Project Déjà vu, which was awarded funding in round 1 of the Market Improvement Fund, aimed to demonstrate the benefits of increased data quality by conducting a pilot project that cleansed and matches a selection of supply points (SPIDs) to their relevant Unique Property Reference Number (UPRN)/Valuation Office Agency (VOA) reference.

The team have now completed the pilot study, demonstrating the success of their methodology, which involved adapting a methodology from a similar project that took place in Scotland. We spoke to Greig Ogilvie, Operations Director at Precision Water to find out more.

What were three of the biggest conclusions or successes from this project? 

“From our perspective, the project has yielded several significant conclusions and successes. Firstly, we are pleased to have demonstrated the effectiveness of our methodology and technology in the English water market. Combining our matching algorithm with our Matchbox system, which allows for expert manual review, gave us a high and accurate match rate despite some poor SPID address quality. This is an important finding as it provides us with confidence that the approach can be used in future projects to deliver reliable results.

Secondly, we were able to identify sites that should be deregistered which will help to clean market data. This is an important step in improving the overall quality of data in the market and ensuring that resources are directed towards the sites that are most likely to deliver value.

Thirdly, we were able to discover missing sites through our analysis. By identifying these missing sites, we can work towards achieving greater coverage in the market and providing market participants with a more complete picture. We believe that these conclusions demonstrate the value of taking a structured and comprehensive approach to data cleanse in the market, and we look forward to applying similar approaches in future projects.”

According to your report, you successfully adapted a methodology used in a similar project in Scotland - how difficult was this to adapt and did you encounter any challenges?

“Adapting the basic methodology used in the project in Scotland was relatively straightforward. However, we encountered some challenges during the process. One of the key challenges was around the different data sources used in England compared to Scotland. For instance, the main data source used in Scotland to match SPIDs was the Scottish Assessors valuation roll which has different contents to the VOA data in England. Another challenge was the ambiguous property eligibility rules in England. This made it difficult to determine which properties were suitable for matching, and we had to rely on different criteria to assess their suitability. Additionally, unlike in Scotland, there were multiple wholesalers in England with different set-ups and rules, which added complexity. Overall, while the basic methodology adapted well, we had to overcome several challenges to make it work in the English context.”

How would you like the findings of your report to be used by the market and its participants?

“Our report presents a comprehensive analysis of the pilot area, which we believe can be valuable for market participants in a variety of ways. Firstly, we would like the findings of our report to give the market confidence that the results of our sample provide benefits that could be achieved if the methodology was extended to the entire market. Our analysis highlights the benefits of using a comprehensive and structured approach to address matching, which can provide market participants with a more complete picture of the market.

Additionally, we recommend specific actions that market participants can take to improve the accuracy and completeness of their data. If these recommendations are implemented, we believe that it would be possible to achieve over 90 per cent coverage rate for Unique Address Reference Numbers (UARNs) and UPRNs. This would be a significant improvement on the current situation and would provide market participants with greater confidence in the accuracy of their data.

We hope that the findings of our report will be used by market participants to improve the quality of their data and to make more informed decisions in their day-to-day operations. By using a comprehensive and structured approach to data analysis, we believe that market participants can achieve significant benefits, which can ultimately lead to a more efficient and effective market.”

How do you think this project can complement the Data Cleanse Service being undertaken by MOSL?

“The Project Déjà vu and MOSL’s Data Cleanse Service complement each other as they are aligned to achieve the same outcome – that of better-quality data. The Data Cleanse Service aims to address issues such as eligibility problems with supply points, poor address data, as well as match and validate VOA references and UPRNs across the market. Project Déjà vu identified and classified these same issues in trial areas, which we documented in our final report. The project demonstrated that addressing these issues is both possible and worthwhile, as we saw benefits in the pilot. The Data Cleanse Service could implement some of the processes, methodology, systems, and lessons learned from Project Déjà vu. Our mixed and scalable approach of automated and manual match review, which we conducted using our data cleanse platform, resulted in match rates of up to 90% for VOA and UPRN, as well as approximately 5% of SPIDs reviewed found to be not eligible. This approach would be especially helpful for the 400K properties where automated matching was unable to match to external data sources during MOSL’s data quality assessment.”

You can find download the final report via the project page here.

Receive the most relevant and up-to-date communications from MOSL by signing up to our mailing list.