With the recent acquisitions of Lombardi and Savvion some are questioning the long term viability of BPMS vendors. They may be right, they may be wrong, but certainly things are changing.
If you look at some of the key players in the BPM market today you see names like Progress, Software AG, Oracle and IBM, all of whom come from an application development background. Interestingly enough in the case of the first three they were also key players in the 4GL market. This brings me back to the title of this post, is BPMS now just becoming the next incarnation of application development.
Of course interestingly enough back in the 1990's we also had three other vendors of note TI, Forte and Dynasty, each of whom were trying to promote the idea of a model driven approach to development with little or no programming involved. At the time these vendors struggled and one by one fell by the wayside, due in a major part to resistance from programmers who suggested that black box application development was not appropriate. I wonder if they were not more right than people realized at the time. One could have thought of them as the BPMS vendors of their time, if it were not the case that largely they took a data-centric rather than a process-centric view of life.
Today the world is very different; many organisations have either downsized or outsourced their development teams. The armies of developers seem to be shrinking. The time taken to bring applications to market is shorter and the need for flexibility is greater.
It was during this potential trough of disillusionment that we saw the rise of package applications, component based development with a high level of granularity. The use of such applications certainly enabled organisations to focus on their own business rather than that of building systems, but the high cost of maintenance, the challenges of implementation and the monolithic style seems to be going out of favour. Again we see the needs for agility and flexibility going against the traditional ERP type approaches. It some ways these vendors helped to make us understand process better, for many it was not until they started to change their organisation to fit the application that the need to understand how people work really came home.
So we saw the rise of the BPMS approach, it was to be a marriage of good process thinking and practice with new technology to enable us to do things we never thought possible. In part this has been true; in part it has missed the objective. The needs of any organisation are complex, to consider process without considering data is highly risky, and now increasingly interest is turning to rules too. So if we are to create systems taking into account process, data and rules, then we are simply going back to plain simple application development, hence the inevitability that that as the market matures we will see the dominant players coming from the ranks of the application development vendors.
So who may yet come to the party? Sybase, Compuware, and Unify are all names that have yet to seriously enter the BPMS fray, while one cannot be sure they will, given their strong 4GL heritage I think they offer good potential exit routes for some of the BPMS vendors and it is hard to see how they can let the others in their space have things their own way for too long.
As the market evolves other vendors such as Magic Software are likely to see resurgence too, they played well in the 4GL, application generation market and are still operating in the market. Interestingly enough we may yet see another twist before Magic comes to the fore. Magic changed their business model many years ago, recognising that actually their clients did not simply want to build applications, they wanted to buy them. So Magic started to use their own products to build applications or solutions for specific verticals. This is likely to be something we see more of over the next few years, BPMS vendors selling not just a platform, but core solution blocks, with the added benefit to clients that they can easily be customised to suit their own environment. This will be important to users as they seek to maintain competitive advantage in their markets, something that the ERP solutions made difficult for them.
So it could yet be that the Object Management Group's (OMG) vision of Model Driven Architecture may yet come true, just that it may operate at a far higher level of granularity than they envisaged. MDA being used to assemble and customise components, rather than create components.
Finally then to the Modelling tool vendors, here my hope is that at least some of the independent vendors survive, they will have to work hard to deliver more by way of pre-built models and to play more easily with a greater number of vendors, but we need them. As mentioned before we need to address issues of process with issues of data and to be able to deal with rules too. When you combine that with the need to understand and mange impact, to communicate effectively and to address proper anlaysis, it is hard to see how we can manage without them. The raft of BPMS Vendor based and free tools are certainly giving users choice and make some things easier, but in the long run to model one piece of the jigsaw without the others is to create another problem.