Hindsight, as they say, is a wonderful thing – ‘If only we’d known then what we know now’. Think about a retailer being widely ridiculed on social media for a couple of days because it couldn’t see a discount crisis looming and nip it in the bud. Or a manufacturer having to recall thousands of units because the first few breakdowns weren’t spotted in the early days of distribution. If only they’d known – prevention is a lot easier on the bottom line than recovery.
The good news, however, is that anticipating the future is perfectly possible if we optimise the way we use predictive analytics. It’s an organisational thing that calls for deploying the right technology, processes, people and disciplines. With these in place and organised for maximum efficiency, we have the ability to anticipate potential mishaps quickly and find answers to hard questions. Moreover, we can invariably do so with the information already at hand.
Let’s look at four proven steps we can take to increase analytics agility – to limit uncertainty and flag potential problems before they take hold.
Step 1.Promote predictive analytics – and its power to know – throughout the organisation. Make sure every employee acknowledges the important part analysis plays in decision making. Most organisations now recognise that their data is one of their most valuable assets – promote the understanding that analytics is the key to unlocking that value.
Simultaneously, take stock of current analytical assets. Find out where the analytics skills are – or can be developed – in all the parts of the organisation. It’s also a good idea to nominate an ‘analytics champion’ within the organisation – someone with the knowhow, authority and enthusiasm to navigate you through analytics best practice. Hire from outside if necessary.
Step 2. Undertake an analytics gap analysis. Look for any shortfalls in the organisation’s analytics skills and capabilities. Compare what is currently available with what will be needed to meet defined objectives. Consider operations such as workflow processes to see what could contribute to more accurate and timely analysis reporting – and for whom and how reports should be designed and presented for action.
Very importantly, think in terms of the whole analytics lifecycle. Look at the quality of data exploration and modification, as well as modelling, deployment, monitoring and feedback. Managing analytics as a complete lifecycle rather than a series of components will make it easier to identify shortcomings to be addressed.
Step 3. Build an analytics team – with the emphasis on ‘team’. Without disrupting proven established structures, designate an organisation-wide analytics team. This ensures analysts in different parts of the organisation know what each is working on so that they can share experiences and learn from each other. This is also a good way to avoid ‘reinventing the wheel’.
Analysts are typically heads-down individuals so sustaining the team culture needs to be worked at. Put programs in place to ensure cross-communication. Regular, formalised face-to-face or virtual team meetings are recommended, together with blog posts and other intra communication measures. This discipline will help keep all analytics projects on track.
Step 4. Match predictive analysis to the organisation’s objectives going forward. With an organisation-wide appreciation of the value of analytics established, the skills and other asset gaps identified and filled, and a cooperative team approach up and running, it’s now time to use analytics to look ahead at all times. The infrastructure is now in place to enable quick reaction to changing business needs. Required new models can be built and deployed faster and more efficiently, to deliver better quality information.
At this stage, analytics has graduated from answering current questions to prediction of what questions lie ahead. Analysis is now ready to test theoretical future scenarios and eliminate potentially unwanted outcomes through simulation, optimisation and other advanced techniques. Social media and big data sources from outside the organisation can be added to internal information and worked on in a centralised analytics hub to clarify decision options and contribute to the bottom line.
In summary – Theses four steps will be big steps for many organisations and ‘Rome wasn’t built in a day’. The ideal won’t be achieved quickly and the fortitude of the key stakeholders will be tested along the way. Sustained effort will be called upon by all.
But advancing the organisation’s predictive analytics capacity and capabilities will pay rich dividends downstream in the form of better, faster decision making when it matters and – moreover – before opportunities are missed and negative impacts on the bottom line have a chance to manifest themselves.
Geoff Beynon is the New Zealand country manager of SAS Institute – the leader in business analytics software and services, and the largest independent vendor in the business intelligence market. See www.sas.com To contact the head office of SAS Institute New Zealand call 04 917 6800.
