Quality of Product
Blue Martini Expertise
A selling point for Blue Martini was the fact that it was a mature product. Unfortunately this doesn't
seem to work in the benefit of the product. There are still too many defects to call the product mature
and stable. Furthermore the product requires older versions of third party software in order to work!
Remote Desktop - Inconsistencies between what is available in the ED client versus RD client
Security- Need for more permissions control such as being able to prevent users from deleting a
product folder (and all it's contents). The current privilege is Create/Delete, but this needs to be
separated into two separate permissions for greater control. A finer grained security model is
Administration frustrations – e.g. the existence of too many patches and all the consequences
thereof. Despite the claim that BlueMartini is a mature product one gets the feeling that parts of the
application have been insufficiently tested and/or rarely used.
GUI of ED should be improved with expected bars/graphs/buttons. Datamining Status - Either a \%
complete\ or better still a \x minutes remaining\ indicator when executing a \Show Table\ for
Delay in upgrade path to newer JVM's and/or systems. We are still *stuck* on 1.3.x and are unable
to make use of new features (such as Assert) and improved performance of JDK1.4x. There is no
Linux version available which would reduce costs for many customers
The quality of the product needs to improve and will only do so when more clients are taken on-
board and/or test suites are set up to test specific parts of the application which rarely gets touched
by QA. By encouraging customers to download code to test and opening up other parts of the code
for view the code will heal itself, ala Open Source model. Porting BM to Linux and up-to-date JVM
can't be that difficult and could be written & tested by customers together with BM.
Despite the *maturity* of the product BM expertise seems to be scarse. Best practises based on
empirical data seems to be non-existent (or worse, exist but aren't published), most solutions to
problems seem to come from at most four different BM engineers. Developer forum is a good start to
them sharing their knowledge but one feels that this is still insufficient.
Lack of documented standards - what should go into DNA files compared to EMS for example.
Basically lack of experience with the product in general. The only sources of information appear to
be a BluePlanet example, PDF's and the developer forum. Where are the tips/tricks from
experienced consultants as to getting the most out of BM??
One feels quite *alone* with the product. Sometimes clients knowing more that the preferred
partners/integrators OR stumbling into problems which could have been avoided such as data
concerning sizing of databases etc
Expanded Blue Martini Support & Developer Network websites with supported / non-supported
downloads and accompanying information. More visibility is required - complete SR threads and not
just the initial post, searchable known bugs in all released versions, etc.
Publishing 6.0 - How to set up a full production environment seems to be not fully covered in the
Improved Training courses for advanced clients, this would of course also relieve support of many
issues, especially for 6.x
Blue Martini should make a concerted effort to aggregate empirical data from customers and
provide findings in courses/PDF's. Everything from tuning large installations to best practises are
vital for the customers.
Client companies want to expend the least effort possible on maintenance so that maximum effort can
be spent on development. Too much time is currently necessary to keep a Blue Martini application
functioning properly. Sanity checks, exception/performance/etc. analytical tools (a DSS report isn't
sufficient), and housekeeping tools are just some examples of what's missing.
Housekeeping - Archiving / removing old content revisions, media and not used content
Change Management - by default Blue Martini installations have limited change management
support and this has forced some clients to try implementing a change management solution using
bmimport/bmexport scripts. A complete change management solution should allow developers to
reconstruct any specific web version in time - code, dna, content, etc.
Housekeeping - Cleaning up publishing temporary tables and files
Housekeeping - Cleaning up production file systems of staged (versioned) files, LTAs, log files etc
Housekeeping - Archiving / removing old user accounts. User management appears to be
incomplete. We have users that haven't logged in for two years or more. I'm unable to delete them
through the EMS
Upgrade issues - Upgrading still difficult to achieve and prone to errors which are sometimes not
evident until sometime after the upgrade (ex. data issues)
Housekeeping difficulties – only some Blue Martini tools lend themselves to (automated)
maintenance activities whereas this should be true for all tools and elements of any Blue Martini
installation (meaning WebLogic & Hummingbird SearchServer too).
The GUI is missing a number of tools to clean up installations and manage users which any decent
config. Tool should have. Instead we are provided with scripts you have to run! The easier it is to
maintain the config the more stable the installation is as a whole.
In the European market it is essential to reach customers in their own language. This means building
a multilingual site for a specific market (in Belgium alone they speak three languages!). However it is
felt that specific functionality is only tested/works best in US locale. This hinders adoption of Blue
Martini as a viable solution for European customers
Multilingual support - Object paths cannot use high-ascii (ex. accented) characters causing
customers to implement conversion routines for their import processes
Multilingual support - Having to provide a value for the default locale even if it makes no sense
requiring input and storage overheads.
Localization – sometimes only the English value works. See the multilingual issue above. It is
related to overrides
Multilingual support - It is possible to enter data that has no value for the default locale which leads
to problems. See the multilingual issue above.
Multilingual support - BMS Remote Desktop HTML editor garbles high-ascii characters
This is essential for expansion in Europe and cannot be stressed enough and should be addressed.
The basic work has been done but appears to be incomplete
The performance of parts of BM suite seem only to have been tested with small data sets. Several
Blue Martini SQL queries have already been identified for optimization when retrieving large result
sets. This fosters the belief that there are more problems in both code and queries which are
DSSGen processing times - Especially full DSSGen runs for sites which have many months of store
data. The biggest problem is when a customer upgrades as they will have to do a complete
DSSGen run due to the schema changes.
Import/Export-Import performance has been tuned over time. Howev right now Blue Martini
Software has still performance issues with export performance
Unreasonable execution times - staging time is too lengthy for most users’ needs
Publishing 6.0 - Publishing in 6.0 generates new caches on each publish causing major
performance overheads, and limits the advantages of what publishing was supposed to provide.
Feedback from customers (especially DBA's) should be incorporated into the product, after all they
seem to be testing the product and are coming with suggestions for improvement based on the
amounts of data that they're working with.
It is felt that choosing Blue Martini as a solution gives strong vendor lockin with no insight in quality
of (BM) code and integration of other tools/API's based on Open Standards. Other companies strive to
make their products interact with tools already on the market and Open Source specific products. This
could help BM immensly by building on knowledge/ideas of others and increase acceptance of
product as a whole
Interoperability (plugins?) with internal/external standard tools. Although it is possible to get BM to
work with external tools such as Eclipse, a plugin (for Eclipse) with simple wizards/templates for
generating/editing code and DNA files (possibly based on XDocklet) would improve both quality and
time to market.
Lack of open framework allowing the client to choose which templating engine to use; Velocity,
Webmacro and Freemarker spring to mind as (faster) alternatives to JSP. XSLT could also be
More conformance to external (= non-specific to Blue Martini User community) Open Standards
Lack of support of Open Standards and/or Sun Java API Interfaces such as List which complicates
integration with other code
Information pertaining to the testing applications based on tools such as Junit/cactus is (almost) non
Lack of a Portal Architecture (based on jsr-168) which is almost a necessity these days when
building an informational site
Lack of insite of community in BM code/ideas. Programmers could help BM extend and improve
software and/or tools
Blue Martini should stop inventing the wheel and start making use of tools/projects that are readily
available for download on the net. Also knowledge of these tools is wide in the Open Source
community and could be tapped/utilised! Instead of making BlueMartini everything, make it at least
pluggable for similiar but diffirent technolgies. Tools need to be available to improve quality and
Blue Martini is perceived to be an expensive product (suite). This could be hindering a number of
potential clients to come on board thereby limiting the total number of clients..
Need for more granular pricing options (ex. prices per quantity such as 100s or 1000s of pieces
Pricing needs to be more in line with actual quality and maturity of the product as it stands today!
Seeing the number of defects that still exist in the product one feels that this should be
compensated for in the price of the product as a whole and the service fee in particular!
Furthermore the fees should be attractive to customers who will scale up their configurations and
fee based on customer demand/use and /or CPU's. This is more in line with other companies.
Overall Ranking Company
Jacob Share, Club Dial
Nice to have 3
Not applicable 0