In part one of this series I described how the Tapscott and Williams Wikinomics book mentioned a few things that gave me ideas for semantic web projects. One was the concept of open source Customer Relationship Management packages, which I hadn’t heard of before. The book didn’t mention any specific ones, but a Google search on “Open Source” CRM gets plenty of hits.
What does CRM software do? Although the Wikipedia CRM entry contains the warning “This article or section appears to contain a large number of buzzwords and may require cleanup” (It’s nice to see that the Wikipedians worry about this) the following passage from it makes sense to me:
CRM software is essentially meant to address the needs of Marketing, Sales and Distribution, and Customer Service and Support divisions within an organization and allow the three to share data on prospects, customers, partners, competitors and employees.
Making it easier to share data across organizational boundaries is a big goal of the semantic web, and if businesses are going to get more value from semweb technology, this looks like a fertile place to plant something. The “R” in “CRM” can have some interesting implications in a semweb application; the ability to identify and track a company’s relationship(s) with a given customer (and related customers) in a way that improves those relationships would be quite a payoff for adding a cool new technology to an existing software infrastructure. If the software is open source, there should be hooks to add such new features.
A few quick web searches don’t turn up an existing CRM ontology or taxonomy, so one may have to be built. As database packages, a lot of naming and relationships will already be taken care of, but you don’t want to extrapolate from the first package you look at to the whole CRM world. If building an ontology, don’t fall into the common semantic web developer trap of building a huge ontology and then telling the world to come and get it—you’re better off with the agile approach of creating a small ontology, developing working code around it, demoing that, and then building from there.
Before committing to a particular open source CRM package to work with, I’d look through the discussion forums for each and see what users are doing and trying to do with them, as well as what the developers working with the code have to say and what resources are available to answer their questions. To see positive results from the addition of new features, you’d want to work with real data, so the forums would provide candidates for partners to work with. If you come up with the right partner and proposal, maybe you could get paid for this work!
A review of multiple CRM packages would provide good input for the development of a CRM ontology, and it would also broaden your perspective on what people want out of CRM systems and what different packages are doing to meet or exceed those needs. For that matter, you may as well look over the promises made by commercial packages. CRM is a big business with its own culture and trade press, so there are plenty of places to do research.
Keep asking yourself “if these people don’t know about semantic web technologies, what are they missing? What could it add here to make this software do more?” You can be the person who shows them.
Me again :-) We are embarking upon an effort (community based) to create Ontologies for XBRL Taxonomies. In the same vein, we will embark on a similar effort re. CRM. I already floated this idea to the Linking Open Data project’s mailing list last week as part of the THALIA project enhancements re. incorporation of SPARQL, RDF, and OWL.
I have been searching without success for an eCRM Ontology (can’t believe there isn’t one out there). Thus, we will build should nothing show up from the public domain in the next week or so.\
To a first approximation, there are no open-source CRM packages. There are many that claim to be so, but on investigation it turns out that they use the Mozilla Public License or some equivalent with the definitely non-Open-Source addition of rules saying that modified versions must keep their logos with specified content and minimum-size restrictions in place on every displayed screen.
That is good business for them (it discourages competitors from grabbing the whole thing, hiking off the logo, and substituting their own), but it breaks the Open Source Definition requirement #10, which says that licensing can’t depend on the particular technology: if you try to reuse any of the code within a headless application, for example, you are screwed.
The matter has been taken up on the firstname.lastname@example.org mailing list lately, and some CRM vendors are changing their tunes. But read carefully.
Kingsley: one of the other ideas I was going to write up was going to be about XBRL, because they’ve worked out so many taxonomies with direct business applications. Keep us posted on what you come up with…