Talk:Project portfolio management/Archives/2013

Definition
There's no actual definition of Project Portfolio management here. Just a description, but not a definition. Reads like a marketing brochure. I'd expect a definition of the type: The practice of collecting and viewing all IT change projects in a consistent fashion that allows decisions based on the financial returns expected by IT projects. That's off the top of my head. I'm sure there are some rational sources for a definition. Also, there should be a link back to IT Portfolio Management and a cross-link to Application Portfolio Management. --Nickmalik 19:58, 17 October 2006 (UTC)


 * I agree &mdash; it does read like a cross between a marketing brochure and a training seminar: the phrase "next generation" in the first sentence gives that away. This article needs some critical analysis (what are the alternatives/competitors to PPM? what are the main criticisms of PPM?) and less promotion.  I've tagged it for POV. David 11:07, 16 April 2007 (UTC)

criticisms of PPM
If you are looking for criticisms of PPM, don't worry - there are plenty. This is mostly a term used by several software firms that have developed tools for managing projects. However, there is almost no basis in economic theory or modern portfolio theory in these tools. I see the author included optimization under constraints for a portfolio but the lack of mention of modern portfolio theory makes me think that the author had no knowledge of this field (it won the Nobel Prize in 1990). I'll add a few comments to even this out. Hubbardaie 18:11, 8 June 2007 (UTC)

PPM as a valid topic
I am not a unilateral advocate of PPM. I do think that there is too much of one perspective represented as yet (even after my limited editing). PPM may have started out as a vendor marketing label, but is it increasingly being used by Project Management Offices (PMOs) as a way to capture and represent the need for a longer term perspective. Classic approaches to selecting IT projects included treating them as captial investments for which once the money was allocated, only massive failures cased deallocations -- in other words, no turning back.

As yet, I find the tone of the primary author to be unnecessarily antagonistic to the lack of financial precision in PPM. The label helps management see ongoing projects in the context of new choices for projects.

BTW, financial portfolio theory as applied in the real world must use imperfect measures of risk sensitivity (on behalf of the portfolio owner) and does not guarantee the outcome for a particular portfolio owner, but does show that over time, for most portfolios, that applying the theory will produce value. A particular limit to rigorous application of financial models for IT portfolios is that an organization is not an open market system -- only a subset of all possible projects will ever be valid for a particular organization. This type of bias is excluded from the financial analysis models by assumption.

I believe that concepts of real options theory would be valuable here (rather than focus on rigorous financial metrics). The concept of a portfolio for IT projects as a set of real options for future organizational outcomes is powerful and reflects what earned new portfolio theory the Nobel Prize.

To give some balance, I added a small section noting that PPM is unlikely to be meaningful for small or young organizations.

lika2know 10:28:10, 22 June 2007


 * First, nobody claims that the value of Modern Porfolio Theory (MPT), the method that won the Nobel in 1990, must produce perfect predictions in order to be an improvement. In fact, anyone who knows anything about risk analysis knows that its not about perfect predictions or "guarantees" but making better bets - so that on average returns are better than they otherwsie would be.  If your criterion would be improvement rather than perfection, then the validity of MPT is a demostrable fact.
 * Second, there is no need to assume that IT must be an open market for such financial methods to apply. All you need is big decsions under uncertainty...I don't think anyone disputes that.  I routinely apply these methods within IT for all of my clients.  The fact that IT projects are not sold on an open market is no obstacle whatsoever to quantifying risk and applying sound porfolio optimization methods.
 * Third, the only problem with the powerful Options Theory is that is usually the only portfolio management theory most IT people can think of. Actually, MPT won the Nobel prior to OT and OT is arguably a subset of the porfolio management issues dealt with by MPT (MPT deals with all uncertain investments, not just strict options).  Also, most people who are attempting to apply Black-Scholes (the theoretical basis of OT) are doing it wrong.  Real Options are almost always better modeled by older decision-theory modeling methods.
 * Fourth, the REASON you have any risk at all is because you have "imperfect measures". You have NO perfect measures at all.  Thats why we need to quantify uncertainty and optimize portfolios under a state of uncertainty.  To say the shortcoming of quantifying risk is a lack of perfect knowledge misses the entire point of risk in the first place.
 * Finally, IT PMO's are decades behind almost every other industry in dealing with these issues and they act as if they are blazing trails. Actuaries have sophisticated methods for measuring risk, portfolio analysts know the optimal solutions, economists know how to quantify the "intangibles", and operations research people know how to optimize limited resources under constraints.  The solution is to quit making up their own solutions and figure out how to catch up with everyone else.  If it were entirely up to me I would just redirect this to Applied Information EconomicsHubbardaie 20:27, 22 June 2007 (UTC)

I am confused by the definition of risk as being due entirely to uncertainty of information. If by that you mean that we could avoid all risk by knowing in advance where not to be and when not to be there (e.g., to not get struck by lightening)...that implies that we have also complete freedom to change our behavior (and the behavior of animate and inanimate others)...which we don't. (Unless we're dead, I suppose.)

Financial markets work because they simplify a huge amount of information. You are right to emphasize the centrality of metrics, but maybe confused about which came first. Financial markets as we know them today can only thrive where standard metrics already exist -- otherwise markets cannot operate effectively remotely from the object that they are trading in. (cf. Linkater, Andro's book (2003) "Measuring America).  The concepts of risk we use arose around games with limited and known possible outcomes.  The ability to trade in company shares arises when individuals are protected from the complete risk by the creation of a limited liability entity; in other words, there's a reason only a few of all the investors due puts and calls.  One of the major contributions of real option theory is to move away from the pure dollars basis and look at the choices that can be provided in the future by investments today.

Back to this entry: The coverage in IT portfolio management has more of the perspective that I would like to see here. That entry reflects that projects are undertaken for a variety of reasons, only some of which have (easily, reliably, agreed-upon) measures. If the "higher level" concept of IT Portfolio Management recognizes the limits, I think this entry should as well. I would like to edit major parts of this entry to remove the somewhat scolding tone as I think it does not encourage people struggling to be more systematic to keep on with their efforts.

It is an open topic as to why IT is so behind other domains in obtaining and applying metrics which would make porfolio theory more applicable. Youth of the domain has been an argument. Rapid technology change is another. But perhaps, the most significan reason can be traced to the definition of a project (a la Project Management Institute) which emphasizes the temporary nature and the uniqueness of the product/service/result. If projects were not inherently different, they would be called "operations".

The creation of PMOs in major corporations has paralleled the deployment of ERP systems in which an attempt is made to replicate a successful project across many locations and at different times. If you are thinking of these types of projects, I am more hopeful that metrics could be used, but they will always be "local" in that they reflect what can be measured and understood within the firm. This makes them difficult to generalize, especially compared to metrics such as human mortality or property damage. Both of these have a stable unit upon which to conduct comparisons. An IT project hardly qualifies as a stable base.

Outsourcing is another potential source for developing comparability and therefore metrics since a dollar is placed on the service; however, the most comparable outsourcing approaches are "operations". Outsourcing system development has no better record of predictable success than "insourcing" does.

One assumption behind portfolio theory that doesn't translate so well is that projects are independent of each other. It is much less likely that a set of investments under consideration would have the types of interlinkages that projects within an organization have. This means in that some cases, it is necessary to accept one project with a particular risk level (and nature of risk) in order to do something else. In this way, the real options theory definitely has value. Conventional portfolio models assume that the outcomes and risks of individual investments are independent (heterogeneity) -- an assumption clearly violated in the pool of projects available to decision makers within an organization.

Projects are so varied and distributed in time, a single organization is likely to have difficulty capturing the needed metrics without experiencing a shift in the environmental variables (which wash out of financial portfolio theory by picking a point in time). I am aware of some efforts to pool information across some organizations (e.g., state departments of transportation) to develop models to define projects such as bridge repair and replacement. Companies who make IT products for others would presumably be applying product portfolio concepts (I think they are more powerful than project portfolio concepts). Finally, since IT projects are enablers of other capabilities in the organization, by the time "must do" projects are identified, the latitude for trying to achieve some kind of portfolio objective is often quite narrow, something that is also not considered in financial-instrument-based portfolio management.

My minimal case conclusion: IT projects are not very comparable (so the metrics are poor or difficult to develop), and if they did have usable metrics, those metrics would violate a key statistical assumption of common portfolio models. I recommend we redirect this topic to: IT portfolio management[[User:lika2know|lika2know] 15:18, 8 July 2007

You seem to hold several erroneous assumptions about several issues. I will try to address all of them:
 * Your first sentence indicates some confusion about uncertainty and risk. See the Uncertainty article for clarification.
 * No, I'm not confused about whether metrics came first nor did I make any claim at all that had any bearing on which came first. The history you describe about trading is in no way at odds with anything I said.
 * The only thing you need to apply MPT is more than one investment with uncertain returns. No formula in MPT implies the need for what you call a "stable base".  It only requires that we be able to quantify our uncertainty.
 * If you mean that not all IT projects have defined measures, I agree. But they should be measured.  If you mean they are not all measurable, I don't agree.  If it is observable in any way, directly or indirectly, it can be observed in a structured way that allows for uncertainty reduction.  See the measurement article.
 * There is no more reason IT investments are limited to "local" measures within the firm than any other type of investment. My stochastic models of IT investments routinely include probability functions of factors outside the firm.
 * Nothing I said is affected, contradicted or otherwise addressed by the relative success or failure of outsourcing.
 * You said that IT project independence does not translate well from portfolio theory. First, MPT allows for both dependent and independent investments.  Investment constraints such as "project A increases the return on project B" are just additional mathematical constraints.  I don't know what "conventional" portfolio management models you are thinking of that assume independence, but MPT never did that since even the 1950's.
 * You are imagining difficulties with measurement where these difficulties don't really exist. And what environmental variables "wash out" when you pick a point in time in a financial portfolio?  Again, you are assuming characteristics of MPT that are not really characteristics of MPT.
 * Again, "narrow lattitudes" are just another constraint that can be modeled mathematically. Other portfolios such as M&A or real estate have exactly those kinds of contraints.  And even when projects are mandated, there are often multiple ways of satisfying the mandate that, itself, is best modeled as a constrained portfolio.  I tell my clients that if they truly are so constrained that they have no real choices, then they have far too many managers on the payroll ;-)
 * IT projects don't have to be "comparable" in any way other than they have uncertain returns. Again, that is all you need to apply MPT - a set of investments and their probability distributions of returns.  Its the same way you compare two completely different companies traded on the stock market.
 * Why would these metrics "violate a key statistical assumption of common portfolio models"? They do no such thing.  I just wrote a book that discusses statistics and I've applied portfolio models frequently for my clients.  I suspect that the assumption you are refering to is only your assumption, not anyone else's.

One final note. Every "obstacle" you point out is refuted by my personal professional experience of doing everything you just said can't be done. This is actually what I do for a living. I've written many articles about it and my book discusses all this in some detail ("How to Measure Anything: Finding the Value of Intangibles in Business" John Wiley & Sons, 2007). Don't just assume that because IT projects are in some ways not like stocks, that MPT doesn't apply. Look at the actual mathematics of MPT and see for yourself if it ever states a requirement for any of the distinctions you make.Hubbardaie 04:31, 9 July 2007 (UTC) I have to admit that as someone with a Ph.D. in information systems (vs. an economist), you have me beat on the details of the models. The question being: is this entry meant to capture what is the current practice vs. chide people for not using an advanced technique (which many firms will have trouble implementing IMHO)? I'll leave that up to others. lika2know 2:45pm CT 9 July 2007

I would not say that it is so much about needing to "chide people for not using an advanced technique." Many wikipedia articles about various methodologies include a section about criticisms, especially relatively new methods. I think this gives a balanced view that an encyclopedic reference needs. Looking at the entire article, I would say there is not more negative than positive material. But it is important to point out, as I think the balanced criticisms here do, that PPM is not at all like the portfolio management used in some other industries. Someone familiar with MPT but not familiar with PPM in IT might think it usually includes more advanced methods they are familiar with unless we explicitly point out that it currently does not. However, many methodology articles are not just about what most organizations attempt to do but also gives a glimpse of the most advanced methods in the field. After all, some IT departments really are using the more advanced methods. Just as many articles on computer science and the physical sciences discuss the cutting edge research, such topics can be discussed in this article, too.Hubbardaie 21:53, 9 July 2007 (UTC)

Misdirected Debate
The threads on this entry have gotten misdirected in a zealous comparison with MPT, which is interesting but not especially relevant. Perhaps the word "Portfolio" has become a flash point, but I see only limited value in that line of discussion. It's a footnote and only loosely related to the issues addressed by PPM. I agree with Nickmalik that a more practical definition would be useful. Demian Entrekin 22:29, 20 September 2007 (UTC)

Wikipedia entry purpose: Prosletyzing, berating or educating?
I am very puzzled by the dismissive and highly academic tone of this entry on Project Portfolio Management. Stepping away from the realm of the PhD and into the real world where most projects fail, and most project investment decisions are based on anything but sound economic principles, I've experienced significant changes in my clients simply by starting this journey toward better decision making. In one client, huge investment was being made in what time has shown to be a very flawed technology direction, only because its chief proponent wielded a great deal of personal influence over other executives in the company. Never was there any serious, dispassionate scrutiny of the risks to the expected ROI of the project, or whether or not the company should continue to invest once those risks started to be realized. Surely any rational, risk-based discussion is better than "my VP can beat up your VP" as a basis for decision making, even if it is somewhat subjective. Let's face it, we'd all like to be able to predict the future, but none of us truly can. In cases of many project investment decisions, there's not a statistically valid basis for economically gauging these risks - the best we can do is leverage our experience to make the best judgment we can.

The French are fond of saying <> - while striving for perfection we often overlook real immediate opportunities for improvement. The engineer as opposed to the academician is more likely to choose a practical approach that improves the decision making process and visibility of quantifiable results of those decisions, even if that approach is not perfect.

I think the chief author of this entry has some sort of axe to grind. It's truly a shame that we luddites using and benefiting from the portfolio management metaphor are so offensive to the author. I respectfully suggest this person a) get over it and b) refrain from trying to use Wikipedia as his/her bully pulpit.

171.71.244.136 (talk) 17:37, 3 April 2008 (UTC)

I have editied the entry to try to moderate this a bit by providing a definition of the PPM process which will hopefully be viewed as "non-partisan"...

171.71.244.136 (talk)

171.71.244.136 (talk) 16:59, 3 April 2008 (UTC)

To the anon editor who made the last round of changes, they look good. I believe that your edits increase the balance of this article and do a better job of reflecting a neutral point of view. Cudos. --Nickmalik (talk) 21:11, 7 April 2008 (UTC)

Reference to P3Portal
While I did not remove the reference to P3Portal's article, I support the removal of that link. I view it as Wikipedia abuse to to add links to articles that add nothing to the content of the Wiki article, especially if they are from anonymous or poorly qualified sources. While I am sympathetic to the difficulty of an organization to "get started" and be relevant, I believe that individual authors should seek to gain necessary notariety through research published in reference journals. That article can be cited in Wikipedia. Simply putting up a web page is not sufficient to cite the source in Wikipedia.

Removing commercial software "list"
I think we should remove the sentence: "Vendors that provide PPM software include HP, Microsoft, Oracle (Primavera) and Planisware." I don't see how it adds anything to the article and it's not a complete list so it doesn't make sense to include some vendors and not to include others. — Preceding unsigned comment added by Guy.other (talk • contribs) 17:54, 18 July 2012 (UTC)