I’ve had the pleasure of building a career delivering analytic solutions within the private and public sector.

These solutions involved providing insights into everything from forecasting energy consumption to building models that help improve safety in our mines. My current role at the Queensland Audit Office is to utilise these techniques to help my auditor colleagues deliver even more efficient and effective audits. This is a great mandate to have but what I love about this role the most is it also allows me to challenge how audits are delivered.

In the coming year I will explore the learnings from my role in a series of blogs just like this one, but for my first blog I thought I would start with what I have learnt about successful analytics engagements.

You think I’m stating the obvious.

What experience tells me is that just because it is obvious does not mean we do it. So what is my advice? Well, it can be grouped into clarity; data governance; analytic technique; and collaboration.

Clarity

This is the principle of having everyone in the engagement team, including key stakeholders, connected to the engagement and on the same page.

It sounds obvious but I have witnessed way too many projects where some really advanced analysis is performed, but the insight was not the insight required, or it took ages to produce (with associated cost to match), or even worse, it was not actionable.

The reasons this occurred usually included:

• the question to be answered was not well defined
• the question to be answered was not collaboratively designed or validated with the team
• the question to be answered was not well communicated.

Basically, how can analytic specialists help solve a problem if they are not clear on what the problem is?

Clarity is also about everyone being across any constraints on delivering the analytic technique. This includes time expectations, budget expectations, client politics, data availability and insight expectations.

Analytic professionals love doing cutting edge stuff, but we also need to be grounded in reality to be efficient and effective. Now you have clarity on what you are trying to achieve and the boundaries you need to operate within, you need to understand the available data. This is where the principles of data governance come in. Data governance is about understanding your data and respecting it. This includes understanding:

• limitations—the data you need to source and any known issues or limitations with that data
• owner—who owns that data, and engaging with the data owner for permission to source
• classification—the classification of that data and expectations on how you secure it
• availability—availability of the data and your ability to source in a timely manner
• retention—how long you are allowed to retain that data and whether this conflicts with any retention requirements in your own organisation.

Now that you have clarity regarding what you are trying to achieve, and you understand the data available to analyse, we can finally start thinking about the analytic technique.

Choosing an analytic technique is less about finding a technique that is cool and cutting edge and more about choosing one that is practical to deliver the insight you need. Emphasis on practical.

This can mean being prepared to challenge yourself during the life of the project on whether the chosen technique is still the best one, and even having the courage to stop the analysis if the required insight won’t be possible. This type of objectivity can be really hard when you have invested time, effort and emotional energy into a particular approach, but I have found it is this unbiased thinking that distinguishes successful analytic specialists from the less successful.

Another important concept is that good analytics is not about the science of the technique but the insight that comes from it. I have spent way too much time trying to educate clients on things like the science of linear regression, neural networks, multivariate statistics or cluster analysis when I should have been focusing my efforts on communicating the insight from those techniques. Analysis is about finding something you did not know and most importantly is telling what you can do with it.

This leads to the most important and most common reason for unsuccessful analytic engagements… a lack of effective collaboration.

Whether you are talking about a successful business, project or sports team, in nearly every case it’s not the individuals that made the success but the collaboration of individuals with different skills. True analytic magic comes from a group of individuals where each person focuses on where they are strong and works together towards a common goal.

It’s probably the most common-sense attribute of a successful engagement but it is easily dropped when you introduce time, budget, politics, and worst of all, ego. Ego is thinking you can do everything, and wanting to do everything.

While most professionals are capable of achieving a lot, the reality is that unless you live and breathe it every day, you will never be as good as someone who has a focus area. I have been around audit for more than 20 years and despite having delivered audits and being a Chartered Accountant, I never install myself as the audit subject matter expert. My focus is analytics. What this experience does give me though is the ability to communicate and collaborate with auditors, clients and analytic specialists. It’s this ‘translation’ role which is the hardest to fill in any analytic engagement. But if people don’t understand each other, how is your analytic engagement going to be successful?

Analytics is a team sport and is only successful if people work collaboratively, playing to their strengths.