How to Plan your Analytics Strategy

We hear that “Data is the new Oil” but how do you extract real value from your data?  Do you exist in a data rich but information poor enterprise? Do you have a business crying out for answers to critical questions but find it difficult to generate those answers? In this article we highlight the key issues to address when planning your Data Mining strategy.

Let’s face it; your data is one of the few assets you have that your competitors cannot replicate. So how do you make a data mining strategy work for you? How do you plan a strategy that delivers insights that can be turned into tangible business benefits; Insights that help you increase your organisation’s top line or drives greater efficiencies?

Predictive Analytics is a Business Process, not a technology. It starts with a business question and ends with an implemented business process. Most failed data mining strategies can be traced back to the fundamental error of focusing on the technology instead of the process.

The key principles to ensure a successful Data Mining strategy include:

1. Setting a realistic vision based on a clear understanding of your current analytics capability
2. Getting the right people – and that means business representatives, not only analysts or data scientists!
3. Designing a Competence Centre that builds momentum
4. Remembering at all times that it’s all about “business business business”.

1. Know where you stand and build a Vision accordingly

Before you set goals and objectives you need to understand your current capabilities.   Data Analytics is a business process. At the core of defining analytics maturity is your organisation’s ability to make decisions based on data across all hierarchical levels and across all functional areas.

Understanding your organisation’s “analytics maturity” will bring you a long way towards setting the right vision for analytics in your organisation. This is as simple (or as difficult) a answering the following questions:

• What business value do you expect to derive from your analytics function? How closely linked is this to your corporate strategy?
• How much demand is there for advanced analytics within the organization?
• Do you have full backing from senior management for your strategy?
• What is the motivation and ability to sustain an advanced analytics function?
• Do you have a central data repository?
• How complex and diverse is the analytics you are planning (CRM and PM)? Analytics needs input from Subject Matter Experts so it should be based close to these.
• Have you identified the size of your investment in analytics?
• How complex is the deployment of the results?

Completing a benchmarking exercise against your Industry peers is extremely useful for assessing where you stand and setting the right level of ambition for Data Analytics in your organisation.

2. Getting the right people – it is not only about the analysts or the data scientists it is about the Libbat!

Believe it or not, this is the single most important person on the analytics team. A Libbat is the link between business and technology and the person who will define and deliver your most important analytical projects. Your Libbat will combine critical key skills (communications, business acumen, knowledge of the analytical process and project management). These people do not grow on trees, so treat them well.

The Libbat is responsible for the communication and coordination of the project from initial scoping (Business Requirements) through to final deployment. This does not negate the key role of the analyst in implementing an analytical model but the Libbat will act as project manager to ensure deadlines, milestones and critical tasks are addressed on time.

This ensures all projects have clear governance but remain flexible enough to change when required.

Business managers will always ask for countless ad hoc, low value reports “as a matter of urgency”. Following this route will swamp and suffocate the advance analytics function and ensure your good analysts leave. In order to build, grow and sustain an Advanced Analytics function the Head of the Analytics function should have the characteristics of a Sherman tank.

The Head of Analytics must:

1. Clearly identify high value projects
2. Clearly define deployment plans
3. Agreed the method and timeframe to measure success or failure and ROI
4. Be willing to say no to protect his team from distracting, low-value requests

3. Getting the right design for your Competence Centre to build momentum

A high-performing Analytics Competence Centre provides a wide range of capabilities that go beyond Data Analytics including:

• Business Analysis and Performance Management;
• Data Analytics and Modelling;
• IT Integration and data;
• Project Management ;
• Change Management;
• Deployment Management (making sure solutions are deployed and are still fit-for-purpose);
• Business Support;
• “Evangelism”.

Central, decentralised or virtual? Where should the competency belong in the organisation?

Advanced Analytics is a cross enterprise function that should have its own department and Head. It should NOT reside within IT or a business department such as Marketing. A key characteristic of analytics is “innovation” and its position within the organisation must facilitate this freedom to experiment.

For those looking to place analytics at the heart of enterprise decision making, an Analytics Competency Centre will pool your best resources and enable you to develop standard best–in-class procedures for all analytics projects. The key benefit is that you will have ‘one version of the truth’ enabling cross-collaboration between departments, which helps pull down the ‘Chinese walls’ that often divide units in the same organization.

A central analytics function also is more effective at driving innovation and new thinking when applying analytics to new and existing business challenges. A by-product of this is the motivation of the analytics resources (please don’t call them scientists), who will have to address constantly changing use cases across the entire business.

4. Remembering – it is about “business business business”

Building an Analytics Strategy and Competency Centre will deliver long term competitive advantage and could put your name in lights.

It is worth repeating that Predictive Analytics is a Business process and that the focus should be always on what is needed in an operational environment to answer your business question. Everything you do should be validated against the test of “how will this contribute to answering our business question?”

In planning your implementation, some key principles will help give you the best chance of success:

• Start with building very senior/capable analytics teams before the scale builds up
• Ensure good communication with senior management to get agreement on what you are doing
• Start with the pain and follow the pain! – Identify a real, costly business pain that is aligned to the company’s strategic plan.
• Agree a KPI to demonstrate Return on Investment – measure the right thing for success
• Continuously re-invest/innovate in your analytics team
• Encourage a culture of collaboration and knowledge sharing
• But ensure the team is disciplined in planning and execution – analysts need to be managed and told when to stop!
• Be ruthless regarding adherence to the CRISP methodology – paralysis by analysis is a dreadful thing
• Finally, don’t hide your success. This is not a function where “build it and they will come” applies. You will need to constantly broadcast your success and capability. You need to constantly remind business managers about what your team can deliver.

Your data is unique to you and how you use it can deliver truly unique insights and stellar ROI.  Like the ancient Mariner, you may be awash with data, don’t try to boil the ocean. Think Big, start small (and grow)!

More Analytics Blog Posts