Find answers and general information quickly about Optimove’s Streams.
Setting Up a Stream
When should I use the Stream Builder instead of the Campaign Builder?
The Stream Builder enables you to build a series of interconnected campaigns, called a Stream treatment.
If you want to test between different strategies, the Stream Builder enables you to experiment with up to four treatments. For example, you can test one treatment using only email campaigns, one treatment using only SMS campaigns, and another treatment using a combination of email and SMS campaigns.
Since campaigns within a Stream are measured as a group, it helps to easily analyze the impact of different strategies within a customer journey.
How do I set up a campaign to be evaluated using Streams?
Using the Visual Stream Builder – via the Create Experiment button.
Can I evaluate any type of campaign using Streams?
Yes. Any campaign can be included in a Stream.
Can I select more than one objective for a Streams experiment?
No, you need to select a single objective when setting up a Streams experiment. Having said that, you will always be able to analyze the performance across all campaign KPIs, regardless of which objective you selected.
What does the LTV metric under the ‘AI Insights Agent (formerly Optibot) Recommendations’ section in the Streams Builder refers to?
This metric reflects the LTV for each new customer that you will acquire. Specifically, this is the Future Value plus the Past Value of customers who made their first order during the past period. This metric can also be found in the Optimove Dashboard.
Which execution channels can be tested with Streams?
Streams can be used to test campaigns delivered via all execution channels.
How long should each Streams experiment run?
With Streams, we recommend using longer experiments (and not shorter ones, such as onboarding campaigns). The recommended duration depends on the size of the targeted group and their response rates to the different campaigns; as a baseline, try setting upstream experiments to run for at least one month.
Can I use an existing target group in a Stream folder?
Yes. You can drag and drop an existing target group into your Streams folder. The target groups will immediately inherit the default criteria of your stream.
How do streams affect campaign prioritization?
They don’t. When working with Streams, you prioritize campaigns as usual, using the standard target group prioritization mechanism.
Are the customers included in a Stream’s campaign excluded from campaigns running outside the Stream?
Yes, as long as you utilize Optimove’s exclusion capability.
Can I test more than two campaign series variants within a single Stream?
This capability will be available soon. In the meantime, you can test two variants per experiment.
Is there a limitation on the number of splits or decision points I can use in a Stream?
The number of elements in a stream, including campaigns, flow controls and splits cannot exceed 150.
If a campaign in the stream has a 3-day duration and an attribute split is added (e.g. players who make a deposit exit), how will the duration of the campaign affect the user?
The campaign's 3-day duration affects the players like any other campaign duration in Optimove. If an attribute split causes certain players to exit the stream, they are effectively removed from the current stream and its associated campaigns but will be locked in the campaign for the set duration.
What happens if two users make edits at the same time to the same Target Group or Stream?
It is possible for one user's changes to be overwritten by another's. If two users are editing the same Target Group or Stream simultaneously, the system will save whichever changes are made last. The final state of the Target Group or Stream will reflect the last set of changes saved. Therefore, it is recommended to coordinate edits to avoid unintentional overwriting of data.
What is the naming convention for Target Groups generated by a Stream?
The naming convention for Target Groups generated by a Stream follows a structured format that reflects the Stream's Treatment and Campaign. The general formula will include: Treatment name Campaign 1, Treatment name Campaign 2, and so on. For example when creating a loyalty Stream with 2 Treatments for high loyalty users and low loyalty users the naming will be : loyalty flow- high- campaign 1 / loyalty flow- low- campaign 1.
Is the merge Feature for Campaigns restricted to the same Treatment?
Yes, the Merge feature is only available for Campaigns within the same Treatment. Treatments are separated by random customer slices for measurement purposes, so merging Campaigns across different Treatments is not supported. This ensures that the integrity of the measurement and analysis for each Treatment is maintained.
Can the KPI in the response split be changed?
No, it’s based on the default KPI which is configured to your Optimove instance.
I duplicated a stream, made some changes, and tried to relaunch it, but I keep getting an error saying "Scheduled time has passed for OptiMail Campaign." Why is this happening and how can I fix it?
A. The error occurs because the duplicated Stream's analysis date range includes past dates. To resolve this issue: 1. **Update the Analysis Range:** Change the analysis date range to future dates. This will ensure that the system schedules the campaign correctly. 2. **Same-Day Campaigns:** If you plan to run the campaign on the same day (e.g., today), make sure the scheduled time is also set to a future time on that day. The system requires that all scheduling times be in the future to execute properly.
What are the response split options based on?
The response split is based on the KPI selected in the stream
What is the maximum number of days allowable for a stream's analysis date range?
There is no limit to the number of days you can set for a stream's analysis date range. You can specify an extensive range based on your analysis needs.
How do I target a Stream's Control Group or Treatment if they never received a campaign so they cannot find them using Campaign History?
The only way to identify and target a stream's control group or treatment is by exporting the data from the stream's analysis
Can a deleted stream be recovered?
No.
Managing Streams
When using the "control" treatment (super control) in streams, are users restricted from receiving campaigns outside of the stream? If so, for how long?
Yes, users are restricted for the duration of the entire stream . The users will receive only "include all" campaigns.
they can only get 'include all' campiagns
Do you recommend having a control group for every campaign within a Streams experiment?
Yes. To allow an analysis of each individual campaign in parallel with analyzing each stream, each campaign should use its own control group.
When using the Visual Stream Builder, can customers enter the stream flow via any target group in the stream?
No. Customers can only enter the sequence via the first target group in the flow.
When using the Visual Stream Builder, how can I make sure a customer will not enter the sequence more than one time?
In order to make sure a customer will not enter the flow more than once, define a Frequency Cap for your stream to limit the number of times a customer may enter a stream. Once defined, a relevant Campaign History criterion will be added to the stream's target group.
Can I edit a Stream once it is running?
You can edit the following elements of a Stream:
- Recurrence settings
- Campaign templates
- Promotions
What happens if I delete or edit a campaign during a Streams experiment?
You may freely edit campaigns, within the editing limitations, that are part of Streams experiments. Deleted campaigns will be ignored for the overall experiment analysis.
If you stop a running Stream, you will have the option to allow customers who are mid-stream to continue recieving the rest of the campaigns. This means no new customers will be targeted by this Stream, however you can complete the series of campaigns for customers who have already been targeted.
Or, you can choose to stop the Stream and it’s campaigns altogether.
I built my experiment using the Visual Stream Builder, and the amount of customers is decreasing. Why?
Using the Visual Stream Builder, the campaigns are dependent on the stream’s conditions and sequence. For example, if, for some reason, users did not qualify for receiving campaign #2 after being targeted with campaign #1, (reasons vary and could be related to users being targeted as control, or excluded by a higher priority campaign) these users will not proceed to campaign #2, and so on.
As a result, the number of customers that will complete a stream can only be the same or lower than the number of customers who started the stream.
How do I avoid all my customers being excluded in the middle of a Stream?
We recommend using manual prioritization for target groups in your Streams.
Manually prioritized target groups will always have a higher priority than target groups using auto-priority.
When changing criteria in the reference target group that's being used in a running stream, should the stream's related first campaign be changed correspondingly?
ou can't change the reference target group in a running stream. However, you can modify the criteria for Campaign 1, Campaign 2, and subsequent campaigns. Changes made to the campaigns will not alter the original reference target group criteria set at the start of the stream
If I create a completely new flow, how can I ensure that users from the old stream do not join the new one?
You will have to set it manually in the target groups of the new stream using the Campaign history attribute.
If a stream is already activated and in the 'running' status, can you make changes to future emails in the stream?
Yes, you can make changes in campaign details.
In a multi-campaign stream flow (e.g., campaign, wait X days, campaign) that uses control groups within each campaign, is it expected for customers who fell in the control group of the first campaign to be included in the second campaign?
No. When a campaign in the Stream uses a control group, the customers assigned to that control group are intentionally held back and not targeted by that campaign. Because they were not targeted, they do not gain the campaign history that the next campaign's Target Group requires — so they no longer qualify for the subsequent campaigns and are effectively removed from the remainder of the Stream.
Can I add an Exit Criteria to an Activated Stream if the campaigns are not running yet?
Once a Stream has been activated, you cannot add exit criteria afterward directly. If you need to incorporate exit criteria that weren't initially included, the current process requires duplicating the entire Stream and setting up the desired exit criteria in the duplicated version. This ensures that the new conditions are applied consistently from the start of the stream's execution.
Is there a way to completely delete old streams from the Optimove instance?
No, but you can use the filters on the streams page to manage them effectively.
How can I remove individual steps from an active Stream?
While you cannot directly remove a step from an active stream, you can effectively do it by reconfiguring the steps around the one you wish to remove. Here's how:
Editing the Stream
You can edit the following elements of a Stream:
- Reconfigure the Stream: Adjust the steps around the one you want to remove e.g. Analysis Range & Recurrence settings
- Update the Templates: You can update the templates as needed
- Promotions
Creating a New Stream
If reconfiguring the existing stream seems complex or unsuitable, consider creating a new stream. You can set up the new stream with the desired steps and updated templates, ensuring it meets your current requirements.
These methods allow you to manage and streamline the flow of your active stream according to your needs.
Why are users receiving the same emails on Consecutive Days?
No frequency cap was set, so the system keeps sending the campaign to qualifying customers, causing duplicate emails. To prevent this, set the frequency cap at the top of the canvas page to limit how often the same email is sent.
The Stream is marked as completed, but I still receive new user lists daily, and the Mission Control shows new users receiving messages from the first day. Why is this happening?
The Stream is still running because in the setting of the stream, the recurrence is set to end by a future date. The Stream is still active and continuing to send messages to new users who meet the criteria as the date did not pass. To stop the Stream and cease receiving new user lists, you can delete the campaign directly from Mission Control. This will prevent further messages being sent to new users.
Can I pause a Stream?
Currently, pausing a Stream is not possible. If you need to make changes or temporarily halt activity, you may need to stop the Stream entirely.
Why Can't I Edit the End Date for a Stream?
Once a Stream is running, you can only edit the following elements:
- Recurrence settings
- Campaign templates
- Promotions
However, you can either decide to Stop the stream on the day you wish for it to end altogether or delete the occurrences of the campaigns belonging to the Stream in the mission control, from the new desired end date onwards.
Can I edit the names of folders and target groups while a stream is running? Will this affect the execution of the stream?
Yes, you can edit the names of folders and target groups while a stream is running. Renaming folders and TGs does not affect the execution of the stream
For a single treatment stream with no control group, will the random customer percentage still be applied to the target group?
Even in a single treatment stream with no control group, the random customer percentage will still be applied to the target group. This means that the random customer percentage will include all 100% of the target group.
What is the duration of the Super Control Group in Streams?
The super control group is defined for each stream occurrence, not the entire analysis range. This means the control group includes the same customers only during each stream occurrence, even if the analysis range extends to 2050.
Is there a limit to consecutive attribute splits on the same day in Streams?
You can perform consecutive attribute splits of the same attribute on the same day. There is no strict limit on the number of splits you can apply in sequence. There are no hard limitation, but it’s recommended to not have more than ~80 elements.
How Does the 'Wait Until' Function Work Between Sequential Campaigns?
The 'wait until -1 day' function operates on a per-calendar-day basis, not a full 24-hour period. For example, if you have two email campaigns scheduled sequentially—one for today at 9 PM and the next for tomorrow at 7 PM — setting a 'wait until -1 day' will not make the system wait a full 24 hours before sending the next campaign. This means the second campaign will be sent on the next calendar day, even though only 10 hours have passed.
What are the numbers that appear on each campaign in the stream after it's been activated?
The number that appears on each campaign element (e.g., #4) indicates its sequential order within the stream.
- Automatic Assignment: The number is assigned automatically based on the campaign's position in the journey flow. The first campaign is #1, the second is #2, and so on.
- Target Group Naming: This number is also used in the name of the corresponding target group for that campaign, making it easy to identify which audience is associated with which step in the stream.
Analyzing Streams
How does Optimove determine the experiment’s winner?
In each stream, Optimove is tracking two metrics:
- Response rate (the rate of customers that took the desired action)
- Average KPI (a KPI specified by you)
The way Optimove calculates the winner is based on the two metrics’ statistical credibility.
Read more about the exact calculation here.
Are customers who enter the stream on the last day of the analysis range counted in the analysis?
While you can analyze each campaign these customers received, their responses will not be included in the Stream Analysis. You can match up the analysis range to when the occurrences are due to end to ensure the analysis covers the length of every occurrence.
Can I see initial results while a stream experiment is Running, or do I need to wait until it completes to receive an analysis? How do I know when there is a winner?
You can view interim experiment results on the Stream Analysis page even before the Streams experiment concludes, starting with the day after the first campaign’s execution. When the Stream experiment is complete, the AI Insights Agent will notify you if there is a winner.
Is it possible that the results of a Stream experiment will not be statistically credible?
Yes. Review the section "No-Winner Scenarios", for complete details.
Why is the "Group Size" so large?
The Group Size will differ based on the setup you choose.
- When using the Visual Stream Builder setup, the "Group Size" in the results table will contain all customers who qualify for the first target group in the stream, regardless of whether they were actually targeted with a campaign or not. For example, even customers in the control group for the first campaign will appear as part of the Group Size.
What happens if one of the campaigns within a stream never executes, for some reason?
Nothing. All not-run campaigns will be ignored when analyzing the experiment.
Will AI Insights Agent (formerly Optibot) still provide recommendations for individual campaigns that are included within a Stream experiment?
Yes, every campaign within a stream is also analyzed separately, this means each campaign will have a distinct Campaign Analysis report and in case it requires optimization, its own AI Insights Agent (formerly Optibot) recommendations.
How long are customers locked into the super control groups locked in for?
The duration of one full Stream occurrence:
The duration of an occurrence = all wait times between each campaign + the duration of the final campaign.
How can I extract the list of CIDs in the super control group of a stream
You can export the list in the Analysis page of the stream in the top 100 customers section.
If the stream is in the 'completed' status, how will it affect the analysis?
If a stream is in the 'completed' status, the analysis for that stream will also be considered completed. Any ongoing campaigns within the stream that were set to never end will not be included in the analysis once the analysis range concludes and the stream is marked as completed. The analysis will only encompass data collected up to the point when the stream's analysis range ends.
How do I monitor and analyze stream campaigns using Mission Control?
To effectively monitor and analyze streams campaigns using Mission Control, follow these steps:
- Grouping by Stream: In Mission Control, you can group your campaigns by stream. This allows you to analyze the overall performance of the entire series of campaigns associated with a particular stream. You can also drill down into individual campaigns to examine specific details and metrics.
- Monitoring Running Streams: For streams that are currently running, Mission Control provides real-time monitoring capabilities. You can check the status of each campaign to ensure they are executing as expected. This includes verifying that campaigns are running smoothly, tracking progress, and addressing any issues that may arise.
- Analyzing Completed Streams: Once a stream is complete, you can analyze its performance within Mission Control. This analysis helps you understand the impact of each campaign within the stream, allowing you to assess the effectiveness of the overall marketing strategy. Key metrics and insights can highlight areas for optimization and improvement.
- Comparing Streams: Mission Control also allows you to compare the overall performance of different streams. By analyzing metrics across various streams, you can determine which customer journeys are more impactful and successful. This comparison can guide future campaign planning and resource allocation.
When Does Stream Analysis Become Visible?
Stream analysis is available when the stream is in the status 'running.' You do not need to wait for the stream to reach a specific finishing spot or flag point. As soon as the stream begins executing and reaches the 'running' status, the analysis should be accessible.
In the stream analysis page under ‘Stream Summary’, why is there a difference in the Group Size in ‘Treatments’ and ‘Campaign List View’?
The difference in group sizes arises because they refer to different concepts:
- Stream Summary Group Size : This number represents the reference target group, which includes all customers who initially met the criteria for Treatment A when the stream was set up.
- Campaign List View Group Size This figure shows the actual target group that received Campaign 1. This group may be smaller due to several factors, such as exit criteria, customers being filtered out, or other conditions that exclude them from receiving the campaign.
What Does Optimove Consider as “Responders” in the Stream Results Section?
'Responders" are customers who match the conditions of the reference group criteria. This typically includes customers who are also in the first target group (TG) for the first campaign of each treatment. Therefore, "responders" are usually those targeted with the first campaign of each treatment, unless the criteria have been manually edited after activation.
How Do I Attribute Response Rate When Multiple Campaigns Overlap in Duration of the same stream using a control group?
The average uplift per customer. works for multi-touch, By going into Optimove’s Stream Analysis you can see the average uplift per customer for both campaigns proportionally. For example, if Campaign 4, 5, and 6 all contributed to the customer’s response, the response rate can be proportionally attributed to each campaign based on their influence.
No matter the combination of overlapping campaigns, if you are using Optimove’s control groups, you are effectively measuring multi-touch attribution. This approach helps in understanding the overall impact of each campaign within the context of the customer’s complete interaction history.
Does the 'Targeted Customers' field hold the sum of all stream occurrences?
Yes.
When deleting a stream, are the stream's campaigns visible in Customer 360?
If a campaign was executed before the stream was deleted, it will still be visible in Customer360.The deletion of a stream primarily affects the stream's configuration and scheduling, not the historical campaign data.
Can the toggle of the ‘series analysis’ can be turned off on the Streams analysis page?
No.
The stream analysis shows a statistically significant winner, but Mission Control shows 0 uplift for the stream's campaigns. Why is there a difference?
The discrepancy occurs because a statistically significant result in the stream analysis does not automatically mean that the uplift for specific campaigns is significant. Here’s why:
- Stream Analysis: Provides an overview of the entire stream, including the combined effects of all campaigns. It assesses the overall performance and significance of the treatment across the whole stream.
- Mission Control: Measures uplift at the individual campaign level. It’s possible for a treatment to be statistically significant at the stream level but not show significant uplift in every individual campaign.
In summary, a treatment may be performing well overall, but the uplift in specific campaigns may not always reach statistical significance