Just like the practice of project management, Project management tools have also evolved considerably over the years. There used to be a time when project managers working on spreadsheet was considered cutting edge. Slowly we started to see desktop based tools gain acceptance amongst the project management community. Though these tools were a definite improvement over spreadsheet, they were found lacking in a lot of areas. The very fact that they were desktop based meant that usually only one person could work on it at a time. Communication among members was still through the email.
Slowly with the advent of internet and high bandwidth these project management tools came online and were providing project management software as service (SAAS). These tools along with providing the ability to collaborate seamlessly have also given users a way to communicate effectively. Overall it can be said these tools have led to increase in productivity.
We are now slowly approaching an era of the third generation of project management tools. With the advent of big data, use of data science and artificial intelligence concepts are being applied to help ensure optimum execution across various PMBOK elements. Given the nature of artificial intelligence and the huge amount of data that is being created in projects it is becoming increasingly easier for machines to extract relevant information and use it for managing the , projects efficiently. Machines can do things quicker, less expensively and more consistently. Many project management steps which are repetitive/sequential tasks (especially those performed during the execution phases of projects) can easily be carried out by machines and will eventually be replaced by machines.
If you look back, most of the early generation tools were focused only on project planning. While planning is an essential part of project management, it soon became evident that lack of focus on execution meant that teams were spending months on planning and when they came across issues during execution they were none wiser. Online project management tools made it easier for users to collaborate, store documents and share communication effectively amongst project group. While all these attributes have made project management easier the problem of optimum execution was still waiting to be cracked.
Clever systems are being currently designed which will not only, extract greater amounts of information from the project environment than traditional approaches but will also and then leverage various mathematical techniques to make probabilistic judgments. One of the key areas that we see a lot of problems during execution is in the project risk management space.
Very often, when you see the weekly status reports of project managers you will see green status light on most aspects of the project. If the project goes down and becomes a mess only then the status light changes to any other color. The status report which was supposed to give the stakeholders leading information, actually lags behind what is commonly known by now in the team. Project risk management systems till today have not yet addressed this concern. However using data science we can tackle this problem. The objective of any project risk management system should be to provide project and portfolio managers with the earliest possible insights and awareness to emerging risks while minimizing false alarms. What kind of system the manager uses to interpret these risks (a good system should have a high ratio of true positives to false positives) is also going to determine whether the team spends time on solving actual problems or on falsely perceived risks.
On a basic level, Automated Project Risk Management system can be broken down into 4 steps
1. Risk Identification:
This is all about capturing risks. Here smart systems depend more on the Voice of the Team rather than just the project managers. This helps to give an accurate picture of the project. Often times there are various contrasting views expressed by the members of the same team. Sometimes a few members can bring your attention to small, subtle change that can give rise to major issue down the line. In effect this way of capturing information is much better than depending upon one person’s views and opinions. The questions asked in this Voice of the team questionnaire are also based on data of thousands of projects which have been executed so far. These are carefully selected based upon prior project performance and answers to these help get maximum information from the project. The system can also use the logs of text based communication between members to catch risks early.
2. Risk Measurement
Next step is basically what can be done once we know the responses from the team. Using predictive analytics we can easily find out the probability of completion of a milestone but also point out if there will be any slippage. But this is just the tip of the iceberg.
In our tool projectrimms.com we use various statistical analysis to compare particular risk with historical thresholds. This helps prioritize the risks and brings the attention of the team specifically where it is needed in a more timely manner.
3. Risk Mitigation
Project risks can be understood better through visualizations and graphs. These can be again categorized based on their severity, likelihood etc. These graphs can then become a part of weekly status reports sent by the project managers to stakeholders. Along with this status report , which is outputted by the system, automatic email notification can be sent to team members depending upon the where the risks are compared to the threshold values. This kind of early warning system enables the team to engage in serious discussions as to how to minimize the risk should it actually materialize.
4. Risk Removal
This final step entails using business rules to identify which categorize risks which have been successfully mitigated and those which haven’t are further escalated using these business rules to the appropriate team member/s. This enables project teams to take risk management to the next level. By going beyond simple risk mitigations the teams do everything possible to reduce the probability of risk occurrence and to reduce the impact the risk should it actually occur.
In our tool projectrimms.com we have combined this new approach to project risk management with Bayesian algorithm to make probabilistic judgments and classifications. The algorithm is based on years of project data and is fine tuned to pick up any incoming disturbances early. The techniques we have employed sift through the voice of team data (i.e. the wisdom of the crowd) to prioritize areas of concern for the project manager. Our proprietary algorithm then tries to set a framework of how these risks will affect on various aspects of the project quality management, cost management, schedule management etc. So in effect the earlier weekly status report which was more or less just ‘record keeping from the project manager is now generated through a proper process by the system itself using algorithms which are more likely to paint the right picture than a subjective opinion of a person based on guess work or hunch.
Eventually you will see systems which will start tracking their own data and will thus be always in a audit ready mode. This will be greatly beneficial for the sponsors and higher management to see how the changes in system have induced changes in the behavior of the team and vice versa.
This is just the beginning of an interesting time in an area which will probably see great changes happen fairly soon.