At DataTools we are in the enviable position of being the market leader when it comes to Australian address capture and validation. It is what we are known for and it is something we have become exceptionally skilled at. But I have come to the realisation that it is not what our customers (or the market in general) really need.
Let me say this again. The market does not need address verification.
How did I come to this highly controversial conclusion? Why would a leader in address verification technology make such a statement?
Quite simply I started to look more closely at the problems we have helped our customers solve. These problems were many and varied, with few actually having anything to do with the need for “verified” address data.
- Some customers wanted the most accurate and correctly formatted addresses to assist with their Master Data Management regime.
- Some wanted a DPID to be able to access postal discounts via Australia Post.
- Some wanted to deliver the best possible customer experience within their online stores.
- Some were chasing the most efficient data capture method for their data entry.
- Some wanted cleansed and standardised data to exchange with third party providers.
Address verification was how we achieved these outcomes. It was just the tool and it is how almost all data quality vendors approach address related issues. But it’s neither foolproof nor does it work 100% of the time. In fact, the reality is that it often causes exception handling processes that are more complicated than the issue originally set out to be fixed.
- What happens when an address cannot be verified?
- What happens when the desired address is not presented to the user?
- What happens when a DPID cannot get applied?
- What happens when multiple sources of “verified” data conflict?
This is where everything gets complicated. This is where “verification” as a tool for achieving the desired business outcome is flawed and why the market really doesn’t need address verification in the context it is promoted today.
This is also the reason organisations who are seeking to improve user experience, increase efficiency, gain better insights and reduce costs associated with address related data must look back to a time before address verification became the default resolution.
Organisations need “the best possible version of an address for their specific use case”.
Verification of an address against a third party reference dataset will only ever be relevant if the specific use case calls for compliance with that data set.
A good example of this is verification of an address against the PAF in order to obtain a DPID for access to postal discounts, as this is the only way a DPID can be obtained. However, it must be remembered that the verification only supports this specific use case and would not deliver the best possible result for other use cases such as address standardisation. If an address cannot be verified against the PAF does it mean it doesn’t exist? I’m sure anyone living in a brand new estate would argue that point!
Prior to verification against authoritative data sets becoming sold as the golden panacea of address data quality, advanced parsing and formatting logics were used as a way of “cleaning” address data. Rule and pattern based matching engines were used to mould and shape address data into “fit for purpose” pieces of useful information. Unfortunately, due to the active promotion of verification technologies over the past 2 decades, the art of address manipulation, standardisation and correction (address cleaning) has somewhat disappeared.
This being said, all is not lost. Some vendors, including the team at DataTools, have continued to develop and refine address parsing and formatting technologies that do not rely on reference data-sets in an attempt to deal with the exceptions generated by the now typical validation driven approach to address data quality.
Now don’t get me wrong. I am not saying that you shouldn’t validate address data. I’m just saying that in most cases you probably don’t NEED to. However, what you do need is address data that is formatted, structured and standardised sufficiently to meet your specific use case. That is, the best possible version of an address to deliver on your desired outcome.
If you can validate an address and it makes sense to do so for your use case, go ahead and do it, but don’t rely on it as the sole solution for address data quality or improved customer experience etc.
Performing address validation then subsequently dealing with the exceptions that arise is putting the cart before the horse. Why generate and deal with exceptions that add no real value to your business and that could have been avoided in the first place?
Standardise, format, cleanse and repair your address data. Make it fit for purpose across the many touch-points throughout your business. Once this is achieved you can explore the value in validation against a reference data set, if indeed there is a valid business case to do so.