Best time to send

Best Time to Send (BTS) provides a prescriptive time slot to send a campaign to the user to increase the chance of interaction or user response. The following are the features of BTS:

  • BTS is defined within MoEngage at a user level.
  • Marketers send campaigns or communication to users at a time when the chances for users to respond and interact are higher.
  • Marketers leverage advanced intelligence and personally target each user at the most preferred time slot instead of sending a communication to all users at a fixed time.

How does it work?

BTS uses advanced data analysis to analyze historical user interactions with the campaigns and channels. The analysis is done separately for Push, Email, and SMS & RCS, and the corresponding best time is determined.

The best time value for each channel is an hourly bucket between 0-24 hours and is available for individual users.  For calculating the BTS for each channel, the following events are considered:

  • Push activity events: All events performed by the user
  • Email activity events: Opens, Clicks, Viewed in browser
  • SMS & RCS activity events: Clicks

For a user who does not have any activity events for the Email, SMS & RCS channels, all the events performed by the user will be used to obtain the BTS hence it will be the same as the best time value for the Push channel. Hence for some users, you may not be using Email/SMS & RCS channels, but MoEngage still calculates Email/SMS & RCS BTS, which will be the same as Push Best time.

Configure campaigns to target users at their best time

You can configure the best time option in a One time or Periodic campaign for Push, SMS & RCS, and Email channels. 

To set BTS,

  1. Navigate to step 3 of the campaign creation flow - Schedule and goals.
  2. Select one of the options: At specific date and time or Periodic.
  3. Select the Best time for user.

For more information, refer to

Example

mceclip0.png

If you select the Start time as 6:58 PM and want to send a message based on the best time till 9:58 PM then this defines your BTS window. All the users in the segment whose best time lies between 7 PM and 10 PM will get the campaign at their respective best time as explained below:

Assume that there were 100 users in the segment and distributed as explained in the following table:

Total users in Segment: 100 

 

 

 

Of 100 users, the users whose BTS is available: 75


 

Number of users whose best time is within the BTS window (6:58 PM to 9:58 PM): 50

25 users have the best time as 7 PM

The campaign will be sent at 7 PM

15 users have the best time as 8 PM

The campaign will be sent at 8 PM

10 users have the best time as 9 PM

The campaign will be sent at 9 PM

Number of users whose best time lies outside the BTS window (6:58 PM to 9:58 PM): 25

Depending on the option you choose under "If user’s best time is outside this time window" 

  • If you choose "Send message at start time or end time", let us assume

15 users have the best time before 7 PM

The campaign will be sent at 6:58 PM (start time)

10 users have the best time after 10 PM

The campaign will be sent at 9:58 PM (end time)

 

  • If you choose "Do not send message" then these 25 users will not be targeted 

Of 100 users, the users whose BTS is not available: 25

Depending on the option chosen under "if user's best time is not available"

  • The campaign will be sent either at the campaign start time of 6:58 PM
  • At the app's best time. App's best time is explained in detail below
  • Do not send a message

Filtering users based on their best time

You can also segment the users based on their best time value. The BTS attribute is available in the User Property dropdown, under the Predictive Insights (Sherpa) section.

 mceclip1.png

 

How is BTS calculated?

The Best Time to Send is calculated for each channel by:

  • Aggregating all specified user events in the last 60 days. MoEngage Intelligence (Sherpa) automatically aggregates the user events by day of the week and the hour within a day to determine the most active time of a user.
  • Weightage is given to the most recent user events and activities than the ones performed earlier. 
  • User best time is computed by considering both the user's intraday activity and the days on which the user was active. Also, all users who have not performed any email or SMS & RCS activity will be populated with push best time by default for the channel best time.

Period for BTS Calculation

  • Best Time to Send value will be generated for any user who has performed an activity in the last 60 days
  • BTS values are updated for all users once a week.

Apps Most Active time

The time when all users of the app are most active. App Most Active Time is captured for different channels.

  • Push app best hour: Hour with max no of users performing activity irrespective of the channel.
  • Email app best hour/ SMS & RCS app best hour: Hour with max no of users performing channel level activity, that is activity related to Email or SMS & RCS. For emails and SMS & RCS, if there is no channel-level activity in the past 60 days, the push app's best hour will be shown by default.

How are Apps Most Active Time calculated?

App Most Active Time is calculated by aggregating the hour in a day when the maximum number of users perform activities. 

Limitation

BTS is not applicable for a triggered campaign because the sending for a triggered campaign is governed by when the user performs the campaign trigger event and the delay between the trigger event and the campaign sending time, set by you.

Best Time to Send is available by default for accounts with the enterprise plan. Contact your account manager or the MoEngage Support team if you wish to enable this for your account.

To contact the MoEngage Support team, you can raise a ticket through the Support Web Form within the MoEngage dashboard. For more information, refer here.
 

FAQs

arrow_drop_down 1. what are the specific data points used to measure user engagement across different channels (email, push notifications, SMS & RCS)? For example, does this include metrics like open rates, click-through rates, or time spent on content?
  • For Email: we consider a function of Open and clicks. The algorithm works on absolute stats of open and clicks count and not the open or click rate.
  • For Push, SMS & RCS: we consider clicks. The algorithm works on absolute stats of clicks count and not the click rate.
    We do not track information around time spent on content
arrow_drop_down 2.What are the specific data points (attributes or events) considered in the context while referring to "all outbound engagement with users"? 

While calculating BTS for Push, we consider click stats on Push notification + any app/web activities tracked via SDK.

Following are the exact events:

  • Notification Clicked Android
  • Notification Clicked iOS
  • Notification Clicked Web
  • Any custom event tracked via SDK

While calculating BTS for Email or SMS & RCS, we consider engagement stats on respective channel i.e.

  • Email Opened [Applicable for Best Time for Email channel]
  • Email Clicked [Applicable for Best Time for Email channel]
  • SMS Clicked [Applicable for Best Time for SMS channel]
  • RCS Clicked [Applicable for Best Time for RCS channel]

In case of Best time in Email or SMS & RCS nor being generated for a user, the Best time of the Push channel is served as the fallback.

arrow_drop_down 3. What are the SDK- events tracked in the BTS calculation?

Every SDK-tracked event participates in BTS calculation. The SDK-tracked events give us an estimation of users being active on your platform. The type of event does not matter here.

arrow_drop_down 4. How are the estimations for Best Time to Send and Most Preferred Channel calculated? 

Best Time to Send

The methodology employs a statistical technique that customizes the time by incorporating factors like recency, event frequency, and significance of the activity period in influencing user behavior. This approach involves two main elements:

  • Activity Concentration: This measures the event frequency over a historical timeframe.
  • Consistency of Activity: This tracks the number of days a user has been regularly active during a specific hour.
  • A time delay factor is incorporated to place greater emphasis on more recent interactions, and the model analyzes all types of user activities from the past 60 days to determine the optimal timing while excluding non-user actions or external triggers such as notifications received, emails sent, and SMS & RCS deliveries. Each event considered in this analysis is given equal importance in the final calculation. Using the factors mentioned earlier, a score is computed for each hour of the day, and the hour with the highest score is identified as the best time for each user.

Most Preferred Channel
The logic involves calculating the number of messages received by each user on each outbound channel (Push, SMS & RCS, or Email) and applying decay to prioritize recent messages. The total messages received by each user and the engagement metrics are summed up. Based on the engagement rate calculated for each channel at user level over the past 60 days of data, the channel with the highest engagement rate is identified and assigned as their most preferred channel.

The following events are considered:

  • Communication sent by you
  • Notification Received Android
  • Notification Received iOS
  • Notification Received Web
  • Email Sent
  • SMS Delivered
  • RCS Delivered

Communication engaged by the end user

  • Notification Clicked Android
  • Notification Clicked iOS
  • Notification Clicked Web
  • Email Opened
  • Email Clicked
  • SMS Clicked
  • RCS Clicked
arrow_drop_down 5. What validation methods are in place to ensure the accuracy and reliability of these estimations?
  • Assessing the campaign performance is the best way to understand the accuracy
  • For Best Time to Send, we generate a metric within the campaign analytics page that compares the campaign engagement between the BTS users and users without BTS
  • For Most Preferred Channel, we do not generate any metric currently.
arrow_drop_down 6. How do the models handle edge cases or anomalies in user data? 
  • Push channel: Push channel uses Push notification engagement as well app activities. The app activities are playing the role of data enrichment here
  • Email , SMS & RCS channels: These channels use the respective channel engagement alone. If some user has no engagement data available in these channels.

If still there are users with no data, you can manually define the fallback the campaign level from the following options.

  • App's best time: This represents the best engaged hour of the day across all users.
  • At fixed time: The campaign will be sent at the start time of the BTS window defined in campaign settings.
  • Do not send: The campaign will not be sent to those users for whom BTS is not calculated.

 

 

Was this article helpful?
1 out of 2 found this helpful

How can we improve this article?