Advertisement

How to Modernize Analytics Strategies with Operational Data

By on
Read more about author Nick Jewell.

We all know data is valuable to our businesses. And we all know that we should be able to do more with that data. Just as Michelangelo said that every block of stone has a statue inside it and it is the task of the sculptor to discover it, so every set of data has more value inside it, waiting to be realized.

The challenge is how to get into the practical processes and steps that turn all that potential data into real value. Any modern data analytics strategy must reflect the overall goals and objectives of the business itself, and must be tightly aligned to an organization’s priorities in order to succeed. A good strategy must include clearly defined and measurable outcomes across the people, processes, and technologies to be brought together across data-driven initiatives. However, research by McKinsey found that only 30% of companies align their data analytics strategy with their organizational goals. This is a significant miss.

Building the Link Between Goals and Processes

Traditional frameworks such as Davenport & Harris or Gartner’s analytics maturity models are frequently used to implement analytics programs. These models explain the progressive value of descriptive, diagnostic, predictive, and prescriptive analytics when solving business challenges. However, do those models link up effectively with business goals over time?

Image source: Gartner

To implement these common frameworks, teams will normally adopt an agile, sprint-based approach and will also identify the gaps that need to be filled in order to improve results. Over time, the team will build on those previous layers to ask increasingly sophisticated questions against data that ranges from the historical (descriptive and diagnostic analytics), to the forward-looking (predictive and prescriptive analytics). 

A great deal of attention has been placed on descriptive analytics, drawing on large volumes of historical data to uncover trends and understand what drives customer and market behaviors. However, these results look back at what has been achieved, rather than lining up with future goals.

There is now more hype and excitement in the market about how to take that backward-looking data and use it to predict future possibilities. Predictive and prescriptive analytics aim to create forward-looking models that can anticipate future results.

In this approach, past data on, say, financial performance and profitability around business activities, is used to judge the relative impact of each investment. This informs future decisions, and data is then gathered for more analysis and to confirm those hypotheses. For example, you could adjust the marketing budget to achieve better returns and more profitability based on past successes. Predictive analytics estimates the return that your budget will achieve based on your choices. Similarly, you might use a more prescriptive model to help you assign your budget in order to get a more effective mix and meet your goals. 

Operational Analytics and Business Goals

There is another class of analytics to consider as well: operational analytics. This is a sub-domain of data analysis focused on improving business operations and processes. Operational analytics analyzes processes and workflows to detect problems and inefficiencies, increase profitability, and reduce waste. Rather than the general questions around “what should happen” that predictive and prescriptive analytics projects tend to focus on, operational analytics typically focuses on more specific questions for line-of-business users about what is taking place right now.

This means being able to drill down from the big picture (such as an overall business key performance indicator, or KPI) all the way through to the individual transactions that create that metric. The data comes from enterprise applications such as ERP or CRM platforms – the typical “systems of record” for business operations such as finance, supply chain, or order management. 

What sets operational analytics apart from more traditional analytics projects is the freshness of the data needed to make decisions. Whereas most analytics environments are fed updated data on a nightly basis, operational analytics requires data that is current. This means different things to different people, so to define this better, current data should be as close to real time as is practical, with updates arriving at intervals between five and 60 minutes, depending on the business. 

Additionally, in order to support the complex, fast-moving requirements of operational analytics, it’s vital that the data is kept as close to the original business source format as possible. This is a significant departure from orthodox data architectures, where data is progressively refined through enterprise data lakes, data warehouses, and dimensional data marts. This makes it easier for users to understand where their results came from, and that those results are repeatable. This provides more confidence to users, as they know that their numbers will reconcile, and it makes it easier for them to find any problems and resolve urgent anomalies using the raw, transactional data. 

Given the real-time nature of operational analytics, any significant additional processing would render the data obsolete and unusable within the window of opportunity that exists to make an operational decision. When you want to make a decision, you want that data to support you in that minute, rather than a day later.

Working in the Moment

There are two primary use cases of operational analytics: the first category is around enabling near-term adjustments to processes and workflows around problems and inefficiencies. By knowing that these issues exist, you can make decisions that counter them, increasing profitability and reducing waste. 

As an example, global security system leader Nortek has to mitigate everyday crises that might affect delivery. This requires thoughtful responses based on data-driven analyses. Operational analytics played a key role in how the company’s employees would resolve urgent, day-to-day data-related issues. When an international trade war threatened to impact 40% of Nortek’s business, operational teams struggled to keep up with constantly changing tariff rates and other factors. Making a decision based on old data could lead to missed opportunities, or leave it exposed to higher costs. Conversely, waiting too long would also lead to the same results, so the company’s team could not stick to its old approach.

By implementing an operational analytics solution as part of their wider enterprise data analytics strategy, Nortek was able to implement thousands of targeted operational adjustments across their global supply chain. The net result of all these decisions, made in near real time based on data that was available to those users when they needed it, was better outcomes for the company’s customers, suppliers, and overall business.

With an operational analytics strategy in place, hundreds of hours were saved by freeing IT resources from tedious manual work, with simple operational dashboards replacing “war rooms” of analytics and managers that were previously drowning in data. 

Getting Consistency Around Transactions and Data

The second category for operational analytics is around delivering consistent, high-quality customer interactions at every touchpoint. This involves analyzing transactions in near real time and informing the decision-making process that is typically managed by frontline employees. 

As an example, the world’s largest coffee retailer had used descriptive analytics over a number of years to visualize and explore store location sales and performance on a nightly basis for over 32,000 locations. This huge volume of data helped the leadership team understand performance and profitability at scale … but what about those that were running those locations? They were not able to get access to that data, nor were they able to use it in their day-to-day or hour-to-hour operations.

By implementing operational analytics the company’s location managers could answer the kinds of questions that would make a difference to them. For example, managers now have the ability to review performance of over 20,000 individual products for their locations in near real-time throughout the working day. For those store managers, they could use the data to implement special offers that would increase sales and minimize food wastage. For regional managers, this data could help them make operational decisions around supply-chain management in order to reduce waste and ensure stores had what they needed and when they needed it. 

Tying Operational Data and Business Goals Together

In complex and fast-moving business environments, it’s vital that modern data analytics strategies deliver solutions for data at all points in the business decision-making lifecycle. This includes providing historical perspectives and forward-looking models. However, it has to include supporting decisions made in the moment and occurring right now in their mission-critical operations.

Without operational analytics, companies will find it harder to link the decisions that their staff make in their working lives to the data that the business gathers and holds. It also makes it harder to move beyond employees working based on hunches and past behaviors, rather than on data.

The ability to make data-driven decisions on operational data within a fleeting window of opportunity is a source of considerable competitive advantage. It also enables companies to scale up their performance by making it easier for everyone to make the right decisions in their given situations. This makes it easier to align data and analytics with those strategic goals for the company, and easier for those decisions to lead to positive results, creating the positive program out of raw data.

Leave a Reply