Although Dialog Insight can generate messages very quickly, up to many millions of messages per hour, the final performance of your campaigns relies really on the message delivery speed, that is the time it takes for Dialog Insight to deliver your messages to your clients' email provider.

Performance depends on various factors, from your reputation, the size of messages, the number of IP addressed used for delivery and the distribution of domains in your contact list.

To better understand how delivery performance is set, here are a few examples to illustrate the impact of these factors.

Delivery limit per domain

Each domain you deliver messages to accepts them at a maximum rate set for their specific domain.

Sometimes limits are technical (such as the limited capacity of a small provider), but most often they are controlled and imposed by important email providers like Hotmail, Gmail and other similar providers. These providers limit message acceptance speed for each source.

Delivery of a standard message could look like this:

In this example, we can see a plateau effect. To understand how this happens, you need to understand how contacts are distributed.

Often, 50% to 75% of your list's addresses are distributed between 3 or 4 majors providers. The other addresses in your list are distributed to hundreds of different domains, each corresponding to a very small number of emails.

The stable level is directly related to this distribution:

  • When starting to send messages, Dialog Insight connects at the same time to all different domains and delivers the messages for their recipients
  • And because the number of contacts in these domains are limited, delivery is performed very quickly
  • Then, each plateau corresponds to the end of delivery of one of the major providers

Some domains accept messages at a very low speed, so the last few percentages of a campaign can take many hours to be completed.

Delivery limit per source

Major email providers apply rules that limit the reception speed of emails originating from the same source, typically, IP address.

The following example shows messages delivered to gmail.com addresses, where only one IP address is used for delivery:

Gmail's delivery speed control can clearly be seen here - about 1000 messages are sent every 5 minutes, that is 12000 message/hour.

This second example shows the delivery process of the next message, after having adjusted configuration to send messages from 4 IP addresses:

We can see here a linear increase to 4000 messages every 5 minutes, that is 48000 messages/hour.

If you use Dialog Insight's shared IP addresses, a significant number of addresses is usually assigned to your messages in order to maintain a good delivery performance.

In the event you use dedicated IP addresses to optimize your reputation, it is important to understand how these factors affects delivery performance - performance depends both on contact distribution by domain and the number of dedicated IP addresses you have. So depending on your performance needs, the size of your contact list and the distribution of your contacts, you will need to plan the appropriate number of IP addresses.

Reputation impact on performance

Your sender's reputation is critical for your success. Your reputation affects the delivery speed of your messages and their final placement (inbox or spam).

For the performance aspect, here is an example with Yahoo. Note that Yahoo manages the emails of other companies (like Rogers.ca), so this example also includes these other domains.

In this example, we see a very low success rate for Yahoo. After more than 12 hours, not even 25% of messages were delivered:

Why?
These are temporary interruptions set by Yahoo. What we see is clear - each hour a few messages are accepted, then Yahoo blocks the mailing for 1 hour. Yahoo applies very advanced reception rules on all messages. These rules are based on client engagement and behaviors to determine how messages will be delivered.

Contacts who immediately delete their messages without reading them, those who have reported spam messages and those who never open messages are all factors to take into consideration. And so is how complaints are managed. For instance, are you still sending message to contacts who have reported spam? When the decision to slow down the delivery of a message, Yahoo automatically blocks the reception of the next messages for a short period (in this example 1 hour).

It is impossible here to eliminate the problem completely with a solution that is only technical, as the problem relates to behavior issues. It is however possible to reduce the problem by taking some actions.

First action: Register to a feedback loop service (FBL)

Yahoo offers a feedback loop service for complaints (FBL / Feedback Loop) that allows the sender to receive a report when a recipient uses the option "This is spam" for any of your messages. Dialog Insight can receive these reports and automatically disable the related contacts in your contact list.

Using the FBL service reduces the number of complaints in your next mailings, and therefore helps to generate better results. But this service does not work instantly as it can take some time to completely clean up your contact list. But with time, the number of spams should reduce and even disappear.

Dialog Insight offers the implementation of Yahoo's FBL service when your account uses a dedicated IP address. Contact your account manager to get more information about FBL.

Second action: Manage client engagement

A second factor to take into consideration is Yahoo contacts' engagement (and of all your other contacts). For example, stop sending message to contacts who don't interact with your messages (contacts who have not opened or clicked in a message for a few months). A re-engagement campaign might be a better way to validate those contact's interest, and if no interest is still shown, you should remove these contacts from your list.

The result of such an action will provide your with improved client engagement, and less risks of having Yahoo block deliveries.