[Gta04-owner] Business strategy

[Dr. H. Nikolaus Schaller]

Hi Martin,

Am 16.11.2011 um 16:32 schrieb Martin Christian:

> Hello Nikolaus,
> 
> by now I've collected some questions regarding the strategy of Golden
> Delicious (I'll call it GD later) and the GTA04. I think they might be
> interesting to others as well, hence I write them directly to the list:

no problem. We try to be "open" as far as we can (it has its limits if third
parties are involved). But this is a question about "us" (GD).

> 
> a) In which role do you see GD regarding the GTA04? Do you want to sell
> a component of a mobile phone only (the mainboard), a complete device
> (including case and so on) or even a complete, ready to use mobile phone?

Our dream is a truly open and well documented handheld personal mobile tool,
that is so open that you as the user have control about what is happening
inside and outside. You can decide which software is good for you. And if
you find bugs, you have the option to fix them. Not like the modern devices
for which you don't get a bug fix for Java Script in the web browser, although it
is a known security hole. And the device is just 2 years old. Or where someone
tells you that some function isn't good for you.

And, a device that is so open that you can modify it and adapt to new functions
and use cases. Or you can take it as a building block for another type of device
(e.g. a tablet etc.). Ever tried to ask one of the major device suppliers to get a
modified hardware? This is what we will provide. Individual solutions. To stimulate
your creativity towards experimenting.

Generally, our idea is to make "digital tools for independent people". Independent
in mind.

Now this was about the general dream, and bringing dreams to reality isn't simple.

So we have started to work on this dream and the GTA04 is the first (big) result.
A smaller one was the Freerunner Navigation board.

And since we had to start somewhere, and don't have really big venture capital
to burn, we started with some first step that did appear feasible two years ago.

It was clear that we can't have new cases because high quality production tools
are very expensive. And since they are expensive the are too expensive to make
beginner's mistakes. Therefore we did start what we all have in the Openmoko
community. We have a working GTA01 or a GTA02 with case. And most urgently
did need a new CPU and UMTS. This was the start of the GTA04 motherboard
upgrade idea.

For the future of a personal mobile tool, we need of course to find out how
to produce nice and robust cases (at low cost). First ideas have already been
discussed on this list. So I guess the future will see a mixture of upgrades
and new cases and complete devices.

> b) What is you time line for the GTA04 project? When do you plan to have
> a production ready version of whatever your goal is (see a)?

Since we think the hardware as defined works well enough to have them duplicated,
we have signed the contract for the production of the next batch of units. The first
units are planned to be built before Xmas and so they will be generally available
from January. A second (bigger) batch is scheduled for April. And then, we will see...

Of course, there will be modifications and upgrades in the future. If we start
dreaming, it is easily described: OMAP4/5/6/7 (or others), bigger capacitive
display, new types of sensors built in, 4G, USB3, (individual) cases. But also
available in components that you can make something out of it.

This is like software releases. When is Linux ready for production? It is in
production since 20 years, but each version has its bug fixes and introduces
new features (and bugs :).

But we have to walk one step after the other. So there is no specific timeline.
New things will happen and will be announced when ready.

> 
> c) What is your target price for the final version of your "product"
> (see a)?
> 
> Honestly, I've been shocked a little bit by your price quotes. I

> expected the board to be something around the beagle board. My idea was
> to start with your board until I have a prototype and then present it to
> my employer as a R&D project. However, 500 EUR is a little bit too much
> for a "hobby" - and that's what it is right now. Hence, my employer
> needs to be involved in a much earlier state. But that would require
> some more details about your plans with GTA04 - and following.

I know that the price tag may appear shocking, if compared to some
mass market devices. We would have been happy to have a much lower
price, but there is no option.

One approach would have been to contact one of these "cheap"
mass market producers in China and ask them to open their design.
This has been tried several times in the past years and has a lot of
big problems. Starting from long-term availability issues over language
barriers to minimum order quantities in the range of 1000-5000 units. 
And no chance to get modifications or you pay a lot of money. They
are only cheap as long as you buy what you get.

So we decided to buy components from major manufacturers who
support long-term availablility (e.g. the TI OMAP is a "Catalog Product"),
and from distributors where we can get them starting at 1 piece. SO
we can control the design and add nice things (e.g. sensors...).

As you can imagine, this strategy has its own drawbacks. The
purchasing price for small quantities is much higher than for bulk
purchases. To see an example, look here (I hope the link works):

http://de.mouser.com/Search/ProductDetail.aspx?R=DM3730CBPvirtualkey59500000virtualkey595-DM3730CBP

This is the CPU we use. You can see from 1 to 100 units, the
price is going down to 82%. And for 1k units, you can compare
the budgetary price from Texas Instruments:

http://www.ti.com/product/dm3730#pricingpackaging

Both are still quite high if compared to the sales price of a
BeagleBoard.

The same is for the UMTS modules, the sensors, GPS, WLAN
and all the other smaller and larger chips we have in the design.

Another factor is that many production steps have one-time cost
(setup cost, shipment etc) that becomes lower per unit if we
produce in larger batches.

This all makes purchasing the components and producing a
GTA04 expensive in the current phase. And, since we don't
have a money-source to subsidize the boards, we have to
ask if you are willing to pay the (higher) price to get something
unique, individualistic and be part of a community effort. Not
controlled by "big money".

There is one way out of this trap of small quantities. It is to
increase quantities. Therefore, we offer the group-buy approach.
By this we increase our buying power for the future and can
negotiate better prices for components and production. If we
can also show that there is future demand, we may even get better
prices than those published like the links shown above. This helps
to reduce the price tag for the future. Or include new really cool
features.

So this is the market- and economical mechanics behind the
GTA04 project.

How did Openmoko make the GTA02 price? I don't know internals,
but I think they did never reach their break-even. Mainly they
were spending the Venture Capital.

The Beagleboard is as far as I know running in 3k batches - and
has only 30% of the components value as we have it. They have
no UMTS, WLAN, GPS, Sensors on board and just a 6 layers PCB.
All this makes it less expensive to produce.

And do you know the OpenPandora? It is sort of a sister project of
our GTA04. An open gaming console (sort of a PDA with game
controller). They share the OMAP3 processor and many other
concepts and ideas.

They started collecting pre-order money approx. 3 years ago.
Last year they started to ship the first units. In the meantime it
turned out that their production cost is higher than the pre-order
money did sum up. So they had to collect new money (loans) to
finalize the devices that already had been paid. So they are in
the situation close to many organizations around the world, who
are heavily indebted and need more money to pay back old
debt. The debtees can now tell what one should do. Including
closing the operation and auctioning the remains.

This is against our idea of "free individualists" and therefore
something we must avoid. We want to stay here and provide
many generations of Openmokos in the future.

Therefore we can only hope that there are enough enthusiasts
around the world who share our ideas and dreams and find
ways to pay for the (more expensive) components and
production. And development of future variants.

Coming back to your question: there is no real "target price".
It always depends on what is inside and how good we can
manage the production cost.

And, if you think your idea is good enough so that your employer
could be interested, I think he should be able to afford to sponsor
some prototype for you. If you need some additional piece of
information to convince him, please let me know.


Hm.

Now, this has become a long "manifest". Much longer than I
had originally thought :)

Finally, I have good news: we have apparently found a way
to get a batch of Si4721 chips in time.

Nikolaus