Today, planners need to know about data science and technology development. Insight teams need to create metrics that predict or shape the future. Customer experience teams need to be as comfortable with numbers as emotions. A sea change for analysts is that ‘predictive analysis’ is now a key area of focus. It’s where the work we all do leads to, and where the value arises. Look at these four stages to see how we can drive data-led decisions that add more value.
1. Report on the past
Most management reports and MI tell us what happened, but often not why. To know where we are now is a critical first step. You can’t plan a journey or track our progress without understanding our starting point. Crucially, each stage of insight is important, but this first step doesn’t change anything, in itself.
2. Explain the past
To move beyond reporting to insight, we need to ask why? Root Cause Analysis (RCA), the five Whys, Control Charts, and many tools like this, help us interpret our data and determine what is causing this to happen. Interpretation requires special care to avoid unconscious or data bias, which we will address later in this chapter.
3. Predict the future
Looking to the future is where we can start to add tangible value and what enables planning, insight, customer experience and transformation teams to join up their approaches. This is not just producing a forecast but painting a picture of the steps and stages that take us towards our strategic goals. Great models allow us to understand the dynamics of the situations we face and propose alternative futures for different scenarios.
4. Shape the future
These same models can be used to shape the future, by breaking down siloes, influencing decision makers, creating action plans, joining corporate project teams. Think of it from the perspective of the business leaders. Why wouldn’t they want you involved if they only understood what you could offer? The key to unlocking this is to start with your organisation’s strategic goals. Then identify the key drivers that are within our control or influence. Work out how to pull the right levers and we start to make a real difference.
The need for change
The need for change stems from the unprecedented pace of change in the underlying digital technologies we all use, which has only been accelerated by the pandemic. Processing capacity has been doubling every year (Moore’s Law), AI capacity is doubling every three months, speech analytics accuracy has reached a tipping point of 92%.
Last year we shared the Gartner data maturity model, leading to discussion in each professional community about how we progress from hindsight to foresight. This year at The Forum, we are sharing the next evolution in our model for the stages of insight. It is not enough to understand what happened in the past, we need to focus on how we shape the future and make a difference.
As we know, there’s huge potential in our data, yet much of it still needs unlocking. As analysts, we are the key to this. So, in our 21st year as a professional community, we call upon all kinds of analyst to pick up this challenge as we come of age. Our success lies in the decisions we enable or make – and the behaviour that is changed as a result. What have we stopped? Started? Adapted? Re-engineered? If no one does anything differently, we’ve delivered no value.
What is predictive analysis?
The danger is that we treat the future as outside our control, not our job. In fact, there is a lot we can influence. Why just analyse the past or create a plan when we can be part of the forces that shape the future? Predictive analysts create a picture of what we can realistically expect at key milestones into the future. If things turn out differently, we are prepared. We learn quickly. We respond fast. This depends on trust in the way we manage and share our data.
Planning teams are one kind of predictive analyst. They traditionally own many forecasts in the budget process, as we see later in the chapter. The challenge in forecasting is to use a wider range of sources and methods. We need to move beyond a narrow focus on volumes, for instance, to be predicting and shaping productivity, customer satisfaction, employee engagement, revenue, costs, retention, and many other things that link to our strategic goals. In an ideal world, any key measure can be used in this way for predictive analysis. This means that in creating dashboards or working with quality assurance, we can start by setting parameters of what to expect.
In this way tools like Control Charts become predictive tools because they give a clear picture of when something lies outside the norm and trigger a clear need for action. In the same way, our learning modules on – Setting The Right Targets are encouraging people to think of the underlying dynamics of all kinds of metrics. In the same way, we need to review underlying dynamics of all kinds of metrics, as we explore in our learning modules on Setting The Right Targets.
The future of dashboards
The next step is for dashboards to include predictive metrics, which look forward. In planning teams, budgets have done this for a long time. We don’t just include forecasts of demand but predict what this will mean when balanced with supply and analysed according to an (ever increasing number of) scenarios and assumptions. In Quality, we can show advisors, managers or product owners what will happen if certain behaviours continue current trends or what would happen if the average performance moved, or the lower quartile were raised to the average or certain outliers were corrected.
Forward-focussed displays and reports can help show the impact of different decisions and actions, especially if they are embedded in some interactive system, as a what-if tool for users. We are drawn to ask: what happens if we do this? The fourth stage of insight is more goal based, working back from strategic objectives to focus on how we deliver what matters most. As a result, everyone can be focussing on the future, as well as asking the question ‘why?’ We crowd-source insight! With the rapid growth of data tools such as PowerBI this is accelerating. We are moving slowly away from EXCEL or Email as the prime means of reporting. Another key feature is the ability to track and report on what you do with reports (access, usage etc). What if frontline advisors could spell out for themselves the difference they would make, if they did some things differently?
Changing roles and the need for collaboration
This is a big change from our separate roles in the past, when planners produced a forecast for scheduling and MI analysts reported what happened. When time permitted, there always has been some great analysis of why something was happening and what we could do on the back of that. Today, best practice sets the bar high, joining up our approaches and making predictive analysis the norm not the exception.
This means much of it needs to be quick and easy to do, even better if it’s automated. This doesn’t mean that we are all doing similar roles. Instead, a team will blend skills, as we see already in many insight teams with technical developers, data scientists, analytics specialists, and those whose role is primarily to engage. Cross-functional teams are also common now, such as planning teams with access to dedicated data scientist or developers, or comms/quality teams with a strong analysis or programming skill set. In any case, you don’t have to be in the same actual team to collaborate!
If predictive analysis is to make it easy to see when we go off track, like a plan or roadmap, or to enable realtime process automation, there are many questions we will need to address. What does accuracy mean for this kind of analysis? What systems, skills and procedures do you need for data collection and interpretation? If we have well-trusted predictive metrics, can we automate more? What needs to be in place to empower the necessary rapid actions, for us to be truly agile? What would happen if more data analysis were automated so that we had this open up for us every morning, like our email inbox does now, or lined up every week for a stakeholder review? This much is clear already: engagement and trust will be the key to unlocking the power of predictive analysis.
Finally, remember that your processes, behaviours, and customer expectations are constantly changing. This means that your models need to be reviewed regularly, within an appropriate governance framework. In planning, this is important throughout the planning cycle, in Budget Planning and Operational Readiness for instance. In insight or quality what is our review framework? The true measure of success for insight or forecasts lies in what we do with this information and how much it is trusted.
Author: Paul Smedley & Ian Roberston
Date: 27th April 2021