Sort and structure your data, and patterns soon emerge. How you then choose to interpret those patterns can be the difference between profit and loss, taking or missing an opportunity or, on a more human scale, giving a vulnerable family somewhere safe and warm to live.
As the volume and variety of data generated by housing providers grows, the cost of dealing with poor quality data is becoming far more visible. Almost every housing provider operates core asset, housing and financial management systems and processes, but how good is the data and are there processes to ensure consistency in the collection, formatting and ongoing management of the data?
Seven pillars of data quality
We recommend starting with a data audit and employing the seven pillars of data quality (DQ) to help a wide range of housing providers cleanse and restructure their data, collaborating with them to look at timeliness, uniqueness, validity, accuracy, completeness, consistency and reasonability as part of a nimble and reliable data quality improvement process.
We encourage them to review the dimensions of data quality management, comprising control, assurance, improvement and planning, and breaking those down into more granular tasks and dashboards for the team so everyone understands what the data shows.
Housing providers aren’t just about buildings, they are also about looking after their communities and being able to spot data patterns can help them to look out for their residents, including ESG reporting, complaints, ‘golden threads’ and disrepairs. However, it’s important to look at the bigger picture. Rather than looking at the challenge facing one part of the business, share your business information and challenges so the organisation has a better understanding of its data-centred operations.
Joining disparate threads
For instance, in the area of complaints, we can aggregate data across all customer interactions and we might be able to see that over half relate to repairs. If we also review data on rents received, arrears, voids and outstanding repairs, we can link these common elements together within discrete datasets. This can quickly pay dividends, not just for the current situation but also for future improvement plans and maintenance programmes.
An example would be if a customer suffered a broken boiler, they would call their housing provider who would book an inspection, send an engineer and perform whatever repairs were needed. In the past, if that customer wasn’t happy with the result, they could raise a complaint and might refuse to pay their rent, at which point they would fall into arrears. Now, housing providers can quickly see the cause and effect of a non-payment and, in the longer term, reduce the likelihood of needing to initiate legal proceedings or settle disrepair claims.
‘What if’ scenarios
Using machine learning to automate processes and run hypothesis-based ‘what if’ scenarios produces incredibly useful results.
For example, by using IoT device-derived data and machine learning, when winter comes it’s possible to see the properties where the internal temperature is still low and mapping this with weather data can give us yet more detail. Using tenancy information about how many people are living in a house and how old the boiler is tells us more. By combining these information streams, housing providers can identify their most at-risk customers and see if more support is needed to maintain a healthy living environment.
Managing, simplifying and streamlining the collection, aggregation and analysis of their data helps housing providers operate more efficiently, cost-effectively and responsively, but without a bedrock of clean, reliable and consistent data in the first place, they will still be making decisions with their eyes closed.
Umesh Parekh is head of public sector at Red Olive.