The new reporting options that are rolling out to Project Online customers allow you to choose how you want your time phased data and give you control of the time buckets you use – and you can choose not to have any at all. My favourite is to use Fiscal Periods as that also limits the timescale of data that gets published – so reducing the data footprint and publish and reporting times by making sure you have just the data you need. For example I could choose just to have fiscal periods defined for 2018 and 2019 and wouldn’t need to worry about my 15 year long mega-project – as the data after 2019 wouldn’t be clogging up my system. But the blog title mentions Portfolio Analysis? Where does that fit in? I’ll explain…
When you are creating a new Analysis you can define the time period and granularity – but the granularity is limited to Calendar Months or Calendar Quarters:
And as I have checked the option to analyze project resource requirements against capacity the next page in the setup of the analysis will check if I have resource requirements – and give me a red banner if it finds none.
Often that message will just mean you haven’t published the projects yet – but I know mine are all up to date. What happened? Let’s take a look at my Reporting options in Server Settings:
Yes, that will do it! Having Never set for my reporting options means I am feeding no time phased data through to reporting – and the Portfolio Analysis actually uses that data for its resource requirements. Which option should I choose? Any of the others will ‘work’ in that they will show you have requirements – but some will give you the results you want – others less so. I’ll start with a bad choice so that you can understand how this all fits together. I’ll go for Fiscal Periods – and I have defined my Fiscal Periods as 4,4,5 (Each quarter split into 4,4, and then 5 weeks). I was going to make a case for 2,3,5 being a better formation – but I’m guessing not too many of my US readers would get the reference.
Now I need to publish all of my projects that are going into the Analysis and build my Portfolio Analysis again. I’ve already built an analysis on the same data set in my Project Server 2013 system and I’d expect my requirement details to look exactly like this – as I have exactly the same data. Here is my Project Server 2013 analysis and I have enough resources to do 3 of my 5 plans. If you are not familiar with the tool my availability for the different roles is shown in the top grid – then my requirments are listed in the bottom grid – along with the priority of the plans based on my pair-wise comparison to other projects and weighted by the driver priority.
How does this compare to my Project Online Analysis – where I have my reporting using my 4,4,5 Fiscal Periods? A quick aside – in Project Online we have a bug where we show you the Enterprise Global (EGlobal) as one of the projects you can select for the analysis – please don’t as it will break your analysis. I have all the same projects and drivers – same resources and used Calendar Months from March 2018 to March 2019 – so it should be the same as the analysis above shouldn't it? Well if it was I wouldn’t be blogging…
A very different picture…
My analysis say I can only resource a single project – and it looks like I have no requirements at all for resources in March, June and September – but April and July look crazy! January to March are also showing no requirements and I’ll deal with that first. My Fiscal Periods were only defined for 2018 – so I publish no timephased data for 2019 – so that makes sense.
But March, June and September? If I look at the first rows from my time phased data from the OData feed for AssignmentTimephasedDataset this gives me a clue – I see data for 2/26, 4/2, 4/30… and these are the start dates of my Fiscal Periods.
Looking at my Fiscal Period Definitions again I can confirm that none of the start dates are in March, June or September. That explains why I see no requirements. But April July and October (and January) each have two of the start dates – so in effect the choosing of the 4,4,5 pattern has doubled up my requirements in some months and isn’t registering at all in others. This is the extreme example – but you will get a similar but not so pronounced effect if you chose weekly because some calendar months will contain more ‘week start dates’ than others. The new reporting options will always show the sum time phased data for the period against the start date of that period. The effect this has on Portfolio Analysis is one thing you need to take into account when making your choice. So what would work? Daily certainly would – but you wouldn’t gain any benefit in terms of faster publishing and reporting. Monthly would also work just fine – as the start date of the period is the start date of the month – and this give good publish improvements over daily. For my money the Fiscal Period option is the best – but with Standard calendar year as the model.
With this new Fiscal Period defined and reporting configured to use Fiscal Periods I’ll re-publish my plans and build my analysis again.
Perfect! Matches up to my 2013 one – except for the 2019 data which I excluded on purpose by not setting the 2019 fiscal period yet. I hope you like the new reporting options and if you are also using Portfolio Analysis I hope this helps you make the right choice!