Ownership, value and effectiveness: The context around analytics platforms


Analytics platformsPlatforms are currently big news in analytics. There are, however, still questions about the platform capability required to support the analytics life cycle and customer understanding of the issues.

1. Getting value from platforms is NOT about technology

When we talk to customers, it is clear that they are not thinking of platforms in terms of technical capacity. Instead, they see them as a means to an end and, particularly, a way to bring people together, breaking down silos and barriers, and encouraging collaboration with a purpose. The focus is very much on “creating business value” rather than “doing analytics.” For companies, platforms create space to complete projects and generate value, instead of spending time unconsciously reinventing the wheel. If we are talking solely about the technical capacity of platforms, are we trying to have the wrong conversation?

When talking about platforms, users don`t talk about technology, but about breaking down silos and barriers, and encouraging collaboration with a purpose.


2. Platform ownership is a tricky question

You would think that platform ownership would be relatively straightforward, but no. In some organisations, the chief digital officer or chief analytics officer is the obvious owner, but in others – perhaps less analytics-centric – these posts may not even exist. If an individual business unit is the owner, this may create friction later when others want to use it. There are also difficult questions about who holds the budget, who is responsible for ongoing maintenance, and who has the right to use the platform. Whoever the designated owner, however, the focus must be on using the platform to generate value for the business as a whole.

3. Enabling data preparation is a key function of an analytics platform

Some estimates suggest that organisations spend up to 80 percent of their modelling time on data preparation. This is not wholly unreasonable because analytics is very much a matter of “garbage in, garbage out.” But an analytics platform can help to improve self-service data preparation. For example, it might have the capacity to allow knowledgeable users to combine, transform and cleanse data before use, or to bring together the right data sources and hand them over to IT for data preparation. The key is ensuring a consistent feed of data to the analytics resources.

Beyond the AI hype, the question is how data-driven innovation can be brought to life. What steps are needed to move AI out of the lab and into business operations? Join SAS experts at an event near you and be inspired to lead your team in the new age of analytics.

4. The most effective platforms are flexible

The key benefit of a platform is that it can be used by many different people, for different purposes. It must therefore, almost by definition, be flexible. It needs to work for both business users and those who program in code routinely. It must also allow a wide range of data sources and types of data, from traditional data warehouses to streaming data and social media. It has to work with existing architecture and processes, both data and analytic. Finally, it needs to be easily scalable, so that complexity and volume can be ramped up or down on demand, and new and different types of algorithms can be added when they become available or needed.

5. Analytics platforms may solve many of the problems of model deployment

Model deployment is a huge challenge for many organisations. This is not surprising: Making the move from pilot stage to business as usual is a challenge in any project or program, and analytics is no exception. Platforms make model deployment significantly easier, meaning that more people can use and get value from the model much earlier. For example, there is no need for recoding for different use cases, such as streaming, batch or on demand. The move from innovation to deployment is seamless and straightforward, and therefore saves huge amounts of time and energy. Instead of taking six months, value can be generated almost immediately – and because the model is more up to date, the value is greater. The key is having the appropriate integration points and the right processes to support this.

6. Users and stakeholders will need reassurance about several aspects of the platform

User acceptance is an important part of implementation. Both users and other stakeholders will need to be confident that the platform is secure and well-governed. On top of that, it has to meet the needs of users for flexibility, accuracy, collaboration, speed and adding value. They also need to know that it will scale, so that there is some level of confidence in its ongoing future in the organisation. It would be helpful if top managers made a commitment that the platform is the future, not least for collaboration, and that it would receive further investment when necessary. This could go a long way towards reassuring users that it is worth investing time in platform exploration and use.

Tags #Road2AI

About Author

Adrian Jones

Director, Global Technology Practice

Adrian is Director for Pre-Sales Support in the SAS Global Technology Practice, with a focus on Big Data, High-Performance Analytics and Enterprise Architecture. Adrian looks at how organisations use data and analytics strategically, with a specific focus on optimizing the data and analytics architecture.

Leave A Reply

Back to Top