Editing Earned value management (section)

== Limitations ==
[[File:Forecasting at compeletion.png|thumb|ER = Earned Regression = 953{{pb}}SPI(R) = ER / AT = 953 / 866 = 1.10{{pb}}MD(R) = 100 / 0.0408 = 2449{{pb}}IEAC(R) = MD(R) / SPI(R) = 2449 / 1.10 = 2223]]Proponents of EVM note a number of issues with implementing it,<ref>{{cite web
 | last = Alleman
 | first = Glen
 | title = Herding Cats : Issues with Deploying Earned Value Management
 | url = http://herdingcats.typepad.com/my_weblog/2012/06/issues-with-deploying-earned-value-management.html
 | date = 2012-06-02
 | access-date = 2013-04-04}}</ref><ref>{{cite web
 | last = Schulze
 | first = E.
 | title = How Earned Value Management is Limited
 | url = http://www.brighthubpm.com/monitoring-projects/10056-how-earned-value-management-is-limited/
 | date = 2010-06-21
 | access-date = 2013-04-04}}
</ref> and further limitations may be inherent to the concept itself.

Because EVM requires quantification of a project plan, it is often perceived to be inapplicable to discovery-driven or [[Agile software development]] projects. For example, it may be impossible to plan certain [[research]] projects far in advance, because research itself uncovers some opportunities (research paths) and actively eliminates others. However, another school of thought holds that all work can be planned, even if in weekly [[timeboxing|timeboxes]] or other short [[iterative development|increments]].<ref>{{Cite book|last=Piyush Solanki|title=EARNED VALUE MANAGEMENT: Integrated View of Cost and Schedule Performance|publisher=Global India Publications Pvt Ltd|year=2009|isbn=978-93-80228-52-5|location=New Delhi|pages=13}}</ref>

Traditional EVM is not intended for non-discrete (continuous) effort. In traditional EVM standards, non-discrete effort is called "level of effort" (LOE). If a project plan contains a significant portion of LOE, and the LOE is intermixed with discrete effort, EVM results will be contaminated.<ref>{{cite web
 | title = The Guide
 | publisher = Defense Contracting Management Agency
 | url = http://guidebook.dcma.mil/79/evhelp/perform.htm#Level%20of%20Effort%20(LOE)
 | access-date = 2011-09-22
 | archive-url = https://web.archive.org/web/20130308083007/http://guidebook.dcma.mil/79/evhelp/perform.htm#Level%20of%20Effort%20(LOE)
 | archive-date = 2013-03-08
 | url-status = dead
 }}</ref> This is another area of EVM research.

Traditional definitions of EVM typically assume that [[project accounting]] and [[project network|project network schedule management]] are prerequisites to achieving any benefit from EVM. Many small projects don't satisfy either of these prerequisites, but they too can benefit from EVM, as described for [[#Simple implementations (emphasizing only technical performance)|simple implementations]], above. Other projects can be planned with a project network, but do not have access to true and timely actual cost data. In practice, the collection of true and timely actual cost data can be the most difficult aspect of EVM. Such projects can benefit from EVM, as described for [[#Intermediate implementations (integrating technical and schedule performance)|intermediate implementations]], above, and [[earned schedule]].

As a means of overcoming objections to EVM's lack of connection to qualitative performance issues, the [[Naval Air Systems Command]] (NAVAIR) PEO(A) organization initiated a project in the late 1990s to integrate true technical achievement into EVM projections by utilizing risk profiles. These risk profiles anticipate opportunities that may be revealed and possibly be exploited as development and testing proceeds. The published research resulted in a Technical Performance Management (TPM) methodology and software application that is still used by many DoD agencies in informing EVM estimates with technical achievement.<ref>{{cite web
 | last = Pisano
 | first = Nicholas D.
 | title = Technical Performance Measurement, Earned Value, and Risk Management: An Integrated Diagnostic Tool for Program Management
 | publisher = Defense Acquisition University Acquisition Research Symposium
 | year = 1999
 | url = https://apps.dtic.mil/sti/pdfs/ADA404646.pdf
 | access-date = 2018-02-06
 | archive-date = 2013-09-16
 | archive-url = https://web.archive.org/web/20130916124041/http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA404646
 | url-status = live
 }}</ref>
The research was peer-reviewed and was the recipient of the Defense Acquisition University Acquisition Research Symposium 1997 Acker Award for excellence in the exchange of information in the field of acquisition research.

There is the difficulty inherent for any periodic monitoring of synchronizing data timing: actual deliveries, actual invoicing, and the date the EVM analysis is done are all independent, so that some items have arrived but their invoicing has not and by the time analysis is delivered the data will likely be weeks behind events. This may limit EVM to a less tactical or less definitive role where use is combined with other forms to explain why or add recent news and manage future expectations.

There is a measurement limitation for how precisely EVM can be used, stemming from classic conflict between [[accuracy and precision]], as the mathematics can calculate deceptively far beyond the precision of the measurements of data and the approximation that is the plan [[estimation]]. The limitation on estimation is commonly understood (such as the [[ninety–ninety rule]] in software) but is not visible in any [[margin of error]]. The limitations on measurement are largely a form of [[Discretization error|digitization error]] as EVM measurements ultimately can be no finer than by item, which may be the [[work breakdown structure]] terminal element size, to the scale of reporting period, typically end summary of a month, and by the means of delivery measure. (The delivery measure may be actual deliveries, may include estimates of partial work done at the end of month subject to estimation limits, and typically does not include QC check or risk offsets.)

As traditionally implemented, EVM deals with, and is based in, budget and cost. It has no relationship to the investment value or benefit for which the project has been funded and undertaken. Yet due to the use of the word "value" in the name, this fact is often misunderstood. However, earned value metrics can be used to compute the cost and schedule inputs to [[Devaux's Index of Project Performance]] (the DIPP), which integrates schedule and cost performance with the planned investment value of the project's scope across the [[project management triangle]].<ref>{{cite book
 | last = Devaux
 | first = Stephen A.
 | title = Managing Projects as Investments: Earned Value to Business Value
 | publisher = CRC Press
 | year = 2014
 | pages = 153–156
 | isbn = 978-1-4822-1270-9 }}
</ref>

Darling & Whitty (2019) conducted an [[ethnographic]] study to see how EVM is implemented, applying Goffman's [[Dramaturgy (sociology)]], they found there is a sham act occurring through impressionable acts presenting statistical data as fact even when the data may be worthless.<ref>{{Cite journal|last=Darling|first=E.|date=2019|title=Sham project compliance behaviour: Necessarily masking the reality of project work from senior management|doi=10.1108/IJMPB-05-2019-0118 |s2cid=210487285 |url=https://doi.org/10.1108/IJMPB-05-2019-0118|journal=International Journal of Managing Projects in Business}}</ref> Findings include sham progress reporting can emerge in an environment where senior management's ignorance of project work creates unworkable binds for project staff. Moreover, the sham behaviour succeeds at its objective because senior management are vulnerable to false impressions. This situation raises ethical issues for those involved, and creates an overhead in dealing with the reality of project work. Further the Darling & Whitty study is pertinent as it provides sociological insight to how a [[scientific management]] technique has been implemented.