Firmographic Match Rates

[MKosmicki]

I've been working on a few D&B Appends and was wondering what others have experienced for a record match rate. I realize not every company is in D&B and every B2B database will be different. 

I thought it would be interesting to see what others are experiencing as far as D&B match rates. So far the highest I've seen is 66%.

What are you seeing?

[mbarone · Accepted solution]

We serve about 600,000 clients across the country.  I get a hit rate of about 38% using the business matching macro that's included in the Join pallate.  This has been consistent since we purchased the Data package a couple years ago.

[anthonyta]

We've only been able to yield roughly a 39% match rate on a dataset that D&B was able to match 100%. Appears that phone, and spacial uniqueness were a big factor causing the low match rate, but have not yet figured out how to increase this to validate the purcahsing the D&B Data package.

Were you using the Business Match (US) tool out of box, or did you do something to improve that match rate?   I'd be interested to know how you were able to attain a 66% rate.. 

[MKosmicki]

I've found that if I run my data filse through CASS or Geocoding first, I have a better match rate. I've found this wtih B2B and B2C data files.

Some of my files also require doing a bit of data cleaning first, which also helps the CASS and match process.

[mbarone]

My guess is it all depends on your population and it's size.  We CASS/Geocode everyone before running them through the macro.  Our 600K clients are mostly small businesses. By nature, they come and go, sometimes several in the same location in the same couple years.  So one address could easily match to a half-dozen businesses in the D&B database.  And the names could be simlar if they use a city or something in the name.  I'd be curious to know the number of businesses in your data that your matching to D&B that gives you the 66%, as well as the type of business (small, start ups, average tenure, etc.).  That's probably priopritary information for you so it's more rhetorical   :)

[LonnieY]

I'm throwing in a bit of info I know about the macro. The BusinessfileMatching_US.yxmc is a great starting point, however, due to D&B license restrictions on returning addresses, telephone number and contacts, a round about way of creating matching keys is employed and does not currently include a key for telephone number. And unlike the Fuzzy Matching tools, there is little flexibility in the matching once the keys are built,..there's no room for tweaking. 

Behind the scenes, the macro creates matching keys from combinations of D&B business name and address components, creates keys for the user file with the same combinations then does a literal join on the keys. Adding a key for matching on telephone is a great idea and we can request this as a future enhancement. 

[Chadfloyd0]

When you say you are running through CASS first, are you referring to processes outside of the macro? I see that the macro CASS's each record before passing through, so wondering if you are doing other things as well.

[mbarone]

We do it separately for business purposes.  But yes, you could do it within the macro itself.

[MKosmicki]

I've been doing it separately. It started as an experiment...then became practice.