I have repeatedly done this on both sides (consumer and supplier) as a database administrator, developer and architect.
As a provider, one of my great achievements (in 1996) was the creation of an installation CD for commercial insurance claims management software for the largest insurance companies (multi-million dollar product). This installation CD installed the Oracle 7.2 RDBMS engine, the FileNet optical storage system (scans paper documents and creates cataloged binary versions) and our custom claims processing application (built-in to VB 4.0), all integrated and ready to run. As part of the installation process, the user can skip the installation of Oracle software or configure it, and the user can configure / override the database configuration in all its basic details (database, schemas, table spaces, sizes, disks, etc.).
I also provided a field service for this product, which included, when necessary, switching to a client site. I tested the installation CD literally hundreds of times in any imaginary scenario that I could reproduce, and we NEVER had a field failure that required even a phone call, not to mention the trip (I traveled four times, but for pre-sale material instead of this).
Most recently (2007) I wrote a script to create an Oracle 10g database for an internal system on a megacasing. In production, the database was 8 TB in size, mainly for a single transaction table with a large amount of data. In the test, the database was about 1 TB in size for a modest server. In development, the database was about 100 MB in size to work on my laptop. EXACTLY SUCH SCRIPTS created all three environments, and I could expand them to handle the new environment / machine in about five minutes. This database included extreme performance tuning, so tuning all the relevant characteristics was absolutely critical.
Back to the insurance claims processing product - let me add that I was initially hired to convert it from a SQL Server database to an Oracle database. This transformation was defined as a business need, as most potential customers did not see the SQL Server-based product as a professional serious solution. This is not so common today, but it still applies in general: a software product has a better chance of penetrating the market if it can use several database options, as preferred by target customers (especially enterprise-class clients).
Similarly, the installation CD was also seen as an important element. However, this situation and many others have shown me that most “real” database administrators will not accept import-based database installations. As a database administrator and architect, I know that I definitely will not for the same reasons.
Simply put, installing an import-based database gives the client virtually no control over the resulting database. This is opaque to the client, and they do not ask what he did. This forces the client to make great efforts to try to realize that little control that they can do. This is known to be fragile and error prone (Oracle imports are well known for ownership and permissions, restriction problems, etc.). Weighing all of these impacts, installing an import-based database is unprofessional - at first it does not pose customer needs.
Database installation scripts provide the right kind of transparency, customizability, selective repeatability and overall client control that requires professionalism. It also encourages you to correctly understand the implications of database design decisions so that imports fail.
Best wishes.