Making the most of it
Written by Wil Cunningham
Most Project Lifecycles and Methodologies explain the types of artefacts an infrastructure project/phase should deliver, but how much of what each project delivers is then re-usable? The normal practice is for each project to be delivered in isolation, the speed of delivery being determined by the project manager’s experience, or their ability to engage with and/or push the support organisation or vendors. They may use previous versions of their own designs, estimates and other artefacts, but this is seldom maximised/measured within installations. The end result is that for two similar projects delivering the same solution, one could take more than 18 months and the other less than a year, depending on the PM.
Establish what you will support today and in the future
If the organisation integrated many companies over the years then it will have inherited every type of infrastructure. The crucial step is to establish which infrastructure is critical for supporting the future direction of the organisation and those that should be phased out. We called our ‘wanted’ infrastructure More Of The Same (MOTS) and the legacy kit we wanted to phase out – Non-MOTS. By having a clear distinction between the past, present and most importantly the future, you can then start to establish the delivery models for each. What is critical is the creation of roadmaps for the technologies so that the future is mapped out and supported by formal standards and governance, new technologies are welcomed as long as they are planned for and not something that has been ‘sold’ to the business as a panacea for all problems.
The delivery factory (Throughput Management)
We knew that the existing model was sub-optimal as the support organisation had no service level agreements (SLAs) for delivery. I created a factory approach. We established eight Centre Of Excellences (COEs) – Plan, Estimate, Design, Capacity, Build, Configure, Test, and Implement. The technical support people were then aligned against the individual factories, and the factories each had an accountable manager who had to establish their delivery SLAs for their artefacts, and tracking mechanisms to ensure that the SLAs were always met. The factories were in addition the central storage libraries for the ‘to be re-used’ artefacts. The delivery SLAs were published so that project managers knew exactly how long infrastructure projects/phases would take – it was no longer subjective and the overall aim was to take at least six months out of the project timescale.
Components, Standardising and Maximising Deliverables
We established the lists of allowable hardware components that were MOTS etc, established their costs and the delivery timescales with our suppliers. During the estimating phase, we were then able to quickly cost and estimate the project. A pre-commence project would receive a high level design and estimate within days to allow the business to either ‘kill’ and/or agree to launch. Establishing re-usable templates (supported by standards), for designs, estimates, build documents, was a real challenge. New technologies took naturally longer to design and deliver the first time and the delivery project was mandated to deliver the supporting templates and standards.
It was initially a challenge to convince and educate our Project Managers that they were not ‘losing control’ – rather they were being given more time to concentrate on their harder unique tasks. They were still assigned technical support people who would work on their artefacts, but the technical staff were owned and managed within the COE. The COE technical staff were able to specialise in many platforms, and staff could move between COEs once they had mastered new skills; by doing the ‘easy’ things faster, they were able to spend time on new technologies.
Measuring the process and benefits
The COE Managers met weekly to review and measure the entire projects’ portfolio end to end. Collectively they were able to identify any potential missed SLAs and were able to prioritise, and or request additional staff if the pipeline was being held up by a particular resource-type shortage. This was the first time that ALL 1,000-plus projects were able to be reviewed within our infrastructure pipeline. They were able to identify opportunities in improving the overall process, timescales, SLAs and organisational structure, which was further optimised after 15 months.
The benefits of the factory process were that we were able to save 25,000 man days’ effort on approximately 1,200 projects in the first year of full operation. Eighty-two projects using pre-provisioned infrastructure saved at least two months on this single aspect per project; we saved on average six months on infrastructure projects that had taken 18 months within the lifecycle. We missed less than five per cent of SLAs within the first year. In the second and following years the bulk of the artefacts were all in the libraries and therefore completely re-usable.