In recent years, we have seen an exponential increase in the amount of data that organisations capture and store. Most organisations typically hold ‘structured data’ – information that is ordered and arranged by categories such as age, gender, etc – which is generally used simply to retrieve specific information rapidly, or to produce historical reports.
Many also hold large amounts of ‘unstructured data’ such as correspondence, web activity logs or call-centre transcripts. It’s more difficult to find specific information within these forms, and they are less susceptible to analysis by ordinary means. However, with the appropriate technology, these sources can be exploited to yield extremely valuable information that can inform crucial strategic and operational decisions. With predictive analytics, decisions that were once made as educated guesses can be justified with convincing evidence.
Organisations that have taken this approach have achieved remarkable efficiency gains and cost savings.
Housing associations have databases holding information about their tenants, but generally use them only for everyday purposes such as retrieving personal or tenancy details. Yet these data sources contain valuable information that can help housing associations in their constant battle to meet their performance targets.
Imagine being able to predict which tenants are likely to go into arrears or suddenly vacate their premises – and being able to know, in every case, the reasons why. Predictive analytics is the key to solving these riddles, providing information to prompt timely actions that not only help meet targets but also save money and maintain revenues.
Within housing associations’ databases is a wealth of information that goes beyond age, sex and marital status to include, for example, income, number of dependants and job type. Accounting records, property type, maintenance and repair history and customer service transactions are also usually available.
All IT departments can run basic queries to confirm a hunch. For example, that young couples are more likely than older ones to fall into arrears. However, if research analysts had SPSS’ statistical tools, they would be able to drill down to a much finer level of detail.
So, while elementary analysis and experience might suggest that young couples are generally a high risk, analysis using predictive analytics might find that those with one child and one member in full-time employment are, in fact, a low risk. Further analysis could reveal, for example, that the highest risk of delinquency is among young couples without children who are unskilled labourers, did not go on to higher education and have a combined income of less than £20,000.
Armed with such detailed, statistically-valid information, managers are much better placed to implement targeted interventions that can save time, money and exasperation. Such targeted measures have helped many RSLs to meet key performance targets, increase their cash flow and benefit their tenants. ALMOs have also used predictive analytics to forecast changes in population demographics for community planning purposes, and councils have provided funding for such analyses because they ultimately benefit the community.
These are just a few examples of the advantages that can be gained by using predictive technologies; there are many other insights that can be drawn from analysing the data available to housing associations. However, only a small number make the best use of their data resources – even though doing so would produce spectacular successes.
Alex Eliades is an account manager at SPSS.