[Marketing] Competition for the XO

[David Farning]

On Sat, Oct 10, 2009 at 5:18 AM, Sean DALY <sdaly.be at gmail.com> wrote:
> This list from early 2008 is way out of date... for OLPC, they are off
> by a million machines or so :D
>
> Nevertheless, it is useful, thanks Tim
>
> I agree that any OEM deal will raise Sugar's profile immensely.
> However, seeing OEMs is time and travel intensive, difficult with our
> limited resources. I should add that as we are dependent on an
> underlying distro, all of which have encountered difficulties inking
> OEM deals.
>
> It's my belief that the most promising OEM deals with GNU/Linux
> distros and "Sugar on top" will happen over non-x86 processor
> architectures such as ARM - in other words, on machines which can't
> run Windows. This is how EeePC started the retail netbook craze;
> Windows couldn't run on the lowest-end netbooks, and even where they
> could, they didn't wish to be price-squeezed at the bottom of the
> market. Their response has been to pressure OEMs to beef up netbook
> specs and drop GNU/Linux distros, to rename netbooks as "ULCPCs" (a
> total failure), and to insure that Windows 7 will be technically able
> to run on netbooks. They are claiming this for the launch later this
> month, but what's missing is how they intend to ease the upgrade path
> from Windows XP (wipe and install necessary) on machines with no
> optical drives. Is online update of the whole OS an option? Or
> requiring price-conscious netbook owners to buy an external optical
> drive just for the upgrade? I myself believe the rumors that there
> will be a Windows USB stick SKU, possibly with a tool for saving and
> reinstalling WinXP data. None of the major tech journalists following
> Microsoft are talking about this problem which leads me to believe
> they have been briefed on Microsoft's plans under embargo.
>
> Our approach up to now has been to establish our identity parallel to
> but separate from OLPC, in order to minimize the impact of their bad
> press, while at the same time supporting OLPC as our primary installed
> base. OEM deciders need to know about us - Mike Lee mentioned how at
> NECC in Washington DC a few months back, he showed Sugar on a Stick
> running on an EeePC to a surprised Asus executive - so we work on
> raising our public profile and building a meme that "Sugar runs on
> everything". Press launches raise buzz and the news reaches NGOs too.
> We are also planning promotional work to NGOs with the FSF.
>
> Concretely, what this means is I try to add the names of OEM execs to
> our press mailing list, and we are making efforts to be present at
> education tech shows and conferences where OEMs can see us.
>
> How do you think we could better reach OEMs? There is of course
> working the phones, which David and Walter do a lot of.
>
> Finally, there is the community aspect. OEMs will take risks choosing
> any GNU/Linux distro and for a Sugar-branded machine in particular.
> They want to know that our community is vibrant, growing, active in
> the distro ecosystem, and well-governed.

Sean is spot on with this assessment.  OEMs are not particularly
concerned with the sugar product.  They are interested in the Sugar
Labs project.  If they just want Sugar, they can fork it.  The value
for an OEM comes from being able to build a competitive product on top
of the 'output' from Sugar Labs.

As an example, 10 years ago there were dozens of kernels and operating
systems for embedded devices. Now we are down to a handful.  The
quality of the remaining kernels and operating systems is good enough
that it is seldom worthwhile for a embedded device manufacture to roll
their own.

Embedded device manufactures collaboratively develop Linux while
competing by selling devices which run Linux.

 The relationship between Sugar Labs and OLPC is a case study in the
nature of a relationship between sugar Labs and any other OEM.  OLPC
(or other OEMs) must:
1. Build the hardware.
2. Pick sugar over (or in addition to) other learning platforms.
3. Pick a Linux distribution.
4. Combine Sugar with a standard Linux distribution to run on their hardware.
5. Optimize the standard Sugar/distribution for their hardware.
6. Sell the hardware/distribution/sugar 'product.'
7. Support the hardware/distribution/sugar 'product.'
8. Generate enough Return On Investment to make it cost effective to
repeat steps 1-7.

An interesting note is that from an operations point of view, SoaS is
very similar to a 'normal' OEM relationship.  But in this case the
hardware is a USB memory stick which runs on other systems.

david