Skip to main content

Making decisions with data – Working through the analytics maturity model

In my past pieces on data and analytics, I’ve looked at issues such as data quality and how companies can improve their use of analytics. For this article, I will take you through how organisations can make more of data across the business.

This use of data follows a maturity model. Similar to other areas of IT, there are initial steps that all companies must take around data that form the building blocks for further, more complex use cases.

While it might seem tempting to try and go around this approach, it can lead to problems, such as business users not understanding how to work with data or IT not being able to support users effectively. Like any maturity model, it is important not to try to run before you walk.

Step 1 – Putting the foundations in place for data-driven decisions

Whether your company has traditional business intelligence platforms in place – or doesn’t yet have BI implemented at all – it’s important to look at how to get the basics right. This begins with understanding past performance and the drivers behind it. This will be based on existing business units and their reports to provide a picture of the business as it stands, describing what has taken place in the past.

Based on this, it’s then possible to start diagnosing how decisions made led to specific results. For marketing professionals, this might be looking at the business impact of their campaigns; for sales teams, this can be analysing how forecast new business compared to results achieved and compensation plans. By getting more insight into what happened, it’s then possible to look at why.

This foundation involves linking data sources together, so that each business team can look at data in context. Rather than relying on individuals to look at stand-alone reports and correlate data in their heads, this process can ensure that everyone is on the same set of data to start off with.

Step 2 – Knowing the right metrics

Using this foundation, it’s then possible to create more detailed reports and dashboards that can show what is really taking place right now. This can also provide better context for performance around areas such as sales and marketing.

For example, before implementing analytics, teams may have been responsible for setting their own goals and targets in isolation. While Marketing would count its Return on Investment (ROI) based on the number of qualified leads passed to Sales, Sales would be looking at deals closed and revenue generated. These two targets should link together nicely. After all, a Sales department can’t create and close all the business on its own. However, there are many instances when there is friction between the teams on roles and responsibilities that may be difficult to spot in each team’s silos of data.

In this example, what Marketing defines as a lead may be different to what the Sales function is used to working with; equally Sales may be having problems converting business where there is interest built up through content and marketing activities as the customers are not understanding the value proposition well enough. Without the right data in place, it is difficult to get to the root cause of problems that might exist.

Using data on past performance, it is possible to look at how each team has delivered on its objectives and see where there are issues. Following this, it is possible to create new metrics for success that encompass the needs of the wider business, rather than the individual teams involved. This information can then be used to drive more collaboration and success through changing the metrics that are used. For Sales and Marketing, this might involve looking at how long it takes to close each customer on average, or how many leads it takes to generate certain revenue targets.

Ultimately, getting to the right metrics involves identifying the key value indicators that each department relies on and turning these into items that are relevant across the whole business. Since each team can link up their data sources, this represents an opportunity to think of measures that will help analyse the business in a more holistic way, rather than simply replicating what has been done before with isolated sources.

Step 3 – Predicting where to invest

The first two stages in this maturity model are about getting the foundations right for the future. Once these are complete they act as a baseline that helps look forward and start using data to influence the direction of the business. Predictive analytics involves using a company’s past data to ‘predict’ with some confidence what might happen in the future.

In the example of Sales and Marketing alignment, predictive analytics can be used to show how investment in certain sales channels and marketing campaigns should generate specific returns. Similarly, it is possible to look at HR data to see how employees are performing and look for patterns that might help future recruitment decisions. By seeking out patterns in the data, it’s possible to make better decisions on areas such as investment in marketing campaigns or hiring of new recruits.

Once you get the descriptive and diagnostic sides of analytics for business teams complete, it’s time to start looking further forward and with greater ambition. All the previous steps in this maturity model have been aimed at using existing data to justify current decisions.

Predictive analytics takes this further by modelling the impact of decisions to see what might happen in different scenarios. By doing this, it helps you load the dice so that each decision made has a greater impact for the company.

This area of analytics is complex to implement. Previously, it was normally the preserve of data scientists or data analysts to use this approach to working with data. The complexity involved made it difficult to deliver the results of this analytics out to business users in a way that they can consume and use.

However, this approach to data is starting to make its way into the lives of business users, as the individual tasks that make up predictive analytics can now be automated. Rather than relying on data scientists to parse the data, make it ready for analysis and then send over the results, a more self-service approach can be offered.

The important result here is that analytics can be made suitable for everyone within a business, rather than it remaining solely useful for an organisation’s data scientists. The aim for predictive analytics is to make it easier for everyone within a business to understand the impact of their decisions and make the choices involved clearer for everyone.

Step 4 - The future is prescriptive

For example, prescriptive analytics can be very useful for staff in sales environments, where data can be used to improve the chances of success. Rather than relying on individual judgment, data can be used to help decisions get made with a greater chance of success. Looking at areas such as price or product bundles, each sales person can benefit from the experience of others across the business in what they recommend. By making recommendations, each sales contact can potentially increase the profitability of their deals and their win ratios, which can have a huge impact on the business.

Over time, analytics can be used by more people across the business, moving from static descriptive and diagnostic data, through to recommendations around what is the best approach to take in given situations.

More importantly, these prescriptions for success can be created and shared across the business to improve profitability and speed of action.

Pedro Arellano, Director of Product Strategy, Birst

Pedro Arellano
Pedro Arellano is vice president, product strategy at Birst, leading development around networked data and analytics. Prior to Birst, he led marketing at MicroStrategy and hosted the Stereo Gol radio show.