Insight Report
The Insight Report can be used to learn about a Seed (or target) Dataset from a Baseline (or comparison) Dataset, allowing you to find attributes that over or under index for your target group. This means you can understand your customer base more so you can deliver more relevant campaigns and creatives, resulting in more positive outcomes.
Companies can use this report to learn more about their audience for:
- 360 customer view: How do my clients differ from the general public? Strategic decision making around positioning, developing new products, sponsorships, or adjusting creative and messaging.
- Insights for media planning and/or activation: By identifying relationships between key variables such as CRM data, publisher content and/or demographic attributes, marketers can segment, prospect, and plan campaigns more effectively.
- Post-campaign audience report: Add demographic and behavioural data to your ad exposures and/or transactional data to gain greater performance insights and learn about potential new audiences to improve upcoming campaigns.
The Insights Report allows you to select the Datasets that you wish to use in the analysis and which filters (if any) need to be applied. You will need to select two datasets:
- A seed Dataset: This represents the group that you want to 'learn' about. Usually this is the brand first-party data or another dataset that represents the brand’s target audience (e.g. a retailer dataset)
- A baseline or comparison Dataset: This contains the wider universe of people, inclusive of your target. This is the data you want to 'learn' from. Usually, this is the publisher’s dataset or a data provider for insights.
Table of contents
Analysing the data in a report
Requirements
Max baseline attribute value count is 100
The user must select which attributes from the baseline Dataset they wish to include in the analysis. The maximum number of attribute values that can be used in the report is 100 - we are in development to increase this volume.
Each baseline attribute must have at least two values
There is no limit on the number of attributes that can be included but each attribute must have a minimum of 2 values. The number of attributes and attribute values included in the report will increase the time to takes to generate.
Data type support
The insight report currently only supports string & integer attributes, floats and date time are not supported.
Only datasets supported
We don’t currently support audiences or segments to be used as the seed or baseline
The seed and baseline cannot be the same dataset
| Important: Permamence of reports |
| The platform will run and save the aggregated and anonymized results of an Insight Report so they can be viewed at any time. This means that even if the original Datasets expire or change, or the permissions change, the report created will still be available with the same aggregated data as at the time of creation. |
Running an Insights Report
On the Infosum Platform, go to the Reports tab under Reporting
In the Reports tab you will see a list of any reports that you have already created for each of your collaborations.
You can navigate between collaborations using the dropdown menu at the top.
To create a New Report, click on New Report in the top right side of the screen.
- Select the Seed and Baseline
- Select the Baseline attribute to include in the report
- (Optional) Filter seed and baseline
- Name and run your report
Select seed and baseline
On the following screen you will still be able switch between collaborations using the dropdown. When you are in the correct context, select your Seed Dataset and Baseline Dataset for a comparison to preview your intersection.
You will only be able to select datasets that are available in the specific collaboration context you are in.
Once you select your datasets, the platform will automatically calculate the intersection count.
Click Continue.
Select Baseline Attributes
In the Attribute Selection screen you will be able to select the attrbutes you want to learn about from the 'baseline' Dataset.
You will see the counter of attribute values selected update as you include further attributes.
Click Continue.
(Optional) Filter seed and baseline
This optional step allows you to filter down your report by specific attributes properties, conditions or ranges.
- Filtering down the seed: this allows you to create a very specific target profile, e.g. loyal customers, recent purchases, high value customers, churned etc
- Filtering down the baseline: this allows you to focus your analysis on attribute values that you really care about, e.g. locations, specific dietary preferences etc
To add a filter, click on the blue + icon next to the filter group & select the specific attribute you wish to filter on. Once you’ve added a filter you can see how that filter will impact the available data for the analysis by clicking on calculate filter size.
Important note:
- You cannot filter on floats or date-time at present but we’re working on bringing this functionality to the platform soon. If your dataset contains any of these data types, they will show in the filter list but there will not be any values available to filter on.
- You will not be able to filter on a value where the number of records in the dataset that contain that value is lower than the reduction threshold of the dataset
You can swap between AND/OR groups by clicking on the ‘swap’ button on the bottom right of each dataset’s filters
Integers can allow for range filter or condition filter (equal/higher/lower)
Strings can be filtered by inclusion/exclusion
Name the report and click on 'Create New Report'
On that same screen, simply name your report
Your report will generate in a max of 30 mins, depending on the number of attribute values and size of the Datasets.
Analysing the data in a report
Once the report is created you can click on 'view report'
The report is made of three tabs:
Charts
You’ll be able to see an initial comparison between seed and baseline for each attribute and values that you’ve selected.
On this bar chart you’ve got the baseline in orange and the seed in blue, and you can compare the % of records that have that attribute value and also the index for each attribute value.
| What is an index? |
| The index indicates the affinity that members of your audience have for a particular attribute value compared to the baseline. An index above 100 means that members of your audience show high affinity for that attribute value while an index below 100 means they show less affinity than the baseline. |
This is great to see at a glance what the most relevant attributes for your target are. You can scroll down the page to see each attribute or click on the left hand side list to navigate to it.
Indexing table
The second tab shows a table view, where you can also see the the volume of the intersection for each specific attribute value and a handy heatmap for the index which helps you visualize what characteristics are most relevant. This table can be ordered by highest to lowest index or viceversa,making it a compelling way of identifying attributes that positively and negatively describe your target audience.
You can use the most relevant attributes for your media planning, to create high-value expansion segments, or to learn more about your audience and apply it to strategic projects such as tailoring the message or creative.
You can use the least relevant atttributes for suppression of out of target audiences, identifying areas of growth or to influence messaging decisions.
Cross attribute analysis
The final tab allows you to compare two attributes at a time in a cross-tab table which can give you more granular insight into different attribute combinations.