The old

If you have any experience of software like Business Objects, SPSS, Jaspersoft or even some of the more advanced features of many spreadsheet programs, you’re likely to understand the basic functionality: Take some data and produce reports or visualisations in order to gain some insight or understanding not apparent simply by looking at the numbers. After that, the most important part is actually doing something with that understanding.

The limiting factors of this type of software are often the processing power of the user's local machine (or sometimes a reporting server) and the fact that they were largely designed with structured, relational datasets in mind. Perhaps more importantly, they are not really designed to work with live data or data at the scale found in modern enterprises. There is also limited scope to use the insights they generate programmatically - i.e. for machines to take action based on reports. Lastly, many of the existing packages rely on users having a reasonably in-depth understanding of the actual data itself, how to structure queries and then how to use the software to turn the resulting information into something useful like a table or a graph.

Whilst vendors have sought to simplify these packages, improve their user interfaces and add integration functionality - some even introducing predictive capabilities and more - the computing capacity required to deal with big data, as well as the accelerating trend towards mobility limits the life of this generation of software.

The inaccessible new

As tech vendors like IBM, Amazon, Google, HP and many others have built out their cloud offerings to include services such as machine learning, predictive analysis, automation, database-as-a-service and much more, web-scale IT has become easily accessible to web and application developers.

This on-demand, elastic way of working has allowed the current generation of web sites and applications to make use of massive computing engines and has equipped some businesses with the tools to manage, understand and act upon data at scale. Examples include the likes of recommendation engines (which can make eCommerce product suggestions in real time, based on thousands of previous transactions, considering everything from the sex of the buyer, to the weather at their location) to applications which monitor tens of thousands of connected devices, predicting when one node is likely to fail based on the behaviour of the others and proactively scheduling services.

The problem is that to utilise these new platforms, users need a level of technical understanding which is beyond most non-specialist IT managers, let alone many of their colleagues throughout the enterprise. This means that while applications and websites have taken a huge step forward, marketing managers, finance managers and many other business professionals are stuck using the last generation of software or in many cases, no software at all.

Intelligence for the masses

With their latest analytics suites, IBM and Microsoft are looking to change all that. If they can somehow connect non-technical staff to their new platforms, they will have found thousands of new customers. As their human-esque names might suggest, Microsoft's Cortana and IBM's Watson are designed to quite literally put a friendly face on the vipers nest of servers, API's and algorithms that make up their services.

The real substance of both suites is the user interface. The most striking change is that users don’t have to understand the datasets, have knowledge of query languages, or even think about how a query should be structured; they just ask human questions. Typical examples might be as simple as "What is our best selling product" or as complicated as "How likely are women to respond to marketing campaigns featuring insurance products in March next year".

Excel spreadsheets and csv files are naturally no problem, but these new cloud-based suites are designed for easy connection to remote data sources, enabling live reporting, monitoring and predictions. In fact, in addition to this, Microsoft and IBM are making it easy for businesses to base the next generation applications and services on Cortana or Watson directly.

Imagine if a bank customer was able to simply ask "Why has my card been declined?" or administration staff were able to "Create a report of all customers who haven't had their car serviced in the last 2 years" inside the web sites or applications they are already using. The new breed of analytics “suites” and the now well-established software that underpins them are designed to be deeply integrated into both customer-facing and back-office applications.

Not just a pretty face

Simply put then, Cortana and Watson are nothing more than a gateway to the powerful platforms which their owners have spent many years developing. This simplified access to machine learning, elastic processing capacity and much more is truly a revelation. They allow businesses to work easily with large-scale data, giving them the tools to gain real-time insight and make the accurate predictions needed to power information-based decision making. Perhaps most importantly of all, they place complex and powerful computing capabilities in the hands of many more staff in the enterprise. In these respects, this new breed of analytics suites really do offer game-changing functionality.

Links & Further Resources

ibm.com/analytics/watson-analytics

microsoft.com/cortanaanalytics