In the first part of our interview with Mark Paulk, Senior Systems Scientist at the Institute for Software Research, Carnegie Mellon University, we talked about high maturity, PSP/TSP, culture, measurement and IT services. In this second part of the interview we take a closer look at agile and sourcing.

What is the market interest in eSourcing compared to ITIL/ISO 20000?

ITSqc looked at existing models and standards and found a niche where we could add value. CMMI for Services is more directly comparable to ITIL. The eSourcing Capability Models looked at a higher level of abstraction, i.e., anyone using IT-enabled capabilities, but not IT-specific, e.g., medical transcription and application providers.

Companies adopting eSourcing Capability Models may already be using CMM and ITIL and ISO 9001 and COBIT and specific standards, e.g., for. security. Some are using the eSourcing models as an integrating framework for other enterprise-wide models for addressing specific needs they might have. Multinationals typically are using multiple frameworks. Jeff Perdue looked at the websites of the Indian Big 5 organisation sthat showed some eSourcing, all used CMMI for Development and ISO 9001; most used ISO 27001. An organisation needs to know how to integrate multiple models across multiple areas.

As part of writing the eSourcing reports, we included comparisons to ISO 20000 / ITIL, ISO 9001, CMMI-DEV, and other frameworks. eSourcing could be used by organisations whether they use CMMI or not. A call centre might use the COPC models for example. The choice of model depends on what you consider will add value in your business domain.

Customers are concerned with evidence of capability for dealing with the kinds of challenges that the modern world is throwing at us. They would use the eSCM for Client Organizations. La Poste in France showed interest in people who do business with the postal service and have formed AeSCM (Association for eSourcing Capability Models). They are interested in the client-side and service provider models. When customers take interest, then this motivates suppliers to take note of the models. CMM adoption was driven because DoD used it for source selection. There is not the same kind of customer drive for other frameworks yet. Such policies would incentivise organisations to pick up frameworks more quickly.

Are there benchmarks for eSourcing and service?

Relatively small numbers, only some case studies. IBM Global Services and Accenture use eSourcing. It does not have the same depth of penetration as ISO 9001 or CMMI. However, CMM has over 20 years of deployment! CMMI-ACQ and CMMI-SVC have the advantage of leveraging an existing user base of CMMI-DEV. For some ISO 9001 variants, it took years for uptake. eSourcing is more f a governance model than an engineering model.

What do you know now that you wish you knew when you started with CMM?

What is still not adequately addressed is capturing product knowledge. One process area to be added into SW-CMM v2 was going to deal with product lines and reuse. This places process measurement work in context and allows you to compare apples to apples and oranges to oranges. It did not make it into the CMMI work. This is an opportunity that companies need to be considering. For high maturity, organisations need to create comparable datasets based on product insight.

We learned to deal with organisations that are truly ready for change. For early assessments, 90% of the organisations did not act on assessment results to change the organisation. For later assessments, we got better at identifying problems, so the success rate went up. Too many organisations don’t want to change behaviour and just hope that things will change. Problems are hard. If it was easy, they would have already changed them. Change is hard. An assessment readiness survey was published in the lead assessors guide to help in confirming that an organisation has senior management sponsorship and cultural readiness.

We also need to steal ideas from other areas, e.g., Chris Argyris on organisational learning and research in innovative organizations. Much has been learned since CMM was originally published.

However, the class on software process that I teach every year has assigned readings of research published recently. The fundamental work published still applies – the foundation for TQM and SPI. It’s just that every year we learn more, but we do stand on the shoulders of giants.

Are agile and CMMI opposing?

The difference is in the emphasis on explicit knowledge management.
For CMMI, you need to write it down.
For Agile, there is tacit knowledge management & face-to-face interaction.

They are points on a spectrum rather than at necessarily at cross purposes. CMMI users can find a point where tacit info will be most efficient. Larger teams/more complex organisations need more explicit capturing of knowledge. Safety/life-essential systems need more explicit information capture.

There is a lot of familiarity in the agile concepts, looking at, for example, de Marco and Lister in Peopleware – good engineering and management practices that were around for many years and are implemented in an “extreme way” in the agile methods. There is a difference in the degree of emerging properties. Differences of a kind that are revolutionary. We can see the seeds of where these ideas brewed. There is a potential for incorporating the good ideas of the agile movement into the more formal/high ceremony organisations where appropriate for kinds of work being done.

At this year’s IBM Innovate conference, Scott Ambler talked about ‘disciplined agile’. Which comes first? Maturity or agility?

Either. From a process perspective, high maturity organisations should be looking for opportunities to build more efficiently and effectively, and agile methods have interesting characteristics that high maturity can actively consider. High maturity organizations could pilot or consider agile methods.

Organisations adopting SCRUM or XP realise that they are aimed at project-level concerns and do not deal with organisational change management, and those companies will need to pick up themes in CMMI or other models to deal with the change management issues effectively.

Both can be complementary. Depends on where an organisation is and the predominant culture.

Final quote or message?

The bottom line for the last 20+ years is that I have been doing what I could to help organisations build software more effectively & efficiently. There are many factors that affect us doing software well: process, hiring good people, innovation, risk management, and so forth. Models such as the People CMM can help with some of these, as can agile methods and other research.

Process and quality will continue to be important for the foreseeable future. Companies will need to pay deliberate attention to process.

People and risk management are already being addressed. The balance in emphasis will change over time for organizations; as old problems are addressed, new concerns predominate. We will never be able to take time off from these issues, they will always be at play.

We started the article, talking about imaginary numbers. If we see how CMM has morphed to keep up with the times, imaginary numbers and models become achievable and tangible through commitment and execution!

To vote for the Alinement Network to organise a webinar presented by Mark Paulk, please provide your comments below.