Analysis Overview

Introduction

Branch Analysis is a powerful, sophisticated analytics experience that takes plain English prompts and uses natural language processing to answer the everyday business questions that come up in your marketing workflows.

Analysis empowers you to quickly see the performance of your active campaigns, so you can effectively optimize your ad spend. By drilling down into specific campaigns, you can break out your data by platform, channel, and more, so you can see what’s really driving results.

Analysis also makes it simple to report out your campaign metrics to internal stakeholders, giving you the materials you need to lead important conversations around your company’s marketing strategies and growth.

Don't see this feature yet?

Analysis is the next evolution in Branch’s marketing analytics and reporting functionality, bringing together data from ads, social media, SEO, Journeys, Email, and more in one cohesive experience.

Analysis is being rolled out to customers in phases. While you're waiting for access, you can explore this article to familiarize yourself with the new functionality.

In the meantime, please continue to refer to our legacy documentation for reporting.

Key terms

The following terms are used extensively throughout this guide. Please take note of their definitions.

Packaging

The following Analysis functionality is available based on your Branch plan’s tier:

• Analysis dashboard creation and editing

• AI dashboard prompt widget

• AI on individual widgets

Please note that some packages do not include the ability to create or edit dashboards; in these cases, you will only be able to view Branch’s default set of dashboards.

For more information, please contact our Sales team or reach out to your Account Manager.

Use Analysis

Branch Analysis is both powerful and intuitive, so you can glean deep insights from the feature without extensive guidance. However, it is helpful to understand the basic functionality of Branch Analysis before you dive into intermediate and advanced queries.

This section will get you started using Branch Analysis quickly, and will use common marketing concepts to inform the types of questions used to prompt Ivy, Branch’s AI.

Dashboards

In Analysis, a dashboard is an individual page that contains a specific set of widgets, which themselves display metrics or charts.

Use each dashboard to tell a different story with your Branch data and to track your company’s marketing growth from different angles.

Default dashboards

Branch Analysis comes with a set of pre-configured dashboards that are ready for you to use immediately.

Part of this set is an important dashboard called Channel Performance, which you will see listed directly under the Analysis section in the left-side navigation of Branch.

The other part of this set is our classic suite of dashboards. These dashboards are meant to replace our legacy data reporting pages, and are useful jumping off points for anyone getting started with Analysis.

Caution

Please note that there is not a 1:1 relationship between legacy data reporting pages and the new classic suite of dashboards in Analysis. However, we have tried to maintain parity where possible.

Also, the default dashboards you see in Analysis will depend on your Branch package.

View dashboards

Under the Analysis section in the left-side navigation in Branch, you will see several of your core dashboards listed by name. Select any of these dashboards to view them.

To view a dashboard not featured in that list, select All Dashboards and select the dashboard you’re looking for from the full list.

From this page, you can also favorite a dashboard by selecting the star icon next to its name. You can then filter the full list by dashboards that are starred or not starred.

Dashboards that have the verified checkmark icon next to their names are dashboards that have been created for you by Branch. You can filter by verified or non-verified dashboards.

To view only the dashboards that you’ve created, you can also select the tab that says Yours at the top of the page.

Create a dashboard

To create a new dashboard:

1. Select Analysis in the left-side navigation in Branch, then select All Dashboards.

2. Select the Create Dashboard button in the top-right corner of the page.

3. Give the dashboard a unique name in the Name field.

4. Give the dashboard a description in the Description field that explains the purpose of the dashboard and how others should read it.

5. Select the Create button.

You are now ready to add widgets to your new dashboard.

Copy a dashboard

Sometimes you may want to use a certain dashboard as a starting point for creating your own version with different metrics. To do this, start by copying the original dashboard.

To make a copy of a dashboard:

1. Navigate to the dashboard you’re interested in copying, and select the three dot icon in the upper-right corner.

2. In the menu that opens, select Make a copy.

3. Give the dashboard a unique name in the Name field.

4. Give the dashboard a description in the Description field that explains the purpose of the dashboard and how others should read it.

5. Select the Create button.

Edit a dashboard

Dashboards can be edited, as long as they aren’t one of the dashboards created by Branch. You will see the author of these dashboards listed as “Branch” and they will have a verified checkmark icon next to their names on the All Dashboards page.

Also, please note that if someone else on your team created the dashboard, they will need to give you edit permissions before you can edit their dashboard.

To edit a dashboard:

1. Navigate to the dashboard you’re interested in editing, and select the three dot icon in the upper-right corner.

2. In the menu that opens, select Edit.

3. From here, you can:

   1. Add notes: Select the Add note button in the top-right corner to add a widget to the dashboard that holds a note you want to keep. You can move and resize this widget. You can also add multiple notes to a single dashboard.

   2. Add filters: Select the Add filter button in the top-right corner to add a new filter and get taken to its configuration menu. When you are finished configuring it, select the Apply button.

      1. To learn more about configuring filters, read “Configure a dashboard filter.”

   3. Edit the dashboard name and description: Select the pencil icon next to the name of the dashboard to give it a new name or description. Select Save to save your changes.

   4. Remove filters: The filters applied to your dashboard are listed at the top in individual bubbles. To remove a filter, select the X icon that appears when you hover over it.

   5. Edit filters: From the filters listed at the top of the page, select the one you want to edit to see its configuration menu. To learn more about configuring filters, read “Configure a dashboard filter.”

   6. Move widgets: Hover over the top edge of the widget, until your cursor becomes an icon with four arrows and the light gray bar appears. Select the light gray bar and drag the widget into place.

   7. Delete widgets: Hover over the widget and select the three dot icon that appears. Select Delete to delete your widget.

      1. Warning: Please note that deleting a widget is a permanent action and cannot be undone!

   8. Resize widgets: There are two ways to resize widgets:

      1. Hover over the widget and select the three dot icon that appears. Select the visual representation of the dimensions you would like the widget to change to from the set of options.

      2. Hover over the bottom-right corner edge of a widget until your cursor becomes an arrow icon. Drag the widget’s corner to make it bigger or smaller.

4. When you are done making changes to your dashboard, select the Save button.

   1. If you would like to discard the changes you’ve made, select the Cancel button instead. Then select Discard changes.

Configure a dashboard filter

Filters help make sure you’re only seeing the data you want to see on your dashboard. To set a filter at the dashboard level:

1. Navigate to the dashboard you’re interested in creating a filter for and select the three dot icon in the upper-right corner.

2. In the menu that opens, select Edit.

3. Select the Add filter button in the top-right corner.

4. A list of possible filters will open. Select the filter that you would like to add.

5. The filter’s configuration menu will open. This menu has four main sections:

   1. Default values: Set the values you want to either include or exclude from your dashboard for this particular filter. Depending on the type of filter, you may be asked to select values from a list, or add your own conditional statements for what values to filter in or out.

      1. Example: To see only your social media campaigns, selected “Feature” as your filter and only include data where the value is social.

         1. Please note that by default, selecting values for a filter will limit the values displayed in any future filters you create for this dashboard. In this example, if you tried to add “Branch Channel” as a second filter, you would only be able to select a channel that applies to social media campaigns, like Short Links. To change this, enable the Show unfiltered values option at the bottom of the Default values section.

   2. Applicable to: Decide which widgets you want to apply this filter to. You can select the checkbox for All, or you can select individual widgets.

   3. Link filters: Connect columns from different models that contain the same values (like campaign names, dates, or regions), so one filter can control data from all sources at once.

   4. Filter options: Choose a display name for your filter, and decide whether you want to make it mandatory. To make a filter mandatory, select the checkbox that says Require users to select a filter value. If this is selected, users will not see widget visualizations on the dashboard until they select a filter value.

6. When you are done configuring your filter, select the Apply button to apply your changes to your dashboard. You will then see your filter populate the list of filters at the top of the dashboard.

Delete a dashboard

Warning

Deleting a dashboard is a permanent action and cannot be undone.

Please note that dashboards where the author is listed as “Branch” cannot be deleted.

To delete a dashboard:

1. Select Analysis in the left-side navigation in Branch, then select All Dashboards.

2. Find the dashboard you want to delete and select the checkbox near its name.

3. If you have the correct permissions to delete the dashboard, you will see the Delete button appear after selecting the checkbox.

4. Select the Delete button to permanently delete the dashboard.

Widgets

Widgets are what make up Analysis dashboards. Each widget contains a key metric or visualization. Widgets can also be used to store your notes.

Use widgets to display important data related to the general theme of your dashboard.

Create and configure a widget

To create a new widget:

1. Navigate to the dashboard you want to add the widget to and select the blue button at the bottom of the screen that says Add new widget.

2. In the top bar, use the dropdown to open the data sources menu. From the menu, select the data source you want your widget to draw from. If you want to use multiple sources, select Enable multiple sources.
   

3. Choose the columns (filters) you want to add to your widget. The categories are:

   1. Measures: These have a measurable, numeric value. Primarily related to quantities of user events and amount of revenue.

   2. Attributes: Qualitative dimensions of your data, like Campaign Name or Platform.

   3. Dates: Check the Date box to include a date column. To configure this column, select the filter icon next to it. Then decide if you want your widget to display rolling data (changes as time passes) or fixed data (static, covering a specific period of time).

   4. Formulas: Add your own formulas to create your own custom calculations. Use either the free-form formula editor or the pre-built functions provided. Use the + Add button to create new formulas. Select operators you’re interested in from the list to learn more about how to use them.

      1. For example, use the formula Ad Partner Name = "Taptica" and contains ( Campaign Name , "spring" ) to create a filter for Taptica campaigns you launched in the spring.

   5. Sets: There are two types of sets:

      1. Column set: A set based off a single column’s data. Use the + Add button to create new column sets.

         1. For example, select “Branch Channel” as the base column, then create a group for paid channels and a group for organic channels.

      2. Query set: A set based off a query that uses one or more columns of data. Use the + Add button to create new query sets.

         1. For example, select “OS”, “Feature”, and “Campaign Name” for your base query, then create a group where OS = Android, Feature = paid advertising, and Campaign Name = spring_solstice.

   6. Parameters: Create a custom name and specify whether the data’s type should be integer, decimal, string, date, or boolean (true or false). Parameters let you decide which values you want to see included in your data, as well as what the default value should be. Use the + Add button to create new parameters.

4. For any columns you selected in step 3, you can add a filter to that column by selecting that column from the list. This way you can include or exclude the appropriate values for that column.

5. Select the Go button to run the data for your widget.

6. A suggested title for your widget will pre-populate for you, but you also have the option to change it to your own custom title.

7. Add a description to your widget to make it easier for others to quickly understand the data being displayed.

8. Your data will automatically be displayed in a suggested format, but you can also manually configure it as this point to be a table or a chart.

   1. If you choose the chart option, you will have the ability to select which type of chart you want your data displayed as by selecting the bar chart icon on the right side. Choose from a library of different types of charts, depending on your needs.

9. Edit the chart, table, or metric’s configuration by selecting the cog icon on the right side. There are two tabs:

   1. Configure: Determine which table columns to show/hide, or customize chart axes.

   2. Settings: Additional display and formatting options, like text wrapping and table theme.

10. Select the lowercase i icon on the right side to see exactly what the query that you’ve built in this process looks like.

11. Once your query looks as expected and your visualization is properly formatted, select the Pin button to save your new widget. You will be able to choose which dashboard you want to save your widget to.

View widget data

To view the underlying data that a widget is based on:

1. Navigate to the dashboard containing the widget you’re interested in.

2. Hover over the widget and select the three dot icon that appears.

3. Select Show underlying data.

   1. Please note that you cannot view underlying data for columns based on aggregate formulas.

4. Select the Edit columns button to edit the columns displayed in this view.

5. Select Download CSV at the bottom-left of the module to download the data.

Tip

Editing which columns are displayed when viewing a widget’s underlying data will also impact the columns included in the CSV download.

You can also explore widget data further without actually making changes to it:

1. Navigate to the dashboard containing the widget you’re interested in.

2. Hover over the widget and select the three dot icon that appears.

3. Select Explore.

4. From here, you can add filters, add columns, replace columns, or compare values.

   1. For example, you can quickly compare data where the “Ad Partner Name” is amazon versus snap to see which ad partner had more activity.
      

Edit a widget

To edit the data displayed in a widget or how it is displayed:

1. Navigate to the dashboard containing the widget you’re interested in.

2. Hover over the widget and select the three dot icon that appears.

3. Select Edit. For more information on widget configuration, read “Create and configure a widget.”

4. When you’ve finished your edits, select the Save button. Then select the Close button at the top of the page.

Display a widget on multiple dashboards

If you’ve already created a widget and you want to copy it to a different dashboard:

1. Navigate to the dashboard containing the widget you’re interested in.

2. Hover over the widget and select the three dot icon that appears.

3. Select Pin.

4. Select the dashboard that you want to add your widget to, then select the Pin button.

5. When you visit the dashboard you selected, you will see that your widget has been added to it.

Tip

Editing a pinned widget in one dashboard will not impact that same widget in a different dashboard.

Delete a widget

Warning

Deleting a widget is a permanent action and cannot be undone.

To delete a widget:

1. Navigate to the dashboard containing the widget you want to delete.

2. Select the three dot icon in the upper-right corner of the dashboard.

3. Hover over the widget and select the three dot icon that appears.

4. Select Delete to permanently delete your widget.

AI

AI is an important cornerstone of Branch Analysis. Your guide through this experience is Ivy, Branch’s AI.

Caution

Please know that Ivy’s answers are probabilistic, which means the same prompt may not always produce the exact same answer. Ivy’s answers may contain errors and should be verified.

We recommend that you manually confirm that the models, filters, formulas, etc. your widgets are based on are what you expect.

Get insights with Ivy

Ivy can answer your questions about Analysis data at both the dashboard and widget levels.

You can prompt Ivy from any dashboard or widget, with the exception of widgets used to store your text-based notes.

If Ivy can’t process the entirety of your prompt, it will let you know and attempt to process part of it.

Once Ivy gives you an answer, you can change the visualization type (table, chart), pin it as a widget to a dashboard, download the answer’s data, or edit the data.

Tip

Selecting the Edit button on one of Ivy’s answers also lets you see in more detail what models and columns it is using for the answer.

If Ivy gives you an answer that is a link, you can select that link to see more information about the columns it is using. In some cases, it may have created a new formula, set, etc. for you, which you will find in the left-side navigation along with the rest of the available columns. You will also see all the columns used listed at the top in the search bar.

If a blue line appears under a column name, you can select it to see alternatives suggested by Ivy.

At the end of every answer, you have the option to tell Ivy whether it interpreted your question correctly by selecting either the checkmark icon or X icon.

Dashboard level

A dashboard is the highest level you can prompt Ivy from, but you can still ask Ivy questions that reference data not currently in that dashboard.

Prompting Ivy at the dashboard level is recommended when you want to create a truly unique widget. Otherwise, we recommend starting with a widget that is similar to what you’re looking for, then prompting Ivy from there.

Widget level

When you prompt Ivy at the widget level, it will start by summarizing and re-iterating your question back to you. Use this output to confirm Ivy understood you correctly.

At this point it may also reference previous answers it has generated for you, to make sure it has built up the correct context between answers.

If Ivy makes changes to a previous answer, it will list these changes in shorthand form, for example Added:Feature,  Branch Channel = 'Short Links'. Make sure these changes are what you expect for your prompt.

General tips for Ivy prompts

Read the tips below and keep them in mind when crafting your own prompts for Ivy.

Start broad, get narrow

Instead of giving Ivy every detail upfront, ask a broader question first to establish a baseline, then build on the answers with more specific questions. Here is an example of a flow that uses this method:

Prompt 1: “How many users completed trials this quarter?”
Prompt 2: “Break that down by OS.”
Prompt 3: “Now just show me users whose OS is Android.”
Prompt 4: “How many of these users are in Australia?”

Use context

The more context Ivy has about your question, the better it will understand what you are asking it to do. This is the reason it can be easier to build new widgets by using widgets you already have, than by creating a widget from scratch.

If you already have a widget that is a good starting point for your question, select the Ivy button on that widget to prompt Ivy from within the context of that widget. It will use the original widget to inform its answer.

Speak naturally, but be aware of keywords

Ivy expects your prompts to be formatted in plain English, not using proper Branch column names. For example, you can ask Ivy “What are the names of my top 5 campaigns?” and not use campaign_name anywhere in the prompt.

However, there are certain phrases worth noting:

• “Channel” and “Branch Channel” are not the same column name. “Branch Channel” includes Journeys, Short Links, Paid Media, and SEO.

• “Paid advertising” and “non-paid advertising” are useful ways to refer to owned and organic channels, respectively.

• The term “rate” is generic based on the context, so try to make it more specific, like “click-through percentage”.

Specify time periods clearly

Be specific about the time period you want Ivy to look at when gathering your data. Instead of using a word like “recently”, say “in the past 7 days”.

Ask for comparisons

Ivy excels at comparative analysis. You can ask it to compare sets of data across different time periods, channels, campaigns, and much more.

Use numbers as filters

Ivy also handles numeric filters well. You can use phrases like:

• “Top 5”

• “Above 10%”

• “Higher than 80k”

Get started with Ivy

Use these prompts as jumping off points to start conversations with Ivy:

• Which feature is driving most of my Short Link traffic?

• Are there differences in audience demographics between Short Link and Journeys clickers?

• Show me installs and opens for my top 5 paid advertising campaigns this quarter.

• What are our email open rates by audience segment?

• Show me click-through percentages for our top 10 email campaigns.

• How many subscriptions did users start last quarter?

• What's our paid advertising versus non-paid advertising install split this quarter?

• How many purchase events came from each Branch channel this month?

• How many registration forms were completed per campaign last quarter?

• What's our install-to-purchase conversion rate by ad partner?

• How many attributed installs do we have per geographic region for the past 60 days?

• Compare iOS versus Android install performance by channel.

• What day of the week has the highest engagement for clicks?

• What proportion of deep link users performed an add to cart event?

Report metrics

Branch Analysis gives you several ways to share your data and discoveries with your team.

Download data

To download an entire dashboard:

1. Navigate to the dashboard you’re interested in downloading.

2. Select the three dot icon in the upper-right corner of the dashboard, then select Download PDF.

To download data from a single widget:

1. Navigate to the widget you’re interested in downloading.

2. Hover over the widget and select the three dot icon that appears, then select Download.

3. Choose your download type from the available options. The types of downloads available depend on the type of data in the widget.

4. Select the Download button.

Share dashboards

You can share one or more dashboards at a time with your teammates:

1. Navigate to the All Dashboards page under Analysis in the left-side navigation of Branch.

2. Select the checkbox next to the dashboard(s) you want to share. Select the Share button.

3. Add either entire groups or individual users to the list of entities to share the dashboard(s) with. Decide whether each entity can view or edit the dashboard(s) you’re sharing.

4. Select the Share button. These users will receive an email and see your dashboard appear in their list of All Dashboards.

FAQ