated VDI and instructions online. Mac
users are a vocal lot :-) and are used to things just working so any
efforts we can make to ease the Mac experience will help us with
coverage. Many journalists write on Macs and some key articles in the
past had screenshots from a Mac VM. Perhaps one of us (Dave B, Gary,
Caroline, myself) could be a Mac emissary to the development team?


* School server.
We mentioned the school server in the PR, which has all the potential
to be a solid solution with SoaS but requires work. Tech journalists
didn't follow up on that in depth, but we need to work out how that
support will be part of the SoaS ecosystem.


* Perception of Sugar Labs as a company, or as part of OLPC.
We slip the word "nonprofit" into our PR often, but we still see
perceptions that SL is a company or is part of OLPC. Our homepage is
clear on the latter, but unclear on the former, so we need to reflect
that.


* No OLPC briefing beforehand.
The authors of the key articles in this launch sought comment from
OLPC, who provided an odd statement from Professor Negroponte, odd
enough that it was heavily edited or truncated by the journalists. The
BBC published more of the quote than anyone, obliging us to respond
that NN hadn't understood the tech. The statement did not make sense
and was out of phase with what is currently under development, namely
a Gnome/Sugar choice of default desktop similar to Apple's OS9/OSX
solution of a few years ago. We are in contact with OLPC's PR people
so as to smooth bumps prior to launches; perhaps Adam Holt can help us
in this regard as well.


* Infrastructure performance issues.
The sites held up well due to infrastructure steps taken by the David,
Bernie, Marten, and the Systems Team (torrent, mirroring, monitoring);
hopefully we will not be swamped by the Chinese coverage which is
growing quickly.


* Coverage reporting.
With each launch, I enumerate links to coverage; aside from
celebrating :-), this serves a dual purpose: 1) enable local language
speakers to evaluate quality of coverage, most useful when there is
negative coverage (happily not the case this time), 2) simplify the
later work of adding contacts to our targeted mailing list. However,
the (unexpected!) volume of coverage was such that one member of the
community complained. The suggestion was made to create a wiki page,
but that doesn't help as far as identifying good, mediocre, bad, and
indifferent (i.e. reprint of PR) coverage - it would never be to our
advantage to publish ratings of journalists' work (all PR firms and
corporate communication departments maintain journalist profiles, but
in the strictest secrecy for obvious reasons). I liked the suggestion
to regroup coverage in batches, and from five days after the release
(when the risk of negative coverage was greatly reduced) I sent the
batches to the marketing list only and not IAEP. I was very pleased to
receive links from Carlo and Raul to coverage I would have missed in
their countries; it's best-practice since I can then add the
journalists to the list. Everyone's help in spotting coverage is
appreciated during a launch and spotting negative coverage early is
critical.


* No coverage on some key tech sites.
Some tech publications did not cover us this time, or came late to the
party despite having received the targeted mailing. I have identified
the sites in question and will add journalists to the list. In the
case of OLPCNews, my impression after an exchange with Wayan (he's
just published by "focus group" post) is that his daily coverage and
proximity to the OLPC/Sugar Labs projects hid the significance of the
release to the wider tech world.


* Press-contactable.
Our press page mail alias and softphone phone number, although
sparsely used, enabled the BBC and a few others to reach us
immediately.


* Site statistics.
I've chosen to not prioritize this for now (aside from worrying if we
will have a China meltdown); Bernie provided me with a graph that said
it all. In future I'd like to check the site stats during the launch
period and report metrics to the marketing team, in particular the
number of SoaS ISO downloads. Dailymotion page views seem small,
indicating we need more content in there.


* Contributor recruitment.
The USA volunteers site, and our wide coverage, worked together to
persuade several people to propose contributing to Sugar Labs. This is
an excellent development and shows the positive impact wide coverage
can have in other areas such as funding and partnerships.


Please reply with comments and suggestions, especially ideas about how
to obtain better coverage in education publications, blogs, social
media, and newspapers.


thanks


Sean
Sugar Labs Marketing Coordinator