It’s sometimes difficult to measure the RoI of analytics, as value often lies in its ability to deliver better outcomes and more accurate predictions. In some cases, these are quantifiable, others the value may be intangible. Nonetheless, data is an asset that can lead to a strategic advantage.

But data itself won’t deliver that advantage. Knowing what to do with it does.

Analytics is not new.  The difference today is we have more data being produced, and better technology to get insights to decision makers where, when and how they need it.

Analyst teams often drive many business-critical projects; production reports, executive dashboards, product forecasts and customer behaviour prediction models are all business-as-usual (there is already plenty of reading material out there if you want to understand these some more).

Though in order to improve performance, organisations around the world are analysing data differently and applying new techniques to age-old questions – increase sales, improve productivity, manage risk and reduce costs.

Here are 5 thought-provoking analytics use cases – ignoring the ‘sector’, how could these processes and applications apply to your business?

1. Cross-sell/Upsell

A coffee-chain which has very healthy same-store sales, wanted to identify additional sales opportunities in order to accelerate growth. Initial analysis demonstrated that rate at which food was bought with a beverage (a key revenue driver in this environment) in the drive-thru, was half the rate of transactions purchased inside the outlet.

Further detailed analysis of individual items and relationship with other menu items, demonstrated that regular brewed coffee transactions had the lowest revenue and profitability, due to its low price and the lowest food attach rate of all drive-thru transaction types. Analysis of in-store transactions provided an understanding of the specific food menu items that had the highest attachment potential.

Pilot sales and merchandising programs featuring the items which provided the highest likelihood of growing sales proved successful and was followed by a significant marketing investment to address this newly discovered opportunity.

The client said, “This seems so simple in retrospect but—without the deep analytic dive—we would have gone on blissfully unaware of this need.”

2. Governance

Most organisations promote cultures of trust, which tends to run counter to creating a strong sense of fraud awareness, often using “that doesn’t apply in my team” to the issue. However, fraud by employees, direct or colluding with 3rd parties, occurs in most financial and business systems including purchasing, payments, billing, payroll, inventory and shipping, P-cards, and travel and entertainment expenses. Studies show a disconnect between actual rates of fraud and what most managers and executives think is the likelihood of fraud in their organisation.

The Association of Certified Fraud Examiners (ACFE) reports fraud costs businesses around 5% of revenues each year.

Analytics provides a way to both prevent and detect fraud and to improve control systems overall. The concept is simple enough: analyse every transaction that flows through systems and flag any indicators of fraud. Each form of fraud may have several indicators – or red flags – so, each transaction can be tested several ways.

Developing automated routines to perform these tests on an ongoing basis (i.e. continuous monitoring) and implementing processes to review and respond to fraud indicators as they arise means actual instances of fraud can be addressed almost immediately, before they have a chance to grow into something far more serious. Not only does this identify fraud on a timely basis, but acts as a deterrent to employees who know they are more likely to get caught.

Similar data analytics techniques are used for many other forms of transaction monitoring and for testing compliance for example SOX, supply-chain management and produce authenticity (who remembers the Italian olive oil scandals and more recent horsemeat in the UK).

3. The Paradise Papers

At 1.4 terabytes of data, containing 13.4 million documents, analytics of the Paradise papers was no small task. That said, some organisations can gather dozens of terrabytes every month.

The unstructured form of the data, with many formats not machine-readable is an issue many organisations face. A second challenge is the silo structure of data, and the requirement to link these disparate data sets. As with most organisations, Paradise Papers investigators brought data in from external databases to deepen the analysis they could undertake.

In order to transform the data into an auditable format, the investigators utilised several phases;

  • Process data into a machine readable format
  • Process the documents into machine readable format • Index the unstructured data into a searchable knowledge centre
  • Load the data into an analytics platform
  • Build analytics and visualisations to simplify the story the data was telling

The result provided unique insights into the offshore interests and tax activities of more than 120 politicians and world leaders, highlighting the relationships between politics, offshore companies and their lawyers.

With this approach, analysts can deal with data complexity and uncover hidden insights across many business challenges.

4. People Management

For most companies, people are their biggest costs and yet there has been little effort put into analysing these costs and getting the most out of these investments. Even now, HR analytics tends to focus on recruitment rather than retention, as most objectives are around time-to-hire.

A US college with 160,000 students across multiple sites wanted to understand the rate of student drop-outs and use predictive analytics to spot warning signs, intervene early and get students back on track. Multiple data sets including student records and finance were brought together for analysis and over 16,000 at-risk students were identified in the first ‘run’.

In the following 2 weeks, over 5,000 students were interviewed and where possible steered towards resources that could help their challenges.

The results;

  • 3% reduction in drop-outs
  • 4% increase in performance for students contacted
  • $,000’s dollars in unclaimed funding recovered

5. Big data applied to an individual customer

Although there were loyalty cards in the UK beforehand, the introduction of the Tesco Clubcard in 1995 changed how businesses engaged their customers significantly. Almost overnight, through analytics Tescos knowledge about their customers, buying habits and preferences increased dramatically and enabled targeted marketing to an individual level.

This approach has been replicated across many industries since and today our healthcare is benefitting from the same approach.  Pooling data across healthcare organisations and capturing individuals health data from tools such as mobile biometric sensors and smartphone apps, enables personalised analysis and healthcare models to be built.

Rather than increasing sales, healthcare is using analytics to profile individuals and more effectively anticipate, diagnose and treat disease. This not only improves the customer experience, but significantly reduces cost, both in terms of direct treatment, and also medical research – moving traditional lab-based studies to broad ‘consumer-based’ analytics and research.

The majority of healthcare organisations already using predictive analytics, believe this will save them 25% or more of their annual costs.

These are examples of challenges faced by many organisations, how can you apply analytics to improve your performance?

If you enjoyed this post, please share with your colleagues and networks.