**** Warning, nerd alert. If you are not into data or algorithms, this article may put you to sleep, or worse, it could make you drink Redbull and have a craving for Cheetos, and play with data and work in the dark and all hours of the night. There is a section in here that will skip to the easy part and no nerd conversions are possible ****
Skip Tracing is a lot more involved than jumping into a paid database and doing a quick check. That is what we refer to as recon and quick wins. In the real world of skip tracing, you don’t always get given the correct details. In some cases, the name you are provided is just the client’s version of spelling. Enter Soundex.
Soundex is commonly used to index names and can also be used for searching databases. There is more than one version of Soundex code and not all are suitable for every use. For example, when it comes to searching databases (assuming you are searching your own dataset), using a standard Soundex algorithm will not give you very good results, why?
Soundex is basically a phonetic organisation tool. It allows the user to put names in phonetic order by Retaining the first character and placing the rest with numbers.
1 = BPFV
2 = CGJKQSXZ
3 = DT
4 = L
5 = MN
6 = R
Ignored = AEIOUWYH
There are a few rules added to this, but this is the basic structure of the American Soundex system.
Most of the standard Soundex formats don’t allow for every variance, for example, Jon, John, Jhon and Jonh will all return as J500. This is great, from a skip tracing or investigation point of view, if you search via sounded it will be searching for all spelling variances, right?
Wrong, I have seen Soundex used incorrectly many times in the past. From a Phonetics point of view, standard Soundex code doesn’t do it all. For example, Steven and Stephen will both return S315, and this is a good thing. But any name starting with PH and F will not, even though Ph and F can have the same sound. So the names Pheonix (P250) and the name Foenix (F520) return very different results.
The point is, with less common names, this tool can help return results for alternative spellings. While it isn’t 100%, it is a helpful tool when you are stuck and when you are not sure about the correct spelling of a name.
From a data nerd point of view, this is very useful. Let’s face it, Soundex code is very simple. And it is very easy to write your own algorithm or create your own set of rules to apply Soundex to a unique problem. Using standard Soundex algorithms for searching databases is a messy and lazy way of doing things, but it can have some good results if you don’t mind all the false positives. Writing your own rules on the other hand, that are unique for the individual application is where Soundex can really take things to the next level.
When it comes to writing predictive skip tracing algorithms, names really matter. And getting a name right can make a big difference in the success of your code or manual investigation.
You can use the two Soundex tools below to explore how it works. This code is based on S. Morse’s standard American format for reverse Soundex and a combination of encoding formats.
Use this tool to convert your subjects name. John For example will return (J500)
Use this tool to convert the result to alternative spellings: so J500 will return John, Jon and a whole heap of other names that are not useful to you.