Geopointe

Data Quality and Geocoding

Data quality is extremely important to ensure accurate geocoding results. To help you understand why, here is an outline of the address conversion process:

  1. An address is entered into Salesforce, broken into five separate fields: Street, City, State, Postal Code, and Country.

  2. The address is passed from Salesforce into the geocoder. The geocoder expects to receive your data broken down as specified above.

  3. The geocoder parses the information and determines if it can find a match. If it does, you’ll see a return very similar to what you entered. If it can’t find an exact match, it will do its best with the information given.

When the geocoder can’t find an exact match, it’s a black box to us, and it has no human help to pull the strings.

Below are some guidelines to ensure the best return possible. Some come from Google, some from Geopointe customer experiences.

Use the right fields - Be sure to enter the correct data in the correct field. For example, do not put the entire address in the Street field, as in the example below. Avoid putting data into the field that does not match its label, such as postal code data in the City field. Even if all of the information is there, the geocoder cannot read it if it is in the wrong field. 

BAD
GOOD


Business Names - A guaranteed way to confuse the geocoder and return bad results is to include the business name in the address. Make sure to avoid this.

BAD
GOOD
Tiffany & Co.
200 5th Ave
New York, NY 10010
US
200 5th Ave
New York, NY 10010
US


Building Names - This one is especially important for countries that use building names for addresses, like the UK or Ireland. Whenever possible, favor the street number over the building name. We understand this you can't always avoid this, but try to avoid it when possible. 

BAD
GOOD
Wilkinson House
45 Cherry Blossom Drive
London NE5 9FS
UK
45 Cherry Blossom Drive
London NE5 9FS
UK


Spellcheck - The geocoder service cannot spellcheck. Chicgo may not end up in Chicago. Missississippi may not end up in Mississippi. Englnad may not end up in England. Furthermore, the geocoder does not recognize all abbreviations (e.g. SF for San Francisco). Even if everything appears to be spelled correctly, double check that correct characters were used; a lowercase l looks a lot like an uppercase I, as does a O and a 0.

BAD
GOOD
4 Yawkey Way
Bstn, Massachewsets 02215
US
4 Yawkey Way
Boston, MA 02215
US
335 PoweII St
San Francisco, CA 941
O2
335 Powell St
San Francisco, CA 94102


Filling in fields unnecessarily - Text like INCOMPLETE, or question marks or hyphens to demonstrate that part of the address is unknown, will weaken the interpretation. Such additional text may actually throw a partially-accurate address even further off course. 

BAD
GOOD


Apartments, suites, and floors - This one is tricky. Google says to not send them across if it’s avoidable. However, Salesforce doesn’t make it easy to store them somewhere else. Best rule of thumb: make sure they are officially part of the address and keep the notation as simple as possible (Example: #3 vs Apartment Number 3).

Hints or landmarks - Descriptors like “third house from the blue barn” may be useful for someone driving to a rural property, but the geocoder won’t be able to parse it.

C/O, ATTN or other mailing notes - The street address is not the place to store mailing information other than the address. It confuses the geocoder and degrades geocode quality. The smarter approach would be to store it in a custom field directly above the address.

Follow your post office - When you’re confused or just not sure, most postal services publish their preferred format for addresses. Locate the format guidelines on the country’s postal service website. The Google geocoder tries to follow these guidelines as much as possible.

IMPORTANT NOTES:
Keep in mind that even if Google Maps can find the location, this doesn't always mean the geocoder will also find it. 

Please also note that while Google has amazing coverage worldwide, it's not perfect. Geopointe data is only as good as Google data.

If you have problems or struggle with getting good geocode readings, you can always reach out to our support team. We’re happy to help, to troubleshoot, and do what we can to improve your geocode readings.