Subject: Re: Motivating support contracts
From: Keith Bostic <>
Date: Thu, 3 Sep 1998 16:27:14 -0400 (EDT)

> Many customers don't want to buy support.
> Customization services and installation support, yes; a straight
> service contract, no.

I think this is a good point in general.  However, it may be
less applicable to my particular situation:  As a single data
point, I find that few of our customers want customization
service, less than 10% at this point.  Installation support is
a great idea for applications.  I'm selling a programmatic
library -- everyone buying it is a programmer, and fully capable
of doing their own installation.

I think that customization services will be less useful over
the long term, there are only so many features that most
applications need (note that I said *most* applications... :-)

> The product (I assume this is Berkeley db) should probably have a
> CD/ROM and prebuilt binaries and so on.

It's not an application, it's a library.  Anyone using it is going
to do their own builds regardless.  The only thing that a CDROM can
add is documentation (e.g., PostScript of the docs, or perhaps docs
that aren't available on the net).  At this point we don't offer a
CDROM, incidentally, for this reason.

> With the support contract,
> send updated software at least once a year, and send out a newsletter
> discussing new uses of the database, new customers if they permit,
> problems solved, and tricks of usage.  Run a mailing list for
> customers only.  Make your customers feel that they are part of the
> community of Berkeley db users.  Then, when the year is nearly up, hit
> them up for a fee to stay in the community.

This sounds good.  I point out that this is probably a low-price
model, i.e., the yearly fee is under 1K.

Thanks for your other comments as well, I'll definitely consider
them further.