2018 is the year of analytics.
It is likely to be the year when data and analytics reaches every industry. In retail and the consumer packaged goods industry, the digital transformation has hugely changed the operating model, generating multiple sales channels to manage. In financial services, the last crisis and the increase in regulation has forced financial organisations to reconsider their data and analytics strategy. In the public sector, budget deficits have forced governments to find new solutions to improve their level of services without spending more money. The management of analytics resources is becoming crucial from this perspective and is probably the beginning of a new analytics economy.
Data and analytics are now at the heart of the operating model in many industries
This is a huge challenge, and forces organisations to reconsider their structure, processes, data and technologies at an enterprise level. Why this level? Ultimately, change need to happen based on a holistic view of the business across all its functions: marketing, sales, finance and risk, R&D, production, delivery and customer services.
Forrester suggested back in 2016 that enterprise architecture professionals were looking for platforms they could trust. There is a wide range of technology available, but it is important to have a simplified way to develop and deploy data analytics applications to meet firms’ requirements for insights at the right moment to support the most important customer and employee decisions.
Open source software does not always deliver business value
From a technology perspective, it is hard to distinguish between open source and commercial software. Data and analytics capabilities, including data preparation, data quality, predictive modelling, machine learning, artificial intelligence, scoring deployment, decision management and model governance, all look pretty much the same. Open source software, however, has the advantage that it looks cheap and innovative, and is therefore very attractive.
It is extremely challenging to compare analytics platforms on paper, especially from a features perspective. You would, after all, not benchmark a new car without driving it. It is the same for analytics. Your platform choice should be based on real-life use cases, with measures of the total cost of ownership and time to deliver the best results.
Companies relying exclusively on open source software for analytics often have to build huge teams to cope with the development and maintenance of their framework. They have to manage their data and analytics projects like a software company, with all the associated costs. Is that really their core mission? Are they really delivering value for the business (and its customers)? What will happen if key data scientists or developers leave the company?
A look at the value of commercial software
This is where commercial software really scores. For example, last year SAS released a new enterprise analytics platform that takes advantage of new and emerging technologies like machine learning, artificial intelligence, Hadoop, in-memory processing and cloud deployment. Companies are finding that they can compensate for the cost of this platform by the added business value it provides, compared to the effort needed to assemble and manage development using open source software. For example:
Value to action
Analytics has no value if it does not provide accurate predictions with the required level of transparency and governance, and if it is not deployed in processes to automate business decisions. The best commercial platforms offer accuracy coupled with governance, and streamlined and automated deployment within the information system.
Speed to action
Speed is important in analytics. You need to make decisions based on recent data, not data from six to nine months ago. If you have to wait that long, the market may have changed and the cost of implementation will be too high. Commercial analytics software like the SAS® Platform provides rapid answers by streamlining the latest analytical techniques, including visual analysis, analytics workflow, machine learning autotuning, and automatic deployment in database, in stream, in Hadoop and in memory. It can also be scaled up as required with cloud deployment.
Trust to action
You need to be able to rely on your analytics platform to deliver the right answers. Analytics has no value if business units cannot trust the results. A software partner is helpful in generating this trust, as it can ensure that you are using the latest technologies and guarantee confidence in the results though documented processes and software. It can also provide more than adequate security for your data. A partner will also offer technical support and professional services, giving you more than just a platform.
It is important to consider all these aspects when deciding between open source and commercial analytics platforms. At SAS we are promoting a hybrid analytics architecture. Please read this research paper.
Open source can be easily combined with the SAS Platform to secure and implement machine learning and artificial intelligence algorithms into the business process and legacy information system. A platform is more than just the initial purchase, and you may easily find that a commercial partner delivers far more in the long term, despite the higher upfront cost. I strongly believe that this winning combination will make your data and analytics use cases happen and therefore deliver the expected value.
