A new approach to enterprise software development

I met with Brian Gentile, CEO of open source BI vendor Jaspersoft. Brian is a self-proclaimed “fan of new generation software”. Here is what he believes constitutes new generation software—and how it differs from traditional enterprise software. This article explores further the tenets of the recent article ‘Is the traditional BI market in decline?’.

First, Jaspersoft has built a club of volunteers who give time and effort to the cause. It’s a bit like the parents’ committee at a well-run primary school where parents want a stake in creating the best learning environment for their kids. Jaspersoft has 175,000 registered programmers in its ‘community’ of fans who have a similar ownership stake in Jaspersoft.

These unpaid end users fix bugs and develop enhancements to Jaspersoft’s open source product, iReport. They champion the product, and provide ‘free’ programming and powerful word-of-mouth recommendations to other potential users. In addition they vote on the new features for the next release—so product development is largely driven by user priorities.

Second, the software architecture is lightweight, web-based, has open APIs, and is easy to install and upgrade. Everything is built on the Java platform and the whole software suite is only 500MB. It requires little installation and professional services consulting support, and is available both for web download and through Jaspersoft’s OEM partners as a SaaS deployment.

Third, it’s low-cost and affordable. The entry level price for the commercial version is $10,000.

Compare this with how the enterprise software vendors tackle these areas:

First, enterprise software vendors have a network of resellers, complementary software vendors, and systems integrators and consultants. These mercenaries add value to the core product only where and when it is in their best commercial interest. Vendor loyalty is typically low—software and services partners often jump ship to where the profit potential is highest. Acts of charity are not high on their agenda.

Product development rests on the needs of a small number of key customers—who may or may not be representative of the market. The vendors control which features are included or excluded from the commercial release.

Second, the traditional enterprise software architecture is heavyweight and leans towards proprietary lock-ins. For example, SAP Business Objects’ BI suite is said to contain c. 30GB of code (ie 60x the size of Jaspersoft’s BI suite). Installation and upgrades are typically long and rely on on-site consulting and services provision over many months.

Third, enterprise software is expensive. Brian reckons Jaspersoft software costs around 1/10 of the cost of comparable commercial enterprise software. The enterprise vendors’ business models demand a premium price in excess of $100K for the most part. Implementation services expenses could double that cost.

So, Jaspersoft has built quite a compelling value proposition. Participative, collaborative, lightweight, fast to implement, transparent and open, and lower cost than its legacy competitors. Granted, it is mainly adopted and used by techie programmers, and it doesn’t offer the sexy front end user experience of SAP Business Object’s Crystal reports, for example. Neither does Jaspersoft offer the comfort of the size and support resources of an SAP or similar. However Jaspersoft is clearly doing something right as iReport downloads are now at the staggering rate of a 250,000 per month.

So what can customers learn from Jaspersoft’s software development techniques?

1) Engage with the hearts and minds of your community (ie ‘The Business’)—not on the basis of “it’s your job to help IT” but rather that “it’s fun to be involved and your contribution will be recognised”. Co-creation and collaboration are the watchwords.

2) Create and choose software that is light, flexible, and easy, and incorporates ‘the wisdom of crowds’. Don’t get dragged down by miles of inflexible code that cannot be re-purposed to reflect the business challenges of tomorrow. SOA is very important here.

3) Experiment with low cost web-based solutions. If they don’t work, junk them, and try something else. Don’t put all your eggs in one big enterprise software investment basket with an uncertain outcome.