We are busily working through the PMBOK® Guide 5th Edition updating Mosaic’s PMI training courses ready for the scheduled examination changes. Three of the more technical changes we have discovered (out of many) are:
Quality management has been tidied up. The seven basic tools of quality management are now dealt with in on place, once, in 188.8.131.52 and referenced through the rest of the chapter. The ‘magnificent 7’ are: Cause and Effect Diagrams, Flow charts, Checksheets (checklist), Pareto Diagrams, histograms, control charts and scatter diagrams. Other specific techniques are discussed in the appropriate process. There is also an attempt to relate the different project/quality cycles including the basic process groups, the ‘Plan-Do-Check-Act cycle, the cost of quality and quality assurance and control.
Monte Carlo is missing from Time Management and Cost Management! One area that needs a major update in the 6th Edition are the aspects of time and cost management focused on three point estimates and variability. Monte Carlo has been moved out of these sections into the Risk chapter, and is defined as the source of ‘contingencies’ and as a means of ‘simulating’ schedule outcomes. Very little is included about the ways simulation through Monte Carlo are developed or used. In particular, there is no discussion of how different distributions should be chosen based on the available data. Understanding the range of potential outcomes is a critical time and cost management process as is the interactions between time and cost. The treatment in the Risk chapter is not bad, just in the wrong place, hopefully in the 6th Edition the consideration of modelling outcomes will move back into the Time and Cost chapters Or there will be far clearer links drawn between the development of the raw data and its use in cost and time management.
For some reason PMI keeps bringing PERT back into consideration, ensuring the unfortunate confusion around PERT will persist for another 4 years at least! The completely inaccurate reference to ‘PERT Cost’ that crept into the 4th Edition has been killed off but the concept has reappeared in Duration Estimating (184.108.40.206), Quality Assurance (220.127.116.11) and the glossary.
There is nothing wrong with PERT being in the PMBOK if the technique is defined properly. PERT is a simplistic technique that applies a modified beta distribution and an approximation of the calculation for Standard Deviation (a polite term for inaccurate), to the activities on the critical path in a single calculation to determine the mean completion date (p50) and the effect of adding 1 Standard Deviation to approximate the p80 completion (ie, the date with an 80% probability of being achieved or bettered). PERT is prone to errors including the ‘PERT Merge Bias’ which describes the effect of a nominally sub-critical path finishing later than the critical path.
However, PERT is not synonymous with three point estimating despite a number of software vendors making the same mistake and using a ‘cute name’ to make their uncertainty calculations sound sexy. Any computation that involves simulation, different distribution options or calculating the whole network is not PERT.
PERT has an important place in history and is a useful teaching tool because you can do the calculations manually. But confusing this venerable technique with simulation and three point estimating helps no one and creates a significant communication problem – when a planner says he has done a PERT analysis you have no idea if this means a full Monte Carlo simulation, a manual PERT calculation described above or something in between. As a professional body PMI has let everyone down perpetuating this confusion.
Overall the PMBOK® Guide 5th Edition is still a significant improvement; our earlier posts have highlighted many of these changes:
To read our earlier comments: