Conversation Merge: Merge data from multiple Conversations

Overview

As a moderator, you can merge data from up to 10 conversations directly on-platform without having to set up Merge Tags in the Excel discussion guide template. The system will automate the mapping process, letting you focus on reviewing the mapping and making edits as needed. Learn how it works below:

A maximum of 10 conversations can be merged at one time.
On-platform Merge is available for all Remesh question types.
On-Platform Merge is available for the following languages: Chinese (Simplified), Dutch, English, French, German, Italian, Japanese, Portuguese, and Spanish.

How it Works

Create a billable or non-billable Conversation.
Build your discussion guide using any of the options listed below. Note that you do not have to use the Excel template, as Merge Tags are no longer required to merge conversations.
When you’re ready, continue creating the additional sessions. The fastest option we recommend is duplicating the first conversation. However, you can use any of the methods listed in the previous step.
After all the sessions have ended, open one conversation and navigate to the Analysis tab on the left. Across the top of the page, click on the Merged Conversation tab. This conversation is your Anchor. Conversations you select to merge will be compared and mapped to the Anchor.
Click on Create Merged Conversation. The module will list all of the completed Billable conversations stored in the Team you’re working in.
Select the conversation(s) you want to merge from the drop-down list and add a name for your Merged Conversation. Then, click Create.
Now, you can see the conversations side-by-side and the automated mapping. Review and make edits to the questions. Changes are auto-saved if you need to exit and come back.
Click Merge Now to start the merge process; it should take a few minutes. You can monitor the status of the merge from this screen.
After your merge process is complete, click View Results to see your new merged dataset.

Frequently Asked Questions

Where can I access my merged Conversations?
After you’ve merged Conversations, you can access the merged data on the Analysis page of any of the individual Conversations that were included in the merge. For example, if you merged Conversations A and B, simply navigate to either Conversation A or Conversation B and then click on the Merged Conversations tab to access your merged data.
What happened to the “Export merged CSV” function?
The new merged Conversation analysis features no longer support the legacy “Export merged CSV”. However, you can still use the standard CSV exports in the upper right corner of the merged Conversation Results tab to retrieve these files. Data for individual Conversations will be represented as new Segments in the files.
What Conversations can be merged?
You can merge Conversations completed on or after August 1, 2021.

Some Conversations will not be eligible for merge due to language incompatibility. For those conversations, no merge will be possible, not even for Poll data. You can review language compatibility here.

Why can’t I access all the features for cutting and visualizing data on my old merged Conversations?
Conversations that were merged prior to July 10, 2023 do not have access to the latest updates, including the AI-powered Summarize feature. To access these features, you’ll need to perform a new merge.
Why is there a new question added after I merge Conversations?
Remesh automatically adds an additional Poll question: “Which conversation did you participate in?” This is added to enable you to filter your data by Conversation during your analysis.
Why is the "Merged Conversations" tab grayed out?
The Merged Conversations tab is only active within conversations if there are other eligible conversations available to merge in.
What if I want to merge multiple Conversations, but I want to add some additional poll or category options to questions in one or more of the Conversations?
The merge will not be disrupted if you add additional poll or category options as long as the order of the original poll options or categories does not change. We would recommend adding any new poll or category options to the end of the options and not move the original options/categories.
If I select the setting to add "Other" or "None of the Above" or enable randomized poll options, will that affect my merge?
No, Remesh takes these settings into account when merging Conversation data.
Why is there a difference between the merged and individual % Agree scores?
A few factors may drive the difference in % Agree scores.
First, individual % Agree scores and ranges differ from conversation to conversation, and population to population. This depends on the number of participants and the kinds of responses fielded. The new population defined in a merge is no different (the merged population is different from the population of any individual source conversation).

Second, the scores for the merged % Agree account for similarities in responses for a new, larger set of overall responses and not just the responses within the source conversation. This means that responses with low appearance and agreement in one session may have similarity to responses that have higher appearance and agreement in other sessions. Likewise, responses with higher appearance and agreement in one session may have similarity to responses that have lower appearance and agreement in other sessions.

Both of these scenarios will result in some overall adjustments in the merged score.
How does Remesh calculate agreement scores for merged sessions across multiple languages?
Participant responses are prepared for the % Agree algorithm by a sentence encoder, which produces mathematical representations of each response. Since this encoder is natively multi-lingual, responses from multiple languages will produce mathematical representations that are directly relatable to each other which means that responses with similar meaning across multiple languages will produce similar mathematical representations. Responses in multiple languages are fully understood by the encoder, which then inputs responses into the % Agree algorithm and produces the % Agree metric as usual.
When merging across multiple languages, will I see the end results in one language?
Our platform does not translate participant responses, and a merged conversation across multiple languages will contain responses in multiple languages. As mentioned above, we use an encoder in conjunction with our % Agree algorithm. We do not use this encoder to display or translate participant responses. The encoder essentially transforms the text into mathematical representations, not text. If you would like to see multi-language merged conversations in a single language, responses will need to be translated off-platform.
How does Remesh’s % Agree algorithm take into account that the number of participants between the merged conversations are different?
When merging conversations, the algorithm accounts for both the new total population size and the totality of responses.
What Analysis features will I be able to use in my merged conversation?
You will have access to our full suite of Analysis tools including Summarize, Thematic Clusters, Sentiment Tags, and more.
Are Segments included when merging a conversation?
Yes, Segments with similar names will be automatically added to your merged conversation. Further management of Segments can be done in the Segment tool found at the top right of your conversation.