David's Primavera Stuff
P6 and the project controls environment. Part one.

Perhaps the most important attribute of P6, to me at least, is the absence of any facilities for the estimate-and-approve stage of a project. What a schedule is for, is to schedule the scope of work. P6 requires you to have a scope of work before you create a schedule.

It would seem obvious that estimate and approval should be part of the ‘project management’ software, probably tightly coupled with scheduling. This has a very difficult history.

Let’s go back in time, to the Lotus 1-2-3 vs Lotus Symphony debate. This was probably before many of the readers of this were born. Lotus 1-2-3 took over the MSDOS spreadsheet market with a vengeance. Visicalc never really migrated successfully from Apple II, Microsoft’s Multiplan did not get the attention it deserved from Microsoft, and the fabulous Supercalc was owned by Computer Associates, enough said. Lotus 1-2-3 was ‘just’ a spreadsheet, to complete your PC desktop you probably also ran WordPerfect for word processing from a different vendor, and Harvard Graphics from a third. Three different vendors each with a best-of-breed software package, which struggled a wee bit to communicate. Then the Lotus company introduced Symphony. It was not as good at spreadsheet, word processing or graphics as the three separate packages, but had a single data source and no issues with data conversion, or an output with text, spreadsheets and charts.

Symphony was introduced, and the ‘best of breed vs integrated’ debate started, in 1984. We are still discussing it.

In our world, well near our world, the best know integrated package is SAP. You can take all of the maintenance workorders in the PM module, and schedule them in the PS module. There are some issues, I would characterise SAP’s PS module as amongst the very best production planning engines. If I was manufacturing 5000 nearly identical BMWs a week, SAP would be a great solution. I have serious reservations about it for project planning.

There have been several attempts to integrate work-scope generation and management with scheduling, mPower from my chums at Monitor is perhaps the best known. I see that in July 2022 the software was acquired by a Canadian company. Canada provided a lot of clients and revenue for Monitor. Artemis is fondly remembered by many in the UK’s energy sector. It offered scheduling linked to a database you could customise, but building a norms-based estimating package staring with ‘just’ a database was just a huge effort and I don’t have any success stories about that most crucial of integrations.

The overwhelming choice amongst my clients is to use a ‘best of breed’ scheduling package – say P6 – and integrate it with an estimate and approval system.





P6 and the project controls environment. Part two.

Counting the clients I have worked for in the last 12 years, I find 16 contractors and 15 owners. The fiasco of ACS’s demise means I have no detailed history before that. I have not gone through the convoluted A&M machinations that will have certainly reduced those numbers a wee bit.

One of many important distinctions in this key differentiator, is the approach to estimate and approval (E&A). Owners are ‘top down’ and contractors ‘bottom up’.

When an owner thinks about a new project, it starts with an idea like ‘let’s develop a new oil field’, and as the project goes through what is now a well-defined process for go/no go, it as if the branches of a WBS are better defined and granulated at each stage of the process before the final decision to proceed is made. P6 from version 1 had the start of a system for doing this, the Spending Plan at the project and WBS level, was clearly built (but not finished) with this in mind:



LOOK! Even a ‘Benefit plan’!! Gosh, so we do projects to make us money, not just get rid of it! Who Knew?

When I first saw this back in 1999, my naïve heart skipped a beat. But. It does not integrate with Excel, and it does not ‘nudge’ – when the project slips, or is brought forwards you have to type all of the values in again. Also, maximum timescale is 3 years. There are other problems, but those three were enough to ensure I have never had a client use it. The principle is sound.

When a contractor thinks about a new project, it is a completely different world. They will have received a tender, and go/no go is determined by the success or failure of a bid. They need to generate a bottom-up estimate, usually from a norms based estimating system. Primavera has never had any software that does this (P6 Cloud’s work management system will be discussed in another rant) REMEMBER, P6 is an ‘owners’ system that ‘contractors’ use because their clients require it of them.

So both owners and contractors require an external E&A system, which have completely different characteristics, to interface with P6.





Primavera – the future. (As of Q3 2022!)

Now boys and girls, I want you to suspend your disbelief if you can. Imagine, I know its hard, but imagine if the Oracle corporation cared more about shareholder returns that fulfilling your functionality dreams.

The most money, and essentially the easiest to forecast money, that is being made in the IT industry is in software as a service (SaaS), aka the old mainframe bureau business. Many of you will have noticed how keen Oracle is to run your IT infrastructure for you, as they desperately play catch up with the big players in this market place led, I kid you not, by Amazon, the Chinese tat and books marketplace. While companies like NVIDIA’s stock can rise and fall on a change in the crypto currency mining industry, if you have three-year SaaS deal with mega-corporations, you can predict revenue with the best reliability in the industry. So where does that take us?

Leaving aside the issue of SaaS vs on-premis, which I have discussed elsewhere on this site, what about the functionality of the SaaS-only P6 Cloud? I shall be starting a full evaluation of it very soon.

To be continued. I am off for a couple of weeks to the country estate ( a wee fisherman’s cottage) and I might take a break from writing this stuff.





Risk

Primavera Risk Analysis (PRA) is, like P6 itself, a magnificent piece of late 20th Century software. It no longer integrates properly with P6, that is the dialogue to read a project directly from the P6 database no longer works (although my colleague Allan Murray can sort this for you) and you require to import XER files. This makes a mockery of the risk assessment process when every tiny change to the schedule to respond to the last simulation needs you to export a new XER file and get lost in a sea of revision control and file names. Also, there are bugs that are unfixed. Yes, Oracle will still sell you a licence for, last time I looked, £5,500.00. Gasp.


WARNING – incoming anecdote. In 1977 Charlie Jordan, then running ‘Aberdeen and Texas Finance’ soon to be Abtex in Carden Place in Aberdeen, invited me over to see ‘Pertmaster’. An Apple II product he had paid to be developed, that ran on a twin floppy machine, with a third party 80-column video card (luxury) attached to a Qume daisy wheel printer. Completely amazing. I was running Project 2 on an IBM mainframe to do the same application. It did a bar made of ‘C’ characters for the critical path and used ‘-‘ for float.


PRA has a not-very-well integrated risk register, but has pretty much all the quantitative tools required to do the job. Shame it is obsolete.

P6 Cloud has a risk module. This supercedes the P6 Prime module, which was horrible. I shall be reviewing it.

What to do? Clearly, a third-party tool is required. See my SaaS vs on-prem discussion in conjunction with this.

Owners


If you are a very large corporation, willing and able to invest the time and money in an enterprise-wide risk management system, the Norwegian product Safran selects itself.

Contractors (or owners needing to dip a toe in the water)

Full Monte from Berbecana offers unprecedented price/performance, and perfect P6 integration. This what I use. There will be a full review, eventually,
here.