In my last article titled ‘Assessing Your Software Project Risk with the Software Estimation Process’, I discussed various risk factors which pose a threat to the success of a software project. It also mentioned the emergence of risk factors associated with software estimation as an important factor in the risk ranking list. Let me now discuss how risk assessment can be done, what questions you need to answer for identifying specific risks to your project and how you would assess them and prioritize them via comparative analysis of historical estimates of tasks in the present context using a standardized estimation tool.
Identifying Similarity In Risks As Per Past Project History
As a project manager, you can study and assess all major categories of risks. Since every project is different from the other, two projects might not be vulnerable to the same types of risks. Review of past project history, evaluating a similar past project and conducting discussions with your project estimation team can help you to identify probable project specific risks.
Risk Related Questions you Should Ponder Over
Pondering over the different categories of problems and risks which showed up in a past (similar) project can help you to identify those risks which are both apparent and hidden. You need to judge and expect the risks which might not be apparent in the first phase but would show up in later stages. Here estimates should be compared and contrasted via last project estimates.
Questions you need to focus upon should be the following –
1.Are my scheduled estimates at risk? A past analysis can enable studying causes of misjudged schedule estimation, frequency of error, impact and associated risks – and effect on cost estimation.
2.What are my problem areas? This would relate to issues faced in the past, how estimates had gone awry due to problem areas. What were the resolution techniques? Had those techniques been useful? Here the degree of variance in effort estimation and its impact can be studied. Present sources of problem areas and risk associated with them can be dealt with greater certainty.
3.Does my team have the expertise and competency to handle this project? Resource analysis would evaluate risk pertaining to incorrect effort estimation. Resource allocation for complex tasks needs to be different. You need to question the credibility of your effort model for determining effort estimation and if the past effort estimates based on it are dependable or not? If the past effort estimates are not dependable, then the estimates for time and cost might not be accurate as you will have to hire costly experts and eventually require more time to handle projects which are more complex and technical in nature as compared to your past experience. The greater the project task complexity, the greater the time to complete it and the greater the chance of the estimate going awry – take special note of this.
4.Do I have a past record of doing estimates which have been dependable?
5.Do I estimate manually with a calculation intensive risk prone model or use an estimation tool with quick estimation and high accuracy enabled capability? Note that using estimation software to do quick estimates enable enhanced flexibility and a high degree of accuracy. If you are bidding for multiple projects, you are supposed to estimate aggressively and accurately for each of them. Manually estimating many projects together is risky and vulnerable to inaccuracy. The only viable and riskless option is using estimation software like Quick FPA to do quick and accurate estimates.
Measuring Your Risks and Setting Standards for Inaccuracy
Answering the above questions will give a broad dimension of the probable risks your project faces. Identification leads to measurement and assigning values of importance to the risk factors to enable resolution of risks in order of priority. Probability of occurrence of the risks and its impact level should be ascertained. A threshold limit of every risk item should be set to determine when (at what stage) to begin resolution/corrective actions. Estimation inaccuracy should be well defined in your organization. This varies with how vulnerable your organization is with respect to allowing risks due to estimation error in cost, schedule and effort. This also varies with project nature.
If you are a small organization handling small projects, you should not allow high estimation errors as this might erode your brand name and may cause software project failure too. You have to be aggressive in your estimates and also maintain consistency of accuracy. If your are a larger organization, you can still allow some degree of estimation inaccuracy if you handle complex projects with well established clients over long term contracts, however focus should be to avoid them. This necessitates establishing permissible levels of estimation error (cost overrun and delay) for each organization, which does not affect its bottom-line and translates to acceptable profit limit too.
How Estimation Software be Used to Define Risk Factors?
Once you have set a standard for your organization, use an estimation tool to estimate for a current project. I am strongly emphasizing using an estimation tool since it is a standardized mechanism based on a model to estimate tasks. If you are bidding for more than one project do it individually for all the projects. Compare and contrast your past historical manual estimates against your present estimates through an automatic tool. Note down the variance in percentage and study the impact of a delay/cost overrun due to that estimation error. If the effect/impact was beyond your defined permissible limit, term that as a risk factor. Various issues, tasks, and risk causing factors can be analyzed with respect to their estimation error and categorized as acceptable/avoidable risk.
Rating the risk for probability of occurrence, creating a scale of probability and impact and estimating the impact of occurrence with respect to accuracy of estimation would enable the project manager and the team to have a contingency plan in place. Risk impact should be measured in terms of nature of the risk factor, severity of the risk distribution and time period of impact.  So if a particular task ‘A’ was estimated for 10 hours instead of 15, the risk impact of underestimation should be measured with respect to its effect on overall sprint, pressure on developers and its ramification on other tasks, whether such pressure lowered effort efficiency and to what extent etc. This can be later applied to all tasks which show variation is estimates (through estimation tool) with respect to past estimates (done manually). The issues at various stages of project management can be evaluated to develop probable risk charts with impact in association with estimation inaccuracy.
Risk monitoring and control measures can be planned and implemented from this stage onwards.
Register at this site to get an invite for accessing Quick FPA – our estimation software which enables limiting risk arising out of inaccuracy of estimation. Use the tool to do cost estimation through function point analysis, which allows incorporating cost adjustment and reuse factor for overall estimation of cost, schedule and effort estimation.
“Software Risk Analysis and Software Risk Analysis and Management.” 18 March 2011. LANE DEPARTMENT OF COMPUTER SCIENCE AND ELECTRICAL ENGINEERING – West Virginia University. 22 December 2015 <http://www.csee.wvu.edu/~cukic/CS430/Risk.pdf>.