Business model for Open Source - advice wanted

[Frank Millman]

Hi all

I would like some advice, and I hope that the good people on c.l.p
will give me the benefit of their experience.

I am busy developing an accounting package, using Python and wxPython,
together with PostgreSQL on a Unix platform or SQL Server on a Windows
platform. I am a qualified accountant and have been developing
accounting software, mostly customised, for over 20 years. I have now
got myself into a position financially where I can take a few months
off and write "the perfect system". No, not really, but I have built
up a lot of experience over the years, I have a lot of ideas about
what I would like to see in a system, and I have decided to give
myself the opportunity to see how good a system I can actually write.

At some point quite soon I will have to start generating some revenue
from my efforts. I have a number of people interested in what I am
doing, and I was planning to start selling it as YAAS (yet another
accounting system) and see how many customers I could get.

As an alternative, I have been toying with the idea of releasing it
under an Open Source license. Part of my thinking is to put something
back into the Open Source community - I love working with Python,
wxPython, PostgreSQL, etc, and it would be nice to contribute
something. The other part of my thinking is that it should be
possible, with a sound business model, to make a viable income out of
supporting an Open Source product. However, as I have no experience of
this, I was wondering how other people survive financially in the Open
Source world. I have no "big daddy" that will pay me to pursue my
dreams, so whatever I do has to bring in a reasonable income.

Here are some of my ideas - any comments will by much appreciated.

The essence of a business model must be that the source code may be
free, but you have to pay for support. I believe that accounting
systems more than most systems need support, and that a sensible
company looking for accounting software would not consider anything
that was not supported. Therefore it is important to decide what kind
of support should be offered, and how to persuade customers to sign up
for it.

The software must be downloadable and runnable by anyone, and it must
be fully functional, at no charge. I realise that there is an option
to restrict functionality in the free version and make it available
only to licensed users, but then you are into security codes, source
code protection, etc, which would at least partially defeat the
object, so I am not thinking along those lines. There will be full
online help (it will probably start off sketchy, but I will improve it
over time), and maybe there should be an FAQ available. That, I think,
is all that should be free.

I would like to open a free discussion forum (probably a mailing list)
to discuss the software from a technical perspective. I am an
experienced programmer, but I am nowhere near being a Pythonista, so I
am sure there will be many areas ripe for improvement. I would like to
remain open to as much of this kind of feedback as possible.

However, any support of the software from a business or accounting
perspective will be available to registered users only. For a monthly
support fee (I am thinking of a range of $50-100 per month - let me
know if it is too high or too low) users will be able to request
support via email, which will be responded to promptly. I have a small
staff at present, and if this idea is successful I will plough money
back into expanding the support infrastructure.

To retain registered users once their software has settled down, I
have in mind a website restricted to registered users which will
provide sufficiently interesting content to persuade them to keep up
their registration. The main content I can think of is enhancement
requests - it is the nature of accounting software that there is a
continuous demand for enhancements, and I intend to keep on enhancing
and improving the product. I envisage a forum where users can forward
ideas, and others can respond/comment/criticise until there is some
consensus. A roadmap of agreed enhancements can be posted on the free
website, which will hopefully attract more users into signing up so
that they can participate.

Here are a couple of pitfalls that I can see, and my possible
solutions to them. Firstly, a user will be able to terminate his
support contract at any time, and the software will continue to
function. Therefore some may be tempted to stop paying until they have
a problem, then pay for a month, get an answer, and stop again. My
thinking is that 3 months must be paid up front to activate a support
contract. Secondly, a consultant may use my software to build up a
client base, pay a single support fee, and use that to support all his
clients. I don't think I can (or even want to) prevent that, but I can
restrict it by limiting each contract to a single email address for
replies to support queries.

These are my thoughts so far. Any comments - suggestions, warnings,
personal experiences - will be much appreciated.

Many thanks

Frank Millman

 

[Terence Way]

Two companies that make money spring to mind:
1. sleepycat software (they support Berkeley DB); and
2. the MySQL folks (they support, um, MySQL)

They have different licenses, MySQL being dual licensed:
GPL and proprietary, while sleepycat is custom and
OSI-compatible.

My understanding is that they both make good money.

--

I think you're right on target that accounting systems
need customization... and you'll make money customizing
it. Your only problem would then be marketing your
product/service. Business people seldom read freshmeat,
I fear.

--

Don't worry about people reneging on support contracts.
Companies get large enough, investors/directors will
blow a fuse if they find out that the accounting system
is unsupported: if you target the right companies the
support checks will flow uninterrupted. If you're still
worried about 'em, offer a pay-per-incident, like MS.

--

If your product is good enough, *I* may be one of those
consultants that uses your software, gets the support
license, and uses it to support all my clients. This is
a *good* thing. For one thing, any enhancements I add
get folded into your product, so you and all your
customers benefit. For another, I won't be claiming I
wrote the thing: support by the original author of a
product is still important.

--

Consider the target company size. Given the $50-$100
support range, I'm thinking small companies < 10 people.
Your support fee is way too low for larger companies.

That is all. Hope this helps.

 

[Shane Hathaway]

I am busy developing an accounting package, using Python and wxPython,
together with PostgreSQL on a Unix platform or SQL Server on a Windows
platform. I am a qualified accountant and have been developing
accounting software, mostly customised, for over 20 years. [...]
I have been toying with the idea of releasing it
under an Open Source license.

I agree with Evan: open source has turned out to be a very effective
model for Zope Corp. Once your product is popular (partly because it's
Free), you can sell your expertise in a lot of ways:

- You can sell to different verticals using your open source product as
a platform. Every business has different accounting needs, but
businesses of a certain category tend to be quite similar. You could
make packages for real estate, law, medical practices, retail shops,
etc. The vertical customizations should be simple enough that it
doesn't cost you much to create them, but should have enough "meat" that
it's more cost effective for a business to buy the customized product
than to re-create it on their own.

- Sell training courses. You might have two levels of training: one for
end users and one for larger companies that want to customize your
product. Training courses can be a lot of fun, and people are often
willing to pay quite a bit for a week's worth of expert knowledge.

- When working with small shops with no IT staff, you could consider
becoming an application service provider, keeping the software running
smoothly and the data backed up without any intervention by your
customer. The recurring income could be enough to sustain the business
for a long time.

The important thing is to build strong relationships of trust with both
customers and potential customers, invite people to use your products
and services, and follow up on commitments. That's a key to success in
any business. Free, open source software gives you an advantage in the
process of building relationships. If you cultivate those relationships
through mailing lists and personal contacts, business opportunities will
surface in the most unexpected places.

Well, that's how it works for Zope, anyway! ;-) I think many open
source businesses struggle primarily because they spend too much time on
the technology and not enough time connecting with people. I am
fortunate to work for a company that has its priorities straight.

Shane

 

[Tim Churches]

I am busy developing an accounting package, using Python and wxPython,
together with PostgreSQL on a Unix platform or SQL Server on a Windows
platform.

Excellent choice of tools - identical to those used by the GnuMed
project (except for MS SQL Server - but note that PostgreSQL 7.4,
expected to be production quality soon, will run natively under
Windows). The common infrastructure raises the possibility of
integrating your package with GnuMed at some time in the future - at the
very least, GPs (family physicians) and others who adopt GnuMed will be
able to install your package without any extra work at all - and every
doctor needs an accounting package...

--

Tim C

PGP/GnuPG Key 1024D/EAF993D0 available from keyservers everywhere
or at http://members.optushome.com.au/tchur/pubkey.asc
Key fingerprint = 8C22 BF76 33BA B3B5 1D5B EB37 7891 46A9 EAF9 93D0



-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQA/Dd28eJFGqer5k9ARAo5CAKDE505aaSgxOwZ3nceIBYhZ8VS1pwCfXXZL
v82w/QAB+EugBO5PwcCpb2A=
=addd
-----END PGP SIGNATURE-----

 

[=?ISO-8859-1?Q?Gerhard_H=E4ring?=]

I am busy developing an accounting package, using Python and wxPython,
together with PostgreSQL on a Unix platform or SQL Server on a Windows
platform.

Excellent choice of tools - identical to those used by the GnuMed
project (except for MS SQL Server - but note that PostgreSQL 7.4,
expected to be production quality soon, will run natively under
Windows). [...]

Unlikely. This is most likely to get postponed to later (version 7.5).
PITR also didn't made it for 7.4.

Apart from this nitpick I agree that the choice of tools is excellent.
They're my preferred ones as well.

FWIW building a Pg/Cygwin installer shouldn't be too much effort,
either. Basically just provide a trimmed down Cygwin.

Native win32 PostgreSQL versions are also available from third parties ...

-- Gerhard

 

[indru]

Your question is very valid and to make money out of your exposed code
is very difficult if you consider that the product you are writing is a
general purpose one, say a loan calculator . But your area is Accounting
the area which requires a lot of customization. So writing a common
engine from where you can build customized solutions for your clients is
one way to make money. Here the question of support doesnt rise up since
your are the person who have written the code and you will be the best
person who knows about it and you can easily enter a contarct with the
people for support.
Wishing you best of luck
Indru

 

[Frank Millman]

Thanks to all for the replies. Much good advice and food for thought.
It seems that this is an idea worth pursuing. I will now keep a low
profile for a few more months to get the software to a stable state,
and then hopefully you will hear from me again.

Frank Millman

 

[Peter Hansen]

Thanks to all for the replies. Much good advice and food for thought.
It seems that this is an idea worth pursuing.
I will now keep a low profile for a few more months to get the software

^^^^^^^^^^^^^^^^^^^^^^^^^^^

Ouch! You clearly haven't adopted "agile" methods such as Extreme
Programming yet. "Keeping a low profile until the software is stable"
is a good, nutshell description of why most software projects FAIL.

Please, for your own good, at least investigate these things and
consider how you might apply them or adapt them to your own needs.

See http://www.agilealliance.org/ and http://www.extremeprogramming.org/
for a start.

to a stable state, and then hopefully you will hear from me again.

I'll be more confident that we will if you actually don't keep a low
profile. ;-)

-Peter

 

[David M. Cook]

I would like some advice, and I hope that the good people on c.l.p
will give me the benefit of their experience.

I was going to point you to this talk on "partnership" billing:

http://www.sdphp.net/node.php?id=10

but I'm having problems unzipping it under Linux, myself; maybe it's OK
under windows.

Dave Cook

 

[Keith Bostic]

Two companies that make money spring to mind:
1. sleepycat software (they support Berkeley DB); and
2. the MySQL folks (they support, um, MySQL)

They have different licenses, MySQL being dual licensed:
GPL and proprietary, while sleepycat is custom and
OSI-compatible.

A minor clarification -- Sleepycat Software also does dual
licensing. The public Berkeley DB releases have a BSD-style,
OSI-compatible license, but we also release under a pretty
standard, proprietary license.

Regards,
--keith

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Keith Bostic (e-mail address removed)
Sleepycat Software Inc. keithbosticim (ymsgid)
118 Tower Rd. +1-781-259-3139
Lincoln, MA 01773 http://www.sleepycat.com

 

[Frank Millman]

I was going to point you to this talk on "partnership" billing:

http://www.sdphp.net/node.php?id=10

but I'm having problems unzipping it under Linux, myself; maybe it's OK
under windows.

Dave Cook

Nope - it will not unzip under Windows either - it seems that it is
not a valid zip file. Thanks anyway.

Frank Millman