Housing Technology interviewed business intelligence, data science and analytics experts from Civica, FLS – Fast Lean Smart, Housemark, Itica and NEC Software Solutions about how housing providers could take advantage of the latest BI and analytics packages and services, what to look out for, what to avoid and how to embed BI into housing providers’ everyday operations.
Why should housing providers adopt BI and analytics?
Jeremy Squire, managing director, FLS – Fast Lean Smart, said, “BI is the use of software to combine business analytics, data mining, data visualisation, data tools and infrastructure as well as best practices to help housing providers to plan more strategically, using technology to manage their assets in a proactive manner.
“BI helps housing providers by showing present and historical data within their business context and enabling their business analysts to predict future trends. Used effectively, BI can help housing providers with compliance, analysis of customers’ behaviour and performance tracking to optimise their operations, discover any problems and to predict pathways to future success.”
John Doughty, director, Itica, said, “It’s important to define what business intelligence (BI) is because it’s often assumed that it is a technical solution. BI is a strategy that generates data value and enables organisations to make better business decisions. The main driver for housing providers to develop a BI strategy should therefore be to inform their business decisions and ultimately drive measurable business value.”
Trevor Hampton, director of housing solutions, NEC Software Solutions, said, “Adopting BI and analytics will help housing providers identify and understand their tenants and their assets better. They can then target resources where they are most needed to support vulnerable tenants and ensure their properties are safety compliant.
“BI and analytics will be game-changing for housing when it’s more widely used to predict and prevent problems before they arise. For example, we’ve worked with a housing provider who reduced its arrears by £500,000 within 18 months by using analytics to identify patterns and trends in rent payments; this helped it determine which of its households were at risk of arrears and so target its support more accurately and effectively.”
What’s preventing the adoption of BI and analytics?
Sarah Paxton, business analyst, Civica, said, “The cost of investing in BI and analytics can be a barrier, combined with the time and resources it can take to implement them. Furthermore, a lack of expertise, particularly around data science and analytics, can prevent housing providers exploring it themselves. While the amount of data captured within housing management systems is vast, concerns over the quality of the data and how it might be used are also a barrier for some housing providers.”
Itica’s Doughty said, “In general, it’s every aspect of data (i.e. quality, structure, accessibility, primacy, ownership, leadership and governance) that prevents housing providers from really achieving business value from an investment in BI.
“Housing providers should forget about investing in leading-edge technologies such as artificial intelligence until they have addressed their data problems, and it’s not as straightforward as trying to fix the problem by just buying some data-cleansing tools; it’s usually the human element of the data lifecycle that needs to be addressed as the first priority.
“Some of the IT suppliers to our sector also make this a lot harder than it needs to be yet they will need very deep pockets to fix that. Examples of this include data structures which are poorly-documented or hard-to-follow, data structures not designed for BI and a lack of support for tools such as Power BI.”
NEC’s Hampton said, “The number-one problem preventing many housing providers from fully adopting BI and analytics is fragmented IT systems. Unconnected datasets leave housing providers drowning in information but not in insights because they can’t determine which version of their data is correct.
“Access to accurate, high-quality data is critical to the successful adoption of BI and analytics. Datasets become outdated very quickly if housing providers’ staff aren’t trained to understand the importance of collecting and maintaining data within the context of their overall business operations.”
Daz Chauhan, principal data consultant, Housemark, said, “The adoption of BI and analytics by housing providers has many challenges related to resourcing, understanding and resistance to change. However, by addressing these challenges and investing in the necessary tools and expertise, housing providers can gain significant benefits from BI and analytics such as improved decision-making, cost savings and enhanced tenant engagement.”
Real-time data analytics & reporting is common in other sectors; why not in housing?
Itica’s Doughty said, “The reason good analytics is not as prevalent in housing as it is in other business sectors is largely down to a lack of investment (or the ability to invest), organisational capabilities, existing IT suppliers’ capabilities and data leadership.
“While we should question the absolute need for widespread real-time data in housing, there is definitely a place for it when it comes to asset-performance alerts, such as via IoT sensors detecting the failure or degradation of an asset component.”
NEC’s Hampton said, “It’s important to recognise that significant progress is being made by our sector in using real-time data analytics and reporting. The fact that it’s less common than in other sectors comes down to either a lack of integrated IT or because the housing provider doesn’t have real-time data analytics and reporting embedded at the heart of its systems.
“For example, housing staff should see their priority tasks as soon as they log on. For this to happen, they need access to real-time information about repairs, complaints, rent accounts and asset compliance. Confidence in the data is also key; for many years, housing providers only had access to low-quality data, leading to an understandable reticence to actually trust their information.”
FLS’s Squire said, “The paybacks from using real-time analytics are huge. For example, since we’ve been working with housing providers’ in-house repairs and maintenance teams, it’s common to see increases of 30 per cent in job completions by operatives, based on (in our case) real-time scheduling and reporting of repairs.”
Do you need dedicated BI & analytics software?
Civica’s Paxton said, “Our role as a supplier is to deliver solutions within our software which make use of the specialist tools so our customers can embrace them, from embedded Power BI dashboards and interfaces to smart sensors through to machine-learning algorithms to predict arrears. Introducing a new module to existing software that’s already familiar to end-users not only makes its adoption easier, faster and cheaper but also means that specialist, in-house skills aren’t needed.”
FLS’s Squire said, “Although off-the-shelf solutions are available, we’ve found that each housing provider wants to combine data from their own range of IT systems, so bespoke solutions can work better.
“Rather than simply analysing the data, predictive analytics helps to anticipate long-term business outcomes more confidently, using historical data, machine learning and AI to predict what will happen in the future. This involves feeding historical data into an algorithm that looks for trends and patterns in the data and creates a model for them.”
Housemark’s Chauhan said, “Whether or not housing providers need dedicated BI and analytics software will depend on their specific needs and circumstances. While dedicated BI software may provide more advanced analytics capabilities, there are alternatives that might be more feasible or cost-effective for some providers.
How do you ensure you’ve the right data?
NEC’s Hampton said, “You need to ensure that the right data is collected so that when it’s analysed, it can answer the questions you want it to. We’re working with customers on several co-design and customer segmentation projects, including building more robust datasets around damp and mould. Properties and tenants can be ranked according to risk if the key initiators have been correctly established. For example, some properties will be more at risk due to their intrinsic structural design, putting them into a higher risk category; knowing this information will make it easier to take preventative action.”
Itica’s Doughty said, “This is where the definition of a good data strategy becomes an absolute priority. A properly-defined and well-implemented data strategy will help you to change the attitudes and behaviours towards data in your organisation. If you don’t invest time in this then you’ll always have a leak in your proverbial bucket, regardless of how much money you spend on data cleansing and other technologies.
“When defining a data strategy, it’s important to take an intrinsic look at your data assets, regardless of whether they are held in databases, spreadsheets or documents. A good data strategy will ensure that an organisation can leverage the potential value of its data, making it clear who is responsible for what across the data lifecycle and ensuring that everyone knows where the ‘version of the truth’ lies. Getting people to care about the data is half the battle in making sure you have good quality data for BI purposes.”
Civica’s Paxton said, “Understand the problem you are trying to solve and work out what data you will or might need. Data integrity is always a challenge – for example, historic or poorly-managed data can cause real problems to your BI and analytics outputs.
“It’s really important to ensure that everyone understands the importance of what data is captured, what that data drives from a BI and analytics perspective and then how it can affect decisions further down the line. The visibility of data at different levels within a BI solution can really help to bring data quality to the forefront.”
What are the bottlenecks to achieving good BI?
Itica’s Doughty said, “BI is only good as the data value it generates; data quality and a lack of a good data strategy are the primary bottlenecks to generating good data value. Some of the IT suppliers don’t make things easy and we would advise spending more effort on this area when evaluating new or replacement solutions.
“The important things to look for are external reporting repositories that are easy to get at (using tools such as Power BI), well-designed, documented and published data structures, data ‘snapshotting’ (a big requirement in housing), data aggregation and data dimensioning, combined with strong, built-in dashboarding and reporting.”
NEC’s Hampton said, “The main bottleneck to achieving good BI is when there are too many disparate systems. Data quality is also a key barrier because inaccurate or incomplete customer and asset data won’t provide an accurate picture. Furthermore, a lack of data sharing between departments can also be a blocker because sharing data improves its quality.”
How do you democratise BI & analytics?
Civica’s Paxton said, “Make your BI and analytics easily accessible to your housing staff. If you can embed it in the software they’re already using and can be accessed at the click of a button, they’re much more likely to use it.
“Make sure your BI is relatable to the jobs they’re doing – it’s not just about KPIs and percentages, there’s a huge amount of data available in housing systems and some of the supplementary or circumstantial data can help paint a much clearer picture. Where possible, use real-time data to demonstrate how end-users are making a genuine difference. And remain open to suggestions about improvements; the journey doesn’t just end once something is live.”
Housemark’s Chauhan said, “Offering training to help housing staff understand how to use data analytics tools and techniques effectively is vital to achieve true self-service. This might include formal training sessions and qualifications as well as ongoing support and guidance. Use data visualisation tools to help staff understand and interpret data more easily to ease the learning curve for people not accustomed to using data on a regular basis.”
Itica’s Doughty said, “This is not a question of technology – housing staff need to be genuinely empowered to make decisions at the appropriate level and have the right skills to understand the data and ask the right questions of it.
“Making sure housing staff can interact with the data is critical to exploiting its potential value. This has been difficult in the past because the cost of some of the BI solutions, particularly with user-based licence models.
“For example, Microsoft is making this much easier for housing providers, with multiple ways of presenting BI data being already included as part of its 365 licensing that most housing providers already have. The cost of Power BI Pro licences is still relatively low but not everyone will need these, particularly if their line-of-business applications have good built-in BI capabilities.”
Good and bad example of BI and analytics
Civica’s Paxton said, “Our Arrears Analytics functionality is a good example of what’s possible. It uses machine-learning to analyse rent accounts and identify those accounts that would benefit from interventions. It’s augmented with embedded Power BI dashboards within our software as a single solution – this lets us show the end-users not only the results of the account analysis but also a much wider dataset from across the system to provide a detailed picture for each account and the circumstances which exist but may not necessarily be immediately obvious. Putting the user at the centre of the solution and empowering them with the right data at the right time allows them to make the most appropriate decisions about how to act.”
FLS’s Squire said, “Insights from AI and IoT, such as using data from Switchee devices to monitor and identify condensation, damp and mould, are being used to take a proactive approach to tackling disrepair problems before residents even report them. For example, Your Housing is using our dynamic scheduling software to enable wider visibility, freeing its scheduling team to focus on other areas. The data from the system also enables optimised tracking and analysis of performance in a way which was not possible before.”
Housemark’s Chauhan said, “Some instances of ‘good’ BI might include using data analytics to identify vulnerable tenants or those at risk of arrears, or to identify trends in repairs and maintenance requests. A couple of examples of ‘bad’ BI include using data analysis to profile tenants and identify those who are likely to be a nuisance or cause problems, and using data analytics to inform decisions around the allocation of housing, which could lead to unfair or unequal access to housing and reinforce existing social inequalities.”
Itica’s Doughty said, “One housing provider went beyond the ‘basic’ skills needed to set up BI reports to a full-on ‘data science’ approach of understanding what its data was portraying; one immediate insight was that over half of the housing provider’s arrears (by value) was due to fewer than 50 tenancies. Another housing provider simply added an extra field to a BI report for customer segmentation; this revealed that around two-thirds of the tenants had some form of disability, leading to a change in how its services were designed and the way in which tenant interactions were handled.
“The two most-common examples of ‘bad’ BI that we come across are, firstly, housing providers that have treated BI as a technology project not a business strategy, resulting in poor returns on their investments, and secondly, prolific use of spreadsheets for manipulating data – while there is nothing wrong with spreadsheets per se, failures in governance relating to user access, version control and data management will create problems with compliance and regulatory reporting.”
NEC’s Hampton said, “The best examples of housing providers harnessing BI and analytics are when a 360-degree view of their tenants and their property assets are combined in one platform. This shows a holistic picture of how the tenant is interacting with the property, giving the housing provider a deeper insight into any problems that arise. One housing provider we work with was able to analyse its entire stock and predict instances of damp and mould at an accuracy of 75 per cent with very little effort.”
Housing Technology would like to thank Sarah Paxton (Civica), Jeremy Squire (FLS – Fast Lean Smart), Daz Chauhan (Housemark), John Doughty (Itica) and Trevor Hampton (NEC Software Solutions) for their editorial contributions.
