Aligning IT Infrastructure to Data Strategy

Given the latest advances in advanced analytics technology, organizations across the board have the potential to move analytics from reactive reporting, to proactive and predictive models. However, many struggle to implement analytics at scale, and for the long-term. According to Deloitte, 21% of analytics projects are canceled prior to being delivered, or are never used. We often see that this is due to a lack of preparedness across the organization, to take advantage of the insights that advanced analytics can deliver.

Advancing an analytics project, from proof of concept to gaining the full benefits of an advanced solution, needs buy-in and engagement from key stakeholders as they adapt to new systems, applications and workflows. It also puts pressure on the IT department, as a fully-fledged analytics capability uses large volumes of streaming, real-time data rather than more manageable historical data.

As well as being at the forefront of analytics innovation, Intel has worked closely with many organizations looking to bring advanced analytics into mainstream use, and implement a lasting advanced analytics capability. Summarizing the key advice from Intel’s new eGuide Ramp up Your Analytics Capabilities, this article outlines the three exercises that can increase the likelihood of success for IT leaders.

Conduct an Analytics Capabilities Assessment

While organizations often have multiple analytics projects in mind (see below), it pays to adopt a capabilities-driven approach that maps your analytics roadmap to your business strategy, around criteria such as innovation, customer focus, leadership, and people focus. By understanding your key organizational objectives, you can orient your advanced analytics program to deliver on these goals, as well as ensuring the executive sponsorship, strategic focus, and organizational buy-in.

You can explore opportunities to bring data-centric thinking into your organizational strategy (if it’s not there already), or start even simpler by considering how different data sources can be united to help streamline decision making in real time. From here you can build and evaluate an analytics action plan, based on available human, IT and other resources.

It’s important to understand the skills your existing team has, or could develop, and where it would make sense to call upon external vendors or consultants. Whatever the mix, your project will involve close collaboration and multiple meetings between business, technical and managerial stakeholders, so having a vendor that can speak the language of all these groups will accelerate progress.

When working with our customers on their technology plans, we recommend they make the most of the open nature of the x86 architecture by considering solutions from multiple vendors for each aspect of their data management and analytics environments. The same goes for software: many open source frameworks—such as Apache Spark*—are available online, and a number of cloud service providers (CSPs) offer analytics tools and capabilities (such as AWS Sagemaker*).

Organizations can adopt a best-of-breed approach, mixing and matching the tools they need, then working with an integration specialist to stitch them all together. Note that security remains a top priority, for internal and third-party solutions.

Build an Insights-focused, Cross-organizational Team

The success of any analytics initiative will depend on the people that design, implement, and use it. It’s critical to think about your organizational structure and how you will build an analytics-enabling team, as well as how to communicate and collaborate with business units and other stakeholders.

For a project to achieve long-term success, it will need a committed executive sponsor to ensure funding and the top-down leadership, not only for the PoC but also to enforce any workflow and cultural changes. Executive sponsorship helps drive:

  • Analytics Leadership: Your executive sponsor(s) will need help from others in the leadership team, to promote the vision and drive others to do their part in helping to achieve it.
  • Funding: As the impact of the advanced analytics organization becomes more widely recognized, more funding should become available at an organizational level.
  • Organizational Design: However you choose to structure your analytics capability (horizontal or vertical), it’s important to make sure roles and responsibilities are clearly defined.

As attracting and retaining top data scientists can be a challenge, it’s also worth developing a plan for how you will do this. You should have an idea of the specific technical skills you need and can focus recruitment efforts on meeting those requirements, and consider how you will retain and nurture that talent over time.

Define an Analytics Process

A successful analytics process needs to align with your business challenges and goals. For example, CRISP-DM13 (a recognized data mining methodology) starts with the business understanding the problems to be solved, and then investigates how to solve them through iterative application of analytic techniques. This approach enables you to evaluate how the right data is creating the desired impact to your business goals, while incorporating a fail-fast approach to learn what is not working and re-evaluate any disconnects.

Your analytics process should also work through the stages of analytics solution maturity, specifically:

  • Descriptive—Understanding what has happened, looking at historical data to describe discernible outcomes or observations of past data behaviors
  • Diagnostic—Understanding why it has happened, helping explain why the behaviors occurred, and starting to form the foundation of a predictive model
  • Predictive—Predicting what will happen, by building models and selecting algorithms based on data types used, assumptions made, and other trade-offs
  • Prescriptive—Understanding why it will happen, understanding why a prediction happened or showing the impact of input features on the model’s outcomes.

As the team works with the data to create and convert it into insights, a pipeline of how the data must be processed needs to take shape. A solution architect needs to work with data scientists and domain experts to determine what raw data needs to be collected, how to integrate it, where to store it and how to query it across data processing, modeling, and visualization. While this could create challenges (including quality checking), up-front investment can mean that integrating new data sources will be much more efficient.

It’s About Working Together

As we have seen, these three exercises involve a variety of roles and expertise, from top-level leadership to domain experts, data scientists, engineers and architects, all of whom need to work together to create analytics solutions that can support successful business insights.

It’s also important to work closely with your technology solution provider(s), to ensure you have the right combination of tools in place to empower your teams to succeed. At Intel we’re working with other industry players to make this easier for our customers by developing Intel® Select Solutions. These are tailored solutions for common workloads like advanced analytics that optimally combine hardware and software.

For more information, read the eGuide Ramp up Your Analytics Capabilities or discover how advanced analytics can help transform your business www.intel.com/analytics.


Intel® technologies’ features and benefits depend on system configuration and may require enabled hardware, software or service activation. Performance varies depending on system configuration. No computer system can be absolutely secure. Check with your system manufacturer or retailer or learn more at www.intel.com.

Published on Categories Big Data and AnalyticsTags , , , , ,

About Ken LeTourneau

Ken LeTourneau has been with Intel for 20 years and is a Solutions Architect focused on Big Data and Artificial Intelligence. He works with leading software vendors on architectures and capabilities for Big Data solutions with a focus on analytics. He provides a unique perspective to leading IT decision makers on why AI is important for 21st century organizations, advising them on architectural best practices for deploying and optimizing their infrastructure to meet their needs. Previously, Ken served as an Engineering Manager and Build Tools Engineer in Intel's Graphics Software Development and Validation group. He got his start as an Application Developer and Application Support Specialist in Intel's Information Technology group.