TIBCO: All analytics endeavours must lead to action

More is not necessarily better.

This is the realisation the big data analytics (BDA) revolution has arrived at after nearly two decades of trial and error.

The vast volumes of data collected over the years, also sees the industry enter a new era of technology to handle and manage these data in its various shapes and forms. Fuelled by machine learning, analytics is now in advanced phase.

Now, organisations realise the big benefits they can gain from big data and analytics and figures by McKinsey and Company show that in the US, there is a total shortage of 190,000 talents with analytical data skills. Analysts and managers who are equipped to understand big data analysis and make decisions based on them are in even lesser numbers, to the tune of 1.5 million thereabouts.

Why do analytics projects fail?

Over half of analytics project fail even, and yet the success of the few like Amazon, Netflix, Google and nearer to home, Media Prima, are outweighing the failure of the many, and demand for data-driven decision-making, and from thereon automation, is at an all-time high.

If you are an organisation in the midst of deploying a data science and analytics project or about to embark on one, what do you do?

In the company’s blog, he stated, “If I had to pick one fundamental reason why, I’d say that it’s because analytics projects are irreducibly complex and multi-faceted: they typically have many moving parts, more than a software project for example.”

As a result, projects very quickly become overwhelmed with technical details, if they haven’t failed from the start due to data being off-limits or hard to find.

Establish business objectives

TIBCO’s Steven Hillion shared the main reasons why analytics projects fail, in the company’s blog citing that projects often move ahead without business objectives being established first.

Instead, the needs of the business have to come first and a good way to kickstart analytics projects is to define business problems that can be solved with analytics. This also means that business users have to be involved in the strategy right from the start, especially because they are the stakeholders who will be using the analytics application and can give realistic representation of their day-to-day requirements.

Circling back to the beginning of this article, data quantity rarely surpasses data quality when it comes to deriving meaningful insights.

Even after insights are squeezed out of all that data, Hillion observed that it is common to assume the more information there is, the better the decision will be.

But that’s the equivalent of having a meeting to address a specific issue and “resolving” it by scheduling another meeting, he wrote.

More is not necessarily better

 For insights to be valuable, it should be predictive in nature, and link to action, for example a pre-emptive sales and marketing action to stem a forecasted decline in sales volume.

Organisations like Netflix and Amazon are masters at this, and Netflix even took things several steps further by not just recommending content to watch according to viewers’ preference. They actually took all the data they have on viewer’s likes and dislikes, and created original content which they are confident will be huge hits.

According to HIllion, “Connecting analytics to actual results demands ‘high resolution’ data and predictive analysis that prompts actions within purchasing, sales, lead generation – whatever the business objective may be.”

“But one final piece is needed to make this all simple, and seamless: integration.”

TIBCO’s newly announced TIBCO Data Science product combines the capabilities of other different products, and has the main objective of operationalising data science quickly for the organisation.

More users however can lead to successful implementation

Hillion recommends starting on one analytics project in the beginning. It should be the one with high business potential and readily available data, if the organisation wants to have a chance of inculcating a data culture.

 Besides involving business stakeholders from the start, representative analytics that is easy to understand, has to be pushed to the business applications that they use like sales, marketing, procurement and more – this is critical for linking insight to action.

Business users should not have to worry about the technical details of interacting with or using an analytics application.

Readying access to analytics content this way, can help to offset the scarcity in data specialist skills.

It’s a journey…

An analytics ad data science project is an on-going work in progress, with new data sources becoming available as business opportunities, challenges and priorities evolve.

To cope with all these moving parts, feedback loops are critical.

Hillion recommends a flexible analytics infrastructure that can meet diverse requirements and adapt quickly to changing business needs. Focus on rapid development of fresh models that are production-ready instead of a “perfect” solution that may not be relevant for very long, he says.

Some companies even give their employees free access to their code bases, for them to experiment with source code, code reviews, and metrics for success.

TIBCO proposes that this can be done via a unified platform that combines the capabilities of their other products like TIBCO Data Science (including TIBCO Statistica)[LB1] , TIBCO Spotfire Statistics Services and TERR. With this integrated platform, data science becomes a team sport, as it empowers collaboration. https://www.tibco.com/products/data-science

…but it doesn’t have to be complex

Data science is already a complex field because of the multiple disciplines involved in finding, extracting and surfacing trends, by leveraging analytical methods, domain expertise, technologies, and other fields like data mining, machine learning, predictive analytics and more.

Business users’ focus should remain on the business, and the complex end-to-end analytics lifecycle can be greatly simplified via a unified platform that also ensures security and governance. Ultimately, it allows businesses to expand data science deployments organisation-wide via flexible authoring and deployment capabilities.