Pros and Cons#

What are the Pros and Cons of using Global Address Public Cloud vs On-Premise?#

In terms of the global address verification functionality, the Public Cloud and the On-Premise both will product the same verification results. Our public cloud offer is after all simply a web application layer built on top of the On-Premise API. You will receive the same global coverage regardless of which one you pick. The main reasons for picking one or the other are due to architectural as well as organization policy reasons.

Public Cloud


Privacy and Security

Data is hosted on Melissa servers in a secure and private environment. We undergo security audits like SOC2 and no PII data is stored.

The most private and secure option. The Global Address library and data files are located on your own servers and nothing leaves your network.


Start verifying addresses with no setup or installation. No maintenance of servers or data updates to worry about.

Updates provided on a quarterly basis that will need to be updated on all machines hosting the on-premise product. Install is included or simply use a copy/replace operation.


Transactional Pricing. Best and cheapest way to get started and only pay for exactly what you use.

Bulk pricing. We are required to charge bulk pricing for putting data on premise. Most cost efficient for high volumes.

Ease of Use

Ready to use web application that supports REST/XML/JSON/SOAP on a globally hosted redundant infrastructure.

API for use in a programming language like Java/Python/.NET, etc. A low-level API that can be integrated to exactly your needs, but is not a ready to go web service like Public Cloud.

Global Address Information Types#

What type of global address information can I send?#

Global Address is designed to take full standard addresses. This means that the address must, in most cases, have a house number, a street (or thoroughfare) and identifiable area data (like postal code and locality). The service will not behave particularly well for partial addresses, point of interest information (like a park or plot of land without a formal address), Directions (e.g. “across the street from ABC Bank”) or non-address information (like phone numbers). The reason for this is that address verification relies on having a good address data source behind it in order to confirm its validity, and non-formal address data is not good enough to power our verification solution.

Minimum Required Input#

What are the minimum input fields required?#

The only field that is really required is the country input. We cannot verify an address without a country. Secondly, almost all addresses must effectively have an AddressLine input. While it is theoretically possible to verify an address with a very unique piece of information like a postal code in certain countries and areas, that is very rare.

Input Data Format#

How should I send in my address input data?#

Your input address can be sent in a number of ways. Our main advice is to try and send the data you have with as little manipulation as possible if you have a full address. The most common ways to send it are:

  1. Full address in the AddressLines:

    AddressLine1: 22382 Avenida Empresa
    AddressLine2: Rancho Santa Margarita CA, 92688
    Ctry: US
  2. Full address in the AddressLines + Parsed Area:

    AddressLine1: 22382 Avenida Empresa
    Locality: Rancho Santa Margarita
    AdministrativeArea: CA
    PostalCode: 92688
    Ctry: US

    Note: What you want to avoid as much as possible is sending in duplicated information, like this BAD example below:

    AddressLine1: 22382 Avenida Empresa
    AddressLine2: Rancho Santa Margarita CA, 92688
    Locality: Rancho Santa Margarita
    AdministrativeArea: CA
    PostalCode: 92688
    Ctry: US

Encoding Type#

What encoding does Global Address use?#

Global Address uses UTF-8 Unicode encoding. It is very important to make sure your data is in UTF-8 when sending to our service, especially for non-Latin countries. Be on the lookup for question marks (?), squares (▖) or other weird characters like �. They may be an indication of encoding issues and may result in data loss.

Address Verification Levels#

Why are there different levels of Address Verification Result Codes?#

The address source data that is available to use as part of our Address validation product can differ from country to country. For most developed countries with a major postal agency like the United States, Great Britain, France, etc, we will have delivery point data available. However, there may be countries where that detailed data is not available, or simply does not exist. In those cases, we will use less detailed information. Here is an overview and inside peak of the types of data available:

  • Delivery Point Data

    Source: 1 Main St Apartment 12, Anytown 12345, USA Max Verification Level: AV25

  • Range Data

    Source: 100-200 Main St, Anytown 12345, USA Max Verification Level: AV24

  • Thoroughfare Data

    Source: Main St, Anytown 12345, USA Max Verification Level: AV23

  • Locality Data

    Source: Anytown 12345, USA Max Verification Level: AV22

Service Speeds#

What can affect speeds from the Global Address Web Service?#

The response time of Global Address Web Service is highly dependent on a number of factors:

  • The country of the input. This is one of the most important factors as each country has its own engine in the background and its own verification paths and algorithms. A well-structured country like the US will be much faster than a less structured one like India. Also, a Latin-based address system will usually be faster than a non-Latin system.

  • The quality of the address. A good address will be faster than a bad or partially bad address. Bad addresses will require extra steps, additional fuzzy matching logic, and extra parses before we either find a match or give up.

  • Real time vs batch. Real time means sending one record at a time, while for batch you will send up to 100 records at a time and loop through your data. Processing 100 at a time will increase your overall per address speed by up to 25%-50%.

  • The location of the client server. The distance between our server to yours will make a difference in overall speed. See here for our current server location list: https://www.melissa.com/company/server-locations-uptime

  • The current load of our public cloud servers. We maintain and make available a large amount of processing capacity. However, all things being equal, the more usage at any given time will decrease overall speed during that time.

  • The protocol used. We have seen SOAP be up to 10% slower than XML or JSON.

Expected Throughput Speeds#

With all of these factors, we cannot give a single throughput number that will cover all the different use cases and types of input. Our expectation is that the majority of users will see varying speeds dependent on the contents of their records and their specific circumstances.

Single Record Address Validation


Single verification for on demand address validation

Throughput Range

50k - 120k records per hour

Use Cases

Form entry Point of entry verification Onboarding clients

Single Threaded Batch


Batch processing for small to medium volume processing

Throughput Range

100k - 400k records per hour

Use Cases

Overnight batch processing Cleansing small-medium sized lists

Multi Threaded Batch


Batch processing for high-volume processing when throughput is important

Throughput Range

300k - 800k records per hour

Use Cases

Cleansing large lists Data integration and ETL processes

As seen in the figures above, multiple threads will improve performance. We recommend starting with 10-15 threads for processing large batch lists.

Response Times#

Response times can vary based upon the postal quality of the country, script, and input quality. For example, validating addresses in Middle Eastern and Asia Pacific countries can cause response times to increase due to the complexity of parsing and validating non-Latin character sets. In contrast, response times in the United States and Canada are expected to be lower due to high postal quality and the Latin-based addresses.

Response times are not guaranteed, but we can provide a general range:


Response Times

Latin Based Countries

30ms - 150ms

Non-Latin Based Countries

50ms - 500ms


Melissa utilizes dynamic scaling to improve throughput speeds during peak usage periods. This behavior ensures that our web services can adjust to changes in traffic and consistently provide optimal performance regardless of demand.

If greater-than-usual volume is anticipated, please consult with your sales representative to ensure that Melissa can work with you to successfully manage your campaign.

Script Handling#

How does the Global Address engine handle different Scripts?#

Global Address has an option called OutputScript that has 3 possible values:




We will detect the script of the input and leave output in that same script.


We will change the output to Latin on output.


We will change the output to the native script of the country (Like Cyrillic for Russia)

Here are some things to make note of when trying to understand this functionality.

Global Address essentially supports up to two scripts per country. Latin and if the country uses a language that is not Latin based, we convert to that language if Native is set. Note that Script and Language are not the same. English, Spanish, French are all Latin Script even if their alphabet and diacritics used are slightly different. We can only change scripts for a record if we are able to verify and validate the address. For Latin based languages, if you specify OutputScript=Latn, we will also remove the diacritics (Gjøvik vs Gjovik)