Periodic Reporting for period 1 - Geowox COMP (Quick, cost-effective and accurate residential property valuations for the European mortgage lender market.)
Reporting period: 2018-12-01 to 2019-05-31
We discovered property valuations work much like they did in the 19th century as manual data collection, structuring and calculus still reigns supreme today. This is an issue for large financial institutions (e.g. mortgage providers) and their regulators (central banks), which need to know the underlying value of their loan collateral at all times.
Between January and June 2019, we have undertaken a full feasibility study on our proposed automated property valuation project (Geowox COMP), to investigate its technical and commercial potential. The objective was to test:
i) product viability, to decide whether it is worth to continue development;
ii) market scalability, to maximise international potential;
iii) innovation potential, to test and define groundbreaking research in the sector.
Between January and June 2019, we have undertaken a full feasibility study on our proposed automated property valuation project (Geowox COMP), to investigate its technical and commercial potential. The objective was to test:
i) product viability, to decide whether it is worth to continue development;
ii) market scalability, to maximise international potential;
iii) innovation potential, to test and define groundbreaking research in the sector.
During the project we were able to identify the need for a ‘hybrid’ valuation model, which is a combination between an automated valuation model and a qualified surveyor review. We were also able to test this on a real property portfolio of a leading mortgage provider, and in collaboration with a leading property valuation provider.
We were able to cross-validate the product across our early prospect markets (IE, UK, BE, NL). This included phone and onsite interviews with a dozen mortgage providers. The result of this process was a clear understanding of local product requirements and feature scalability. We also validated our model with prospect hybrid partners and entered into conversation with regulators. From these conversations, we were able to confirm that our stakeholder requirements were largely similar across all markets, due to EU standards and regulations.
We also presented the results of our R&D in address matching, boundary detection and satellite image recognition. We were able to perform tests in these categories, delivering promising results for clients and uncovering clear goals to becoming the pan-European market leader in AVM technology.
We were able to cross-validate the product across our early prospect markets (IE, UK, BE, NL). This included phone and onsite interviews with a dozen mortgage providers. The result of this process was a clear understanding of local product requirements and feature scalability. We also validated our model with prospect hybrid partners and entered into conversation with regulators. From these conversations, we were able to confirm that our stakeholder requirements were largely similar across all markets, due to EU standards and regulations.
We also presented the results of our R&D in address matching, boundary detection and satellite image recognition. We were able to perform tests in these categories, delivering promising results for clients and uncovering clear goals to becoming the pan-European market leader in AVM technology.
We have completed the testing phase with two Irish banks, BOI and Ulster. We tested both our AVM, and the Hybrid valuation (with our partner CBRE). We were able to reduce time spent per valuation from 45 minutes to below 5 minutes, a productivity increase of 90%, whilst improving portfolio valuation turn-around time by over 80%.