We have analysed the data and concluded that the Multiple Residence data cannot simply be added to PAF, especially when automatically matching addresses. Therefore, we have blended PAF with Multiple Residence, by removing some conflicting data, which gives a much higher match rate than other solutions that have merely added Multiple Residence to PAF.
In doing so, this enables us to increase address matching by 0.36% for a mixed dataset of PAF and non-PAF addresses, as opposed to a simple added version only seeing the match rate increase by 0.18%. Without our approach, this would have a significant effect on how many addresses can be uniquely identified.
For example the following Multiple Residence address conflicts with that already contained in the PAF:
Flat 1-3, 31 West Street, Leicester, LE1 6XN (PAF)
Flat 1, 31 West Street, Leicester, LE1 6XN (Multiple Residence)
Flat 1 will now occur at two separate addresses making it difficult to uniquely match. However, Hopewiser’s blended approach will eliminate these problems.
Multiple Residence is a great addition for organisations who are taking orders online and/or over the phone. The blended dataset will enable users to select the correct address quickly and easily, thus preventing irritation or abandonment of shopping cart, whilst removing duplicate information that could cause differences within your data.
Read more about the importance of data cleansing.