Software Cost Estimation

I had expected to present my webinar,  “Best Practices for Cost Effectiveness Studies using TruePlanning” in early August. As you might know, I was planning to show a real world example from a recent engagement with a government customer. Unfortunately, since the Source Selection has not concluded with a downselect, I was not able to obtain the public release in time. However, for this month’s blog I will continue share some of the highlights of the webinar.

In last month’s blog we explored the uses of TruePlanning during Source Selection from the Supplier’s (or Contractor’s) perspective. This month, I would like to share with you some of the uses of TruePlanning cost estimating software from the Customer’s (or Government’s) perspective:

• Analysis of Alternatives (AoA) – Is proposed technical baseline cost-effective against other competing alternatives?
• Cost Realism – Are the performers bidding within an accurate range
• Data Driven Estimating – Are the performers bidding based on appropriate, traceable historical data points.
• Independent Cost Estimate (ICE) – Using the performer’s technical configuration, what does a completely independent look say about the performer’s bid?
• Risk Analysis – Is our bid over conservative, how much risk are we willing to take and how much cost exposure can we absorb?
  
• Schedule Estimating – Can we really do the job within the schedule constraints?
• Growth Estimating – What other configurations, materials or technologies might we consider?

While TruePlanning is useful in estimating project costs and all of the above types of analysis, I have found it most useful in developing AoAs. During Source Selection, TruePlanning provides a structured, repeatable framework that can rapidly develop cost estimates across a range of alternatives. These results are often incorporated directly into Operational Effectiveness models along with Measure of Effectiveness and Measures of Performance. In some cases, clients using tools such as Phoenix Model Center can directly link TruePlanning with optimization tools as well. 

Over the past several weeks several users have inquired about the best way to estimate costs associated with porting existing software to a new hardware environment. Normally for this situation some of the existing software will require some amount of adaptation to operate on a new server. However, a large portion of the existing software will only require integration into the new environment.  

Estimating software costs associated with the above will require the use of several cost objects:

- Systems cost object if program management, Quality Assurance, configuration, and    documentation costs are to be included in the estimate

- Assembly cost object to estimate integration costs

- Software component cost object to estimate software adaptation costs

- Software COTS cost object to model all existing software requiring only integration effort 

The main input parameter that should be reviewed with care in the system cost object is project complexity. The default value for this input is nominal. Recommend the low value be reviewed since the team size will be small.

The main input parameter that should be reviewed with care in the assembly cost object is system complexity. If the software is just be ported to the new environment, recommend a low value be selected, since the requirements definition system design has been completed previously. 

The software component input parameters will be treated in a normal fashion. All adapted software will be modeled as adapted new design, code and test. 

The main input parameters that should be reviewed with care in the software COTS cost object are software size, integration team maturity, and external integration. This cost object should be used for all software that will only be integrated in the new environment. Functional size should be used if the size is unknown. The input parameters for integration maturity and external integration must be reduced. The recommended value is 0.05, which is the lowest value accepted by the model. This value will account for much lower integration activity for the existing software.

If you have any further questions, please contact a PRICE Systems Support.

Over the past several weeks several users have inquired about the best way to model software code translation when estimating software costs. Software code translation consists converting existing legacy software source code that was developed in a specific high order language (such as Ada or C) and translating it to a new high order language (such as C++). The translation does not add new functionality. For example, translating Ada code to C++, simply results in C++ code. The major benefit of the translation is maintenance. Ada experts are not needed to maintain C++ software for this example. The code and unit test activity is the major cost driver for a software translation.

Modeling this type of code translation is very straight-forward in the TruePlanning for Software model. Two input parameters should be evaluated for modeling this scenario: Auto-Translated Code Size and Auto-Translation Tool Efficiency. The translated code size should be entered for Auto-Translated code size. However, the translator may or may not be very efficient. This is the purpose of the translated tool efficiency input parameter. A drop-down dialog box is available.. The range is from “Very Low” to “Very High.” The efficiency of the translator should be selected from this dialog box. 

The purpose of the dialog box is to account for additional cost or effort due to a poor translator. For example, if the code translation size entered is 10,000 and the translation tool efficiency is Nominal or 80%, then the model will assume that 20% or 2,000 SLOC (source lines of code) will have to be manually translated into the new HOL. 

If you have additional questions concerning software code translation, please contact one of our PRICE Cost Research Analysts.

Jim Otte
Solutions Consultant, PRICE Systems
 

While sitting in the operatory chair yesterday, my dentist said something that made me stop. He was complaining about an increasing rate of incompetence and apathy he observes in those delivering services to him. And while I do agree with him in principal, he and I are of the age where some folks label us as grumpy old men. So, it may not be as bad as we think. Regardless, the statement he said he made to the an unfortunate poor-quality service provider was, “If you don’t have the time to do it right the first time, when are you going to have time to fix what you did?” Apparently, the excuse offered by the provider was that he didn’t have the time to do the job the correct way, so he tried a short-cut.

Wow! This hit me right between the eyes. I immediately expelled all the material and apparatus he had placed in my mouth so I could tell him how impressed I was with his logic. Plus, I had never heard that expression before. But how true it is and in it I see a glimmer of hope for realizing a key initiative of WSARA (Weapons Systems Acquisition Reform Act of 2009) - getting things right from the start with sound systems engineering, cost-estimating. The path to getting this accomplished is through the people doing the work and not through DoD leadership. As I reflect on my experience with DoD Systems Engineers and Cost Estimators, most I know want to do the job right the first time. If every one of us involved in planning and estimating software, hardware, and IT projects pledges to never take short-cuts, we can make the difference.

There are two reasons for why I don’t think this is just rhetoric: 1. getting things right from the start can’t happen if we are not committed to it; 2. I am unable to conjure up a single example of anything I have experienced in any dimension of my life that started out wrong and turned out right.

I recently had the opportunity to work directly for one of our clients on a high visibility, must-win proposal. The contractor was just about ready to commit to the bid number, but wanted to know the likely bids of the other two performers. We were asked to do a “Ghosting the Competition” study where we ethically collect open source data on two competing designs and combined with engineering technical data to develop a best cost estimate of the competitor’s bid positions.

Unfortunately, not much intelligence was known about the competing configurations, but the engineers recently noticed the other companies displaying their latest technology at a tradeshow event. Based on that information, plus some published marketing data, the engineers arrived at reasonable technology configurations and weight statements. We also established a cost baseline for the client by calibrating past data and using the client’s direct rates and overheads.  Using the TruePlanning estimation software we were able to complete all of the cost estimates in about ten hours including detailed reporting down to the resource and activity level.

Once we were done and happy with the results we were asked to brief the company President. Taking one glance at our estimates, he remarked that what he really needed was an Independent Cost Estimate or ICE in addition to our Ghosting study. The purpose of the ICE was to determine if they were really offering the best value possible to the government…and not just the lowest bid. He requested the ICE within four hours. This caused great concern for the engineers and finance as they had never done an ICE in this short amount of time.

However, using the same cost estimating tools and models (TruePlanning) we built for the Ghosting the Competition study, we were able to quickly and easily generate the ICE in less than four hours. This was done by removing the calibrated inputs for manufacturing and electronics complexity and replacing them with values from the PRICE knowledgebase. We also reverted back to the industry average rates contained in the worksheet sets.

Finally, we completed a producibility study on process improvement using the Manufacturing Process Index input. Using TruePlanning we quickly generated all of the reporting formats at the activity/resource level and export the results to EXCEL, WORD and Powerpoint.

The bottom line was that the versatility of the TruePlanning tool allowed the client to quickly and efficiently develop several types of analysis required for a must-win proposal with minimal impact on staff and in a very short amount of time.

Like titanium and other exotic metal-materials, “composites” (by definition, combinations of materials) offer significant weight-savings and reduced part counts, but at a price of high production cost. Sound contrarian to our parametric cost estimating view?   Not really. Complexity of manufacture is quite higher. Likewise process index and structural tooling values grow. Plus, design lead times drive developmental cycles.

That said, understand that composites represent more than a material type. They can involve a highly labor-intensive approach to preparing, braiding/ winding, molding, bonding and modular assemblage. Yes, some aspects of braiding and molding lend themselves to automation—which then drives tooling investment. Composite development offers design flexibility, weight savings as well as advantages in long-term deterioration. But not all pre-curing processes are the same, to include recent advances in structural co-processing before subsystem cure.

At this juncture, rather than get ON a soap box, I’d ask that you join me in help getting us INTO a soap box. 40+ (!) years ago, my engineering Dad challenged my brother & I to make a soapbox derby car from fiberglass. (My uncle worked at little known Owens-Corning at the time, and material costs were cheap!) To make a long story short, Dad required that we not get “gluey” too early and instead had his two very young sons learn the benefits of (pre-CAD/CAM) design drawing. So draw, sketch, and describe we did. Talk about information entropy! But the more we had to draw and detail a component’s design with the physical and functional features Dad needed to make it exactly right, the longer (more time) it took him with each piece… and the more our “cost” (waiting anxiously to ride) went up.

25 years later, MIT’s Hoult & Muter would have been proud: we realized that amount of information communicated was the driver in our composite manufacturing process. Multi-dimensions are one thing to grasp at a young age. Communicating corresponding tolerances is a bigger challenge. The latter are typically known as “feature parameters” in engineering circles. Suffice then that to estimate composites processing, the more effective predictor of cost is entropy between design and build.  

And how would we propose to count all relevant exchanges of information, including these latter parameters? The same way perhaps that parametricians characterize early stage software concepts visualizing inputs, outputs, data stores, elements (toleranced dimensions), operators (processes), etc. Just like using Function Points in software cost estimation! Over the next few months, we’ll examine this new approach following more composite cost research and predictive modeling using an information entropy statistic. Stay Tuned!

John Swaren
Solutions Consultant, PRICE Systems
 

Parametric modeling is excellent for all aspects of early-concept cost estimation, including go/no-go decisions downstream. So, in the spirit of bringing a transparency to (ethical) financial engineering…
why not apply our craft to pricing “real-options”?

The latter are essentially strategic opportunities for engaging resources (cost/schedule) into projects, ventures, investments, or even abandonments. The opportunity choice has value itself! 

Unlike static project Net Present Value (often, but not exclusively, approximated with Discounted Cash Flow) assuming pre-defined decisions, real-options reflect the merit of flexibility. If an R&D or proof-of-concept presents viability/ marketability learning, the option has positive value, above and beyond DCF. The more the flexibility, the higher the value. Likewise, a real-option appreciates with more uncertainty.

By now, you’re asking—“Wasn’t this a parametrics blog? I’m an engineering/ computing/ math/ science type, not a quantitative-finance geek. How could the above possibly help me any”?

Answer: In some situations, specifically go/no-go, the value of your flexibility created with the strategic choice to move forward (or not) can exceed its “option” cost. Not all options should be executed, just as all go/no-go decisions aren’t go’s. But, over time, continuing to pay less than their market value creates an opportunity to average out with total economic-value creation.

But, you say— “How do I find the cost of this flexibility/uncertainty option? It sounds great that I can make investment decisions based on buying/ executing (or not) these options, but do I really need to learn fancy finance stuff like Black-Scholes, Value Trees, Binomial-Risk Neutral Pricing… based on risk-free rates of return and (expected) discounted cashflows…. Yikes!”

Answer: (& bottom-line, for now) No. Use your parametric estimating tool! Concerned about the hardware cost of pilot-production/ tooling? “Buy” an option priced as the cost of preliminary design. 
{Note that the latter cost is your option’s premium, and the go-ahead cost is your option’s strike price.}
Interested in the nonrecurring cost of large-scale full software development? Buy an option for the
cost of first iteration increment. Concerned about COTS versus assembly? Estimate the development (and integration) of both scenarios. 

The point is take an economically-disciplined approach to valuing your strategic choices downstream. Parametric modeling works here and is “data-driven” defensible.  It is certainly applicable to strategic investment, capital -budgeting and new business decisions within both the public and private sectors. Transparency through mathematics is a good thing.

Now that we know the background on the original concept of TRL's (Technology Readiness Levels - reference Arlene Minkiewicz's earlier blog post), we now want to address estimating costs associated with different TRL levels. It is important to realize that a model cannot estimate TRL costs by simply changing an input parameter. Rather the only way to estimate costs associated with different TRL levels is to model the scenario. For example, if you are estimating costs for TRL level 2 phase the input parameters would be very different than if estimating costs for TRL level 6. 

PRICE Systems has prepared a white paper, along with web presentation on just this topic. The web presentation can be found on PRICE University, and is available to all clients. If interested in the white paper, please contact Amanda Rivera, at 856 608-7201 or send an email to info@pricesystems.com with "TRL Paper" as the subject.  

Jim Otte
Solutions Architect, PRICE Systems

Currently we are exploring the best approach to including a more comprehensive cost estimate for Total Ownership Costs (TOC) into TruePlanning. The current version of the software has focused on development and production costs with some life cycle costing including. The life cycle costs included are focused on the system specific O&S costs such as initial spares for priming the supply pipeline, maintenance, replenishment spares, etc. It is a system view as opposed to a program view of TOC.

As we better understand the need to conduct affordability studies it has become clear that design decisions which can be assessed in regards to their total ownership cost will result in more affordable systems. In order to do this we need to include a greater range of cost elements such as construction of system support facilities, O&M of infrastructure and system facilities, some personnel cost which are enterprise level not necessarily systems specific and so forth. What is not clear is how these elements differ from different customer perspectives.

For example, our current lifecycle costs have an original equipment manufacturer (OEM) flavor to it. Whereas I speculate that a PEO, PM, or service staff analyst might have a need for a broader set of elements. Feedback in this regard from our customers would be very beneficial.

Bob Koury
Senior Cost Research Analyst, PRICE Systems

One of the great features of the TruePlanning cost management software is the fact that it makes it easy to handle complications of inflation and estimating projects performed in different countries and currencies. The costs associated with doing work in different countries, and the relative value of different currencies is constantly changing. To address this, the cost research team at PRICE does an annual economic update performed by the cost research team, and this blog will introduce some of basic concepts and research that goes into maintaining this feature every year.

The price of goods and services changes over time, and this value is measured by inflation rates (if prices have gone up) or deflation rates (if prices have gone down). These inflation/deflation rates are constantly updated for many different currencies by the International Monetary Fund (IMF). In TruePlanning, we use past and predicted inflation rates published by the IMF  to determine the cost in today’s currency of doing work in the future, for whatever currency you want to use. Within the TruePlanning tool, these factors can be applied to help produce a cost estimate of work that takes place over a long period of time, with the cost estimate shown in today’s currency values.

In addition to inflation/deflation rates for different currencies, we also recognize that the price of goods and services may be different from one country to the next (i.e. a U.S. Dollar exchanged and spent in India will buy more haircuts than a dollar spent in the United States). The degree of these price differences can be measured by looking at Purchasing Power Parity (PPP) data of a country. Standard PPP values are calculated or compiled and published by the Organization for Economic Co-operation and Development (OECD). The PPP rates published by the OECD are used in the TruePlanning tool to estimate costs of performing projects in different countries, as the cost of doing the same amount of work in different countries can vary significantly.

By automating the application of these and other economic complications of cost estimating, the TruePlanning tool helps create a process in which the effects of these economic factors can be applied consistently and correctly for all your cost estimates.

Last month I blogged about the importance of cost realism, its roots and how as estimators we must always reflect the truth, no matter how unpopular. This month I want to share with you a recent experience on a Source Selection. As part of the Source Selection team, my role was to conduct a Cost Realism estimate on each of the performers submitting bids. I want to share with you a few insights from that experience.

One of the first rules I always follow is to never ask engineers to provide data that you can readily find. I rarely will use parametric data forms, as you will get to be unpopular fast with those you task with the honor of filling them out. Rather, I dig into the technical and cost volumes to derive the configuration, technology, weight statement, rates/overheads and other juicy information to populate TruePlanning.

Once I have the TruePlanning cost management software populated with all of the data I can derive from existing documents, I will then ask the subject matter experts in each area to discuss the other qualitative/quantitative factors about each performer. I can usually guide this conversation to derive inputs such as requirements stability, engineering complexity, integration and others. I find it effective to hold this meeting remotely using Go-To-Meeting so everyone can see my desktop displaying the TruePlanning model and no one has to leave their desk. When inputs are critical cost drivers, I usually pop up the wizards or generators so the engineers can see the choices available.

Once the model is fully populated and all inputs agreed to, I will then produce a very well documented estimate with all assumptions as the final deliverable. In this particular case the Source Selection team was very pleased with the result and asked if we could extrapolate a new configuration based on some technical changes proposed by the performer.

Using our well qualified TruePlanning model, we ran the extrapolation and reported to the Source Selection team that we expect to see a large impact. Initially, no one believed the results; it could not be so they said! However, several weeks later, when the performer’s new estimate arrived, we were within 1% of the revised bid. During this time period, the Source Selection team was able to successfully prepare a negotiating stance based on the expected bid coming in as predicted.

In all my years as an estimator, I have only seen estimates go up as more is known about the technical configuration. Bottom line, as estimators we are the “front line” for telling management the hard to hear truth. In this case, at least they were prepared!

Earlier this week I conducted a webinar intended to make PRICE users aware of the Cost Research Services available to them as part of the license fee they pay to use PRICE products. I thought I would recap the highlights of this webinar for those of you who might have missed it.

At PRICE we understand that cost estimating tools, while useful and valuable, do not always present the complete solution. Every single cost estimation projects presents new and unique challenges.  We think it's important that in addition to solid, time trusted cost estimating models, our users have access to the many years of experience we have as seasoned costestimators, subject matter experts and operations researchers.

This service is nothing new.  For the 30 years that PRICE has been in existence, we have worked as partners with ourcustomers to optimize their use of our models and methodologies.  This was just an opportunity to formalize the offerings and remind the community what services are available.

So what does the Cost Research team at PRICE have to offer the cost estimating commmunity?  On average our researchers have more than 24 years of experience with hardware estimating, software estimating, operations research or some combination of the three. We are constantly engaged in cost research projects addressing market needs as they arise.  The results of these studies vary depending on the need they attempt to address.  In some cases, data collection indicates custom models should be developed.  These can be developed and deployed in TruePlanning, the flexible cost estimating framework.  Some results are published as updates to tables or calculators in the PRICE Software and Hardware cost estimation models.  White papers, webinars and the PRICE blog are all means we use to communicate the results of cost research studies.  

PRICE's Cost Research Team is available 24/7 to address users cost estimating question on an as needed basis.  Some issues require more attention than a single phone call.  Users are encouraged to schedule working sessions with one or more of our analysts to take a deeper dive into cost estimating issues that perplex or intrique them.

Some areas the team is currently studying include Total Ownership Costs, Joint Confidence Level, Performanced based models for technologies such as FPGAs and ASICs, semi-rigid cables, and  Operations and Support costs for space systems. 

The most important thing the cost research team at PRICE wants to do is make our clients better estimators while adding value to the cost estimating community as a whole.  We can't do this without input from our clients.  Share your cost estimating challenges with us.  Call, email or comment on our blog.  856-608-7201 or info@pricesystems.com


 

From my perspective as a cost researcher, the calibration tool is one of the most powerful analysis capabilities built into the TruePlanning cost management software . One way I can use this tool is to go back to an old estimate for a project that is now completed, and analyze the correctness of the previously entered input values. With this analysis, I can find ways to improve our methods of soliciting input values from the user to ensure the best values are entered the next time. This way, the TruePlanning models keep getting “smarter” as new information becomes available.

TruePlanning models calculate a cost estimate based on a set of inputs about the project being estimated. After the project is completed and we have actual cost data available to us, we can go back to TruePlanning and explore the various reasons why an estimated cost may have been different from the actual cost.

Let’s say we are estimating a project involving a brand new electronic technology, and we are uncertain of the values we selected to describe its complexity. The calibration tool allows us to enter the actual costs of the project, and reverse-engineer the value of an input we were uncertain about (i.e. manufacturing complexity for electronics) that would lead to the correct answer. This allows us to identify incorrect input values so we can identify problems, then review and refine our methods used to pick the value in the first place. 

Over the years, we have used this (and many other) TruePlanning analysis capabilities to further our understanding of our cost estimates and our cost estimating methods. There are many analysis tools built into TruePlanning, and many ways to slice and dice information about a cost estimate. This makes it quick and easy to standardize the cost estimation methods used, and continually review and refine our methods to get the most accurate cost estimate possible. With more analysis tools and more information becoming available all the time, TruePlanning will keep getting smarter.

I am excited about the efforts described in Derek Kaufman’s article on the AFMC website.  For the complete story please see this link http://www.afmc.af.mil/news/story.asp?id=123202181. The DoD is investing in the rapid build up of a new foundation of acquisition workers focused on estimating costs and negotiating prices through continuous learning. 

Mr. Assad notes that Wright-Patterson Air Force Base, Ohio, was once the preeminent DoD recognized leader at estimating costs and negotiating prices with defense contractors. It's a skill that has been allowed to atrophy, he noted.   "We need to build that pricing capability back up," he said. "It's extremely important for us to have world-class pricing talent, so that we make sure that we get a reasonable deal for our taxpayers."

For over 30 years, PRICE Systems has passionately supported this same vision.

Every year we train hundreds of analysts and acquisition professional across the globe. We proudly offer training on topics such as source selection support, software cost estimating fundamentals, data collection and total ownership costs.  And recently we introduced an on-line acquisition training center, known as PRICE University, where students can gain continuous learning credits through self paced study, right from their own desktops!

If you haven’t had the chance to hone your skills as Mr. Assad suggests, I invite you to start today by taking a quick tour of PRICE University - http://university.pricesystems.com/index.jsp

I have been frequently asked to do crosschecks on other people’s software cost estimates which are potentially done in a variety of tools from spreadsheets to SLIM.

One of the common operator errors I see from other users is not understanding what activities and resources are included in the outputs of the particular tool that they are estimating with.

This is akin to deciding between two cars and not knowing if both come with the same sets of features (stereo, AC, heated seats).   With software estimation tools you need to know what work is getting estimated by the tool. (Requirements, Design, Code, Test, Documentation, System Test) and what resources are included in the effort estimates to accomplish those task/activities. (Programmers, Analysts, Architects, Systems Engineers, Testers, Tech Writers, Project Managers, QA and CM)

A common mistake I see is generating an estimate with a tool then adding in Program Management effort and cost when the tool utilized already includes those activities and resources in the estimate.

That’s why I drive a pearl black TruePlanning 2 door coupe.

With so many acquisition programs over budget and behind schedule, the term “Cost Realism” is suddenly very popular. In my experience as an estimator on many major acquisition programs, two things have remained certain over years (besides death and taxes). First, the probability of the program ever achieving the original cost estimate is exactly zero and second, the more information that is known about a program, the more it will exceed its original cost estimate.   

With that said, the move to Cost Realism is so important because it recognizes these two fundamental facts. Much of the interest in Cost Realism is driven by the Weapon Systems Acquisition Reform Act of 2009. According to US Senator Carl Levin “Report after report has indicated that the key to successful acquisition programs is getting things right from the start with sound systems engineering, cost estimating, and developmental testing early in the program cycle. The WSARA also calls for a “Director of Independent Cost Assessment to ensure that cost estimates for major defense acquisition programs are fair, reliable, and unbiased.”

If we go a bit further and look at definition of Cost Realism in the Parametric Cost Estimating Handbook we find  that “no one expects a cost estimate to precisely predict what a hardware or software project costs or a time and material service will cost. So, cost realism is not about the exact cost estimate. It's about the system of logic, the assumptions about the future, and the reasonableness of the historical basis of the estimate. That is, it's about the things that make up the foundation of the estimate.”

So while we recognize that at its core cost estimating can never be exact, cost realism seeks to ensure estimates are closely tied to the fundamental realities of the program as we know them at the time, without regard to politics or undue optimism.   As I finish this blog, I’m struck by the line in a Few Good Men “You can’t handle the truth!” As estimators, we need to always reflect the truth, and that is what cost realism is all about.

The newest version of TruePlanning has been released and distributed to customers. The new features were designed to make it easier to estimate entire systems, not just individual components  or sub-systems. TruePlanning is an excellent solution in this regards.

Systems that require you to estimate software costs, hardware costs, and the integration of multiple pieces of each can all be done in one framework. There are  specific features in this release, like input inheritance, that allow you to make changes at the top Project level which then flow throughout the entire system.  This saves an estimator or an engineer lots of time as typically they have to dedicate time to making sure that there is consistency from the top level of a systems throughout.

Other enhancements include improved help and guidance. While using TruePlanning’s built-in calculators to come up with good inputs, you are also informed through pop-up messages and alerts  about what inputs need to be entered or should be entered based on what you have already told the software. We’ve added more charts and graphs that give you visual clarity of how specific inputs of the models effect certain estimated costs. There is also a new Help System that is easy to navigate so you can  locate information that can help answer questions you may have as you are building your estimate.

For more information visit our New Release Info page.  

It’s a common practice to measure failure or success of a project based on the initial functionality requirements and initial cost and schedule estimated.

The Standish Group publishes its Chaos report for software projects which terms a project as a "Success" if it is completed on time, on budget, and satisfying all the initial requirements. Projects are deemed a "Challenged" if functionality is achieved but cost and schedule over runs occur and "Failed" if a project is cancelled while in execution.

However there are other factors e.g. Tom DeMarco’s  Estimation Quality Factor  and Boehm’s Cone of Uncertainty (COU) or combination of these two (a brief analysis of these two factors can be found in the Jan/Feb IEEE Software magazine The Rise and Fall of the Chaos Report Figures: http://www.computer.org/portal/web/csdl/doi/10.1109/MS.2009.154).  

The questions is, does your company measure project success/failure based on initial software estimation? If yes , what methods are used at your company to establish the matrix (Standish, EQF, COU, any other)? In your opinion what is the value of software estimation accuracy?

Failed software projects are always bad but there are additional complications when there is a contract in place to deliver the software.  Disputes over failed software can result in costly litigation that generally damages both the vendor and the buyer. According to observations of Capers Jones in "Conflict and Litigation Between Software Clients and Developers" (2007) , 5% of the projects his clients were involved in either had litigation pending or were currently involved in litigation over project failures.  His findings indicate that it is very large projects, over 10,000 Function Points that are most prone to litigation, which makes sense if you consider the relative increased complexity, as well as the large amount of money such a contract would involve.  Although one occasionally hears of pending litigation over software projects in the news, determining the background and outcomes of such litigation is difficult because they involve large customers and software contracting organizations, both of whom would rather keep their failures out of the spotlight.  This is likely a significant contributor to the fact that such litigations still occur since the industry is denied opportunities to examine and learn from the mistakes of others.

A report in Computer World [6] details a law suit filed by health insurer Highmark against KPGM for malpractice, fraud and brief of contract for a software system that Highmark had outsourced to KPMG.  While the article does not detail the outcome of this lawsuit, it makes it clear that the disagreement stemmed from misunderstandings and the failure of the two organizations to communicate and collaborate.

In light of this I have compiled a list of four responsible actions that both bidder and buyer should engage in when software projects are outsourced.  These may seem like common sense and the same old story, but clearly they bear repeating......

1. Create and Communicate Good Requirements ... misunderstood, constantly changing and new requirements are a common source of angst in software development projects.  Bidders need to understand the buyers requirements, document them fully, and incorporate the risk of requirements growth into their plans
2. Create and Maintain a credible cost estimate - the buyer needs to know what the effort should cost to evalute bids.  The bidder needs to properly assess the cost to them in order to make the project profitable and to avoid having the project cost them money
3. Understand and document risks; Plan up front for risk mitigation - bidders should work with buyers to understand their uncertainties.  These uncertainties need to be incorporated into the estimate and documented in the contract.
4. Create and maintain a cooperative relationship... when bidders and buyers collaborate and work nicely together, bumps in the road and set backs are handled sensibly, reducing the possiblities that things get out of hand.