A tale of the tail wagging the dog

The other day I a saw a preliminary enquiry from an organisation with a large property portfolio for a Microsoft Access condition survey system. I held my head in despair. Why specify the technology and not mention what the condition survey is to be used for? I wonder if the same people would limit their choice of a new washing machine to a specific one at an early stage without considering whether it was needed for a person living on their own or someone who regularly washes all the gear for the local football team.

My experience is that with modern IT systems the issues are rarely technology but functionality and data. It is much more important that the reason for the system and the outputs from it are fully understood. It is also important that the system should adopt recognised data structures and standards. This latter point will enable data to be more easily exchanged with other systems, both internal and external to the organisation, as well as measure performance that can be readily understood. For any property system the property structure should comply with BS 7666, particularly BS 7666-2 Spatial datasets for geographical referencing: specification for a land and property gazetteer; and for condition surveys the asset structure should comply with BS ISO 15686, particularly BS ISO 15686-5:2008, Buildings and constructed assets: service life planning: life cycle costing. Any system that fully complies with these standards will naturally require a sophisticated relational database. Interestingly I failed to fully represent these requirements in a Microsoft Access database.

A few days after seeing this enquiry, on a lovely sunny morning, I reluctantly boarded a train from my local station, where on one side I watched contented cows in a field that runs down to the river and on the other young lambs gambolling, to go and give a presentation at a conference in London. Tearing myself away from this idyllic scene I was rewarded when I sat and listened to one of the other presenters describe planning and delivering an estate strategy. An essential part of this was collecting data to benchmark the estate, and one of the principal benchmarks was condition. Perhaps more importantly though the presenter interspersed his presentation with shots of the property users and quotes from them about the service being delivered. Not surprisingly then the principal conclusion was that the estates strategy should be service [business] led NOT estate led. In turn the data that feeds the strategy should be service [business] focused. With that approach there is a person who will deliver the efficient and effective property portfolio that meets the needs of the organisation.

So to recap:
• First: service [business] requirements;
• Second: estates requirements; and a long way behind
• Third: technology requirements.

Identifying condition based work for a property portfolio – where do I start?

For most portfolio holders with little data this is a big problem. The collection of appropriate data to measure the performance of the property portfolio and ensure that required actions and work can be programmed effectively can be both expensive and time consuming; but without data the consequences are very much more costly. The starting point, therefore, must be data.

The two principal issues with data are reliability and cost of collection. Unfortunately these two issues are in conflict, that is reliable data is generally more expensive to collect.

There are a number of ways property data can be collected but for the purposes of this discussion I have classified them into four:
1. Judgement forecasting – relying on local managers, surveyors and contractors knowledge to assess the condition of properties and forecast what actions and work is required to them;
2. Modelling based on existing data sources – building a symbolic model of each property and associating data from, for example, planned preventative maintenance records;
3. Non intrusive assessments – a typical condition survey; and
4. Schedule of work required – a list of tasks required to be carried out.
The reliability of these is illustrated below:

Many of the organisations I meet currently use judgement forecasting supported by an out-of-date assessment. This is not surprising as to move from this position to, say, one of up-to-date real-time assessments is both costly and time consuming.

Rather than attempting a big leap forward, with all is implications on cost, resources and risks, I often find myself recommending an approach that incrementally improves the position. This approach is to initially model the data using as much of the available reliable data as possible including any judgement forecasting data. The outputs from the model can then be used to identify properties where action is a priority. Assessments can then be carried out on the properties identified. Based on the assessments costed schedules of work can be prepared to finally feed the decision as to what should happen to each identified property. Throughout the whole of this process, which should be continuous, the model should be updated with the most reliable data available

This approach has the following benefits:
1. Concentrating money and effort where the benefits are highest;
2. Continuously improving the property data at the lowest cost;
3. Improving the situation almost immediately; and
4. Capturing structured data thus improving accessibility and usability.