I’m not going to try to compete with Sandy’s wall-to-wall coverage of this event. Mine will be more impressionistic.
Simon Hayward keynote. Gartner likes to sell futures on technology. It’s what they do. Simon has a chart of the value realization from BPM over time, with 3 curves. Today the “productivity” curve is highest. In 2012 (safely over the horizon) the “visibility” curve overtakes it. In 2017 (I’ll be dead by then) the “innovation” curve reigns supreme. After that, I don’t know, maybe global warming wipes out the earth. Does this kind of chart really advance the ball?
An interesting difference between the Gartner and Brainstorm BPM conference is that at Gartner the keynoters assume and universally assert that if you’re not going through the whole model-design-deploy-execute-monitor-analyze-optimize thing you’re not really doing BPM. At Brainstorm the keynoting class generally advances the notion that BPM ends with modeling and “process thinking”… although the vendors who sponsor the thing really wish they would stop saying that. I like the Gartner approach, but which one is addressing the “real” BPM marketplace?
Hayward and Janelle Hill both make a big deal of the idea that BPM has to make the process “explicit” but leave activities decoupled from their actual implementation. Explicit here is kind of a code word. I think it means fully defined at the logical level, i.e. almost executable. They also say that BPM must empower the business side (not IT) to make changes to the (executable) process model. They make this sound like reality, even though it is a vision (one that I support)… rarely fulfilled by actual products.
Hayward lists 3 alternative paths to explicit (i.e. executable) process: BPMS, “process-aware middleware”, and the composite app evolution of packaged apps. The latter two I think are code for Microsoft and Oracle/SAP, respectively. I’m not sure. He does not see the fact that these paths are alternatives as a bad thing, although he admits it forces buyers into tough choices. Not a great answer.
A recurring theme is the requirement for a new role in the organization, what they call a business process analyst, with “crossover” skills blending business and IT. I think it’s someone with the smarts of IT but with a business perspective on process.
One of the SPA’s (Gartner predictions) to be discussed later in the conference says that process modeling will be adopted on a scale half as big as spreadsheets. As someone now invested in BPMN training, I can only say, “I wish…” Following that theme, I went to a session called Process Modeling for Profit by Marc Kerremans from the consulting side of Gartner, i.e not an analyst. He is one of those consultants that enlarges “process modeling” to mean all types of business modeling from business goals and strategic alignment to organizational modeling, process selection, etc. The “for Profit” part is clearly meant for someone whose life is measured in billable hours. At the end of the session I asked Mr Kerremans how he saw the adoption of BPMN versus high-end proprietary BPA tools. He said many companies had BPMN on their requirements list but nobody used it and he didn’t recommend it. In other words he’s sticking with the $15,000-per-seat tools. I guess he didn’t get the message about the SPA. Let’s see, 30 Million seats x $15,000 per seat…
Next, Janelle Hill, definitely the class of the lot. Reiterating several of Hayward’s points… BPM is built to change by the business itself, with real-time feedback of performance improvements resulting from the change, via management dashboards. Business-IT collaboration on implementation, including logical design, UI design, and business rule design. I agree with all of this. But again, I think more a vision than reality today.
She also talked about their BPM magic quadrant, due for update in September. Says you can’t evaluate BPMS purely on a feature list, given the wide range of design skills available in the organization, and the diversity of process types. These are themes I emphasize myself in the 2006 (and forthcoming 2007) BPMS Report series.
The other thing fascinating about this conference so far is the huge emphasis Gartner places on simulation as a tool for process analysis. As I’ve written about extensively, the simulation tools available to business process analysts seem to me in fact nearly useless, and in any case users have no idea how to make simulation analysis meaningful. So I wonder how much of this is reality-based versus faith-based.
I would like to follow up on your comment regarding simulation. Here down under, we’ve been working with organisations using BPMN that have ended up buying Systems Architect for their BPM projects. One of the arguments for this buy is that they wanted to do simulation. But once they got the tool, they realised that the crux of simulation is not so much the tool, but rather the methodology. As you mention, one problem is that people don’t know how to exploit the output of the simulation. They typically stop at measuring top-level indicatiors such as average execution time. But another major issue is that people pay little attention to setting the simulation parameters based on real data. It’s a shame, because many of these organisations have a lot of data dispersed across their systems (e.g. CRM, ERP, job tracking tools, web logs), that could be used to set these simulation parameters. They just need to dig into their systems’ tables, logs, reports, etc. and they would find a mine of data that they could use to tune up their simulation parameters, if only they knew where to look for. After all, it’s not that their processes do not exist. They exist and they are run every day.
This is to say that what we need is to have comprehensive business process simulation methodologies. Tools should be designed (and bought) to support these methodologies, rather than the other way around (meaning: we buy the tool, then we ask ourselves how will we use it).
Also, as surprising as it may seem to us, some people still confuse simulation with “animation” (i.e. watching little bubbles go from one task to the next in the model interactively or in batch). But this is another story…
marlon
process simulation…
Hi. Very nice blog. I\’ve been reading your other entries all day long..lol….