I finally caught up with Pentaho, which along with Jaspersoft is one of the two most visible open source business intelligence companies, Actuate perhaps excepted. Highlights included:
- Much like Jaspersoft, Pentaho’s initial focus was mainly on embedded, operational BI.
- However, Pentaho now feels it has a decent end-user GUI as well, and traditional-BI is a bigger part of sales.
- Also, some sales are focused on data integration, perhaps in support of more traditional BI products. Pentaho has even had an Ab Initio replacement in data integration. (Can there be any change more extreme than going from Ab Initio to open source?)
- As an example of technical breadth, Pentaho says that its Mondrian OLAP engine is used by Jaspersoft.
- Pentaho has Excel output, but not in the form of live formulas.
- Pentaho does XQuery.
- Industries with more Pentaho adoption than average include:
- Financial services (traditionally open-source-friendly, according to Pentaho)
- Government (ditto)
- Web 2.0 (obviously ditto)
- Travel/transportation (cash-strapped)
- Frontier Airlines is a Pentaho/Greenplum customer.
- TradeDoubler is a Pentaho/InfoBright customer. (Pentaho thinks that TradeDoubler reloads its warehouse every day, which if true frankly casts some doubt on InfoBright’s architecture.)
- Data mining is something of a Pentaho sideline. There’s some university in New Zealand that built data mining capabilities in Pentaho, and some data mining research is done in that. Separately, Pentaho has been integrated with R.
- Community contributions are concentrated in the areas you’d expect — features some user or system integrator needs for a specific project, connectors, bug reports, and the like.
The briefing included one of the better slide decks I’ve seen in a while, which Pentaho gave me permission to share (in somewhat abbreviated form) here. In particular, Pentaho provided customer examples illustrating most of the use cases cited above.
Pentaho facts and figures include:
- Pentaho was founded in 2004. The first dozen or so reference customers were acquired in 2007. Before that usage of the product was mainly downloads of a free version.
- Actually, Pentaho’s free usage is more focused on embedded libraries, while paid usage is more skewed to traditional BI.
- Pentaho’s average selling price is $24-25K for first year revenue, which is extremely close to Jaspersoft’s figure.
- There are 100,000+ downloads per month, but Pentaho cautions that’s a very misleading figure. Some users download over 100 different pieces of the product, including for example all the national language support and all the different platform-specific support pieces.
- Pentaho doesn’t offer much in the way of more realistic metric of company size or success.
- Europe provides 35-40% of Pentaho revenue.
- Pentaho has at least one Asia/Pacific reference.
- 50% or so of Pentaho customers are on MySQL. Oracle and Postgres are in a rough tie for #2. That appears to be PostgreSQL rather than EnterpriseDB’s Postgres Plus.