The PMO needs to get to grips with Big Data

Big Data in the PMO

It seems that everyone is talking about Big Data. Organizations across the world are receiving more information from their IT systems, websites, and tools than ever before. The PMO should play a pivotal role in navigating it.

As a profession with so many Analysts, why do we not do more Analytics?

It has always seemed bizarre to me. Many PMO professionals have ‘Analyst’ in their job title, yet they never seem to spend much time analyzing things!

We are in an age where we can (and do) gather increasing amounts of data about our organizations, projects, suppliers, markets, and people. This data resides in CRM systems, PPM tools, HRIS databases, and document management systems. Arguably, these sources could be used to make better data-driven decisions about how we deliver projects and seek to extract value from them.

Despite the apparent potential, few PMOs have made progress by tapping into this pool of data. I believe this is down to three key reasons:

  1. There is a lack of understanding of analytical thinking in PMOs. Many people who work in the PMO space have been Project Managers in the past and have a Project Management mindset. The focus is on the ‘now,’ getting this project delivered and hitting the next milestone. As an ex project manager – I am certainly guilty of this!
  2. A lack of skills or access to the relevant skillsets. Too much focus is on training PMO Analysts in Project Management techniques and methods. Analysts are routinely sheep-dipped with PMP or PRINCE2 certifications, rather than being encouraged to develop Data Science, Big Data, or Statistical methods.
  3. The wrong tools. Portfolio Project Management (PPM) tools focus almost exclusively on planning, capacity management, and budget tracking. While a well-customized PPM system can certainly help with project delivery, they have their limitations. For example, they are woefully inadequate when it comes to an understanding of people, sales flows, and the kind of business metrics that should be used to assess benefits and value.

Big Data Analysis matters. If PMOs don’t know and understand the data, they cannot be the ‘single source of truth’ that they often claim to be.

The PMO needs to understand the data that drive the organization. If the PMO does not know and understand the data, they cannot be the ‘single source of truth’ that they often claim to be. The credibility will reside firmly with Finance teams and BAU teams who understand their data far better than we do.

The good news is that despite what consultants and suppliers may tell you, PMOs don’t need to make significant investments to get started with Analytics – spreadsheets will suffice. They often offer advantages in that there is a low learning curve, spreadsheets are flexible, and you can quickly mash data together from a variety of systems and sources. Then, of course, there is the number one advantage – you can start immediately, as your PMO analysts already have the software on their laptops.

Starting to get to grips with data immediately is essential. It is to easy for inertia to slip in. How many times have you heard people say that ‘once they get their Salesforce/SAP/Oracle/Clarity system installed, they will be able to start analyzing data’? In practice, once the systems are delivered, they are often too complicated for all but the highly trained business intelligence experts to interrogate.

So, what can PMOs do to take advantage of this array of Big Data?

 So what can PMOs do to take advantage of this array of so-called Big Data? First, PMOs need to develop that analytical mindset. You need to be able to gather data, analyze it, draw conclusions, and present those conclusions as intelligence to drive decisions. Is Technical Debt slowing delivery down? Are some areas of the business better at deriving value from projects than others – why is that? Should the company be investing more in automated and multivariate testing, or training? What skill sets will we need on our projects in three years?

Step 1: Define the problem

Defining the question (or set of questions) is the first step. Once the problem is established, the next step is to build an analytical model – first, by identifying the causal factors, then identifying the data and data sources required.

Step 2: Tackle data quality

After this process of model building, PMO Analysts will need to get to grips with the more mundane issues of data management and governance. Is the data good enough to answer the question? If not, what do we need to start gathering, and how? What data do we need from other areas of the business, such as sales and HR? How will we persuade them to work with us?

Data quality is a massive issue. I’ve experienced it first-hand at several Project Hack events. Project Hack is an event where project data analysts and data scientists get together at a hackathon to solve project challenges. The events generally run over a weekend, with teams competing for top prizes. One of the biggest reasons teams come unstuck is when their innovative ideas fall foul of project datasets that are simply not up to scratch due to data entry or compliance issues.

Step 3: Adress skills and capability gaps

The next problem is one of the skills and capabilities. Most PMOs do not have people with expertise in analytics and statistical software packages, such as R and JMP. Investing in training is one possibility (I suggested an online data science course in my article 5 online training courses that will boost your PMO Career). Still, PMOs could also consider hiring resources with statistical backgrounds or tapping into local universities to find students looking for projects and internships.

Step 4: Software (but don’t rush into it)

The final problem is tools. This is the last of the three challenges to tackle. While it is often tempting to rush out and invest in a tool, if you are too hasty, you WILL end up with buyers remorse. Focus on building the models and skills first. The insight and capabilities you will develop will help you make more informed decisions about the type of tool you need when the time comes.