On October 4 the IT Cost Management Summit took place in Haarlem, The Netherlands. In the tranquil atmosphere of the Philharmonie in Haarlem we have shared our experiences on estimating and managing IT costs. Were you not able to attend the conference? Here is a short recap of what we have learned with links to the presentations and the LinkedIn accounts of the speakers.
Project success is a choice
Joop Schefferlie, president of IPMA, one of the larger organizations for project management, bit the bullet. Many of the changes in our society are realized by projects. Projects represent already 40% of all economic activities and this is expected to rise to more than 50% in the next few years. But an average bachelor student does not receive more than 3 weeks of training on project management within their curriculum. Project success is highly dependent on the leadership capabilities of the person in charge. A good project manager needs to have a servant leadership style to be able to attract the right mix of expertise to deliver the expected results on time and within budget.
The key pitfalls of agile software development
Daniel Saroff from IDC showed that a pessimistic estimate and especially a optimistic estimate have a dramatic effect on the final result of software development. Many agile endeavours do estimate the amount of work, but fail to have a good view on what part of the work is bringing real value. This increases the risk that too big a portion of the budget is spent on activities that do not bring real value. At the end of the budget there is insufficient product developed for a successful business case. For that more budget and time is required. By using targeted metrics and monitoring them regularly negative trends can be discovered early and mitigated before they become a real problem.
Agile Cost Estimation – A return to basic principles
Paul Marston from SPA reminded the audience of the positive side of agile: working in small teams towards concrete results with short feedback loops. This appears to be incompatible with traditional project management that is about the proper execution of a plan. But even when you are agile, you still need to plan and estimate. Also in an agile context we should be able to answer questions on when a product will be ready to release and at what cost. What cost estimators can learn from agile is to let go the idea of creating a complete and exhaustive estimate. Let’s start with a Minimum Viable Product for the cost estimate and then refine it.
The Cost Estimation Body of Knowledge for Software (CEBoK-S)
Megan Jones from ICEAA presented the Cost Estimation Body of Knowledge for Software (CEBoK-S) that the organization intends to release later this year as a separate knowledge domain for cost estimation. Estimation of software related costs has a number of specific elements that are additional to the existing CEBoK. In the course of next year i twill be possible to do an online certification exam. Although it is not mandatory to get the CEBoK certification first, the exam is set up in such a way that experience in the software domain is a prerequisite to be able to pass the exam.
Increasing predictability of large government IT projects
Koos Veefkind from the Dutch Tax and Customs Administration shared his experiences in increasing the predictability of their IT projects. The organization has a portfolio of nearly half a million function points of which is over 10% is in transition (new development, changes, modernization) every year. The Dutch tax legislation is extremely complex, due to the many exceptions that are included to repair negative effects for certain groups of citizens. Also the split of the Tax and Customs Administration into three parts (Tax, Allowance and Customs) means that the IT landscape must be split up as well. This complicates the predictability of changes in the landscape even more difficult. To be able to deliver more predictable, IT projects are reduced in size to ascertain a Minimum Viable Product can be delivered within 12 months. Next to that the likelihood of delivery planning is verified with historical data. Over the 17 technical domains over 3000 datapoints of finished projects is available.
Successful estimating – It’s how you tell the story
Amritpal Agar from The Cost of Everything podcast has interviewed people from all over the world who estimate the most diverse costs, from producing a movie to the development of nuclear powerplants. What emerges from this is that most of these people ended up in their role by accident. No matter how different their backgrounds and whatever they estimate, the similarity is that they are curious about the story behind the numbers. One important message they all have is that the estimate must of course be well-founded, but in order to be effective, you must tell the story to the right person(s) in the right way.
The COSMIC battle between David and Goliath
Paul Hussein from the EU Securities & Markets Authority (ESMA) has been involved with the outsourcing of their IT. ESMA does not own any on-premise IT and maintenance of al their cloud software is 100% outsourced. As an EU agency their IT budget is fixed three years in advance. The ratio between run and change within ESMA is 60:40. When the run cost increase this means that there is less budget available for change. To keep the cost of IT predictable ESMA has outsourced maintenance and development based on catalogue prices. In the earlier contracts ESMA used IFPUG function points as a sizing base for the catalogue prices. For their new cloud-native data platform ESMA is now making use of COSMIC function points as a sizing basis. For different types of projects and applications different categories have been defined, based on the size and complexity. With this type of contract a small agency as ESMA is able to control contracts with large suppliers like Accenture.
Automotive Software Cost Estimation – The UCE approach
Emmanuel Mary from Unison Cost Estimating (UCE) took us to the realm of software cost estimating in the automotive industry. The expectation is that in 2030 50% of the total production cost of a vehicle will be software. Although this is a substantial part of the cost, estimating these costs is still in its infancy. Although never expressed out loud this is most likely due to the widespread myth that these costs cannot be estimated. When automotive manufacturers want to get a grip on the cost of software development, they need to start working on a sizing basis to be able to understand, compare and control these costs. UCE has developed a pattern-based sizing approach to make a quick estimate of the size of the software as a basis to derive the expected cost of development.
Do you want to stay updated on the latest news and events that we organize? Make sure to register for our newsletter and be the first to be informed!