Creating common segments
To create a contact, email behavior, geolocation or Smart Insight segment, proceed as follows:
1. Select Contacts > Segments, then click Create Segment.
2. Choose Contact, Email behavior, Geo, Smart Insight segment.
Smart Insight segments can be created both here and on the Smart Insight interface.
3. Give your segment a Name and add a Description to it.
This can help you identify segments easily on the Overview page.
4. Choose your Source list.
All segments are run against either all contacts in your account, or on a specific contact list. You cannot run a segment directly on another segment, but you can combine multiple segments together.
This is where you define which contacts should be in the segment. You can base this on various criteria, see below. Within each category, you can create any number of conditions, but all on the same level and all linked by the AND operator.
Conditions are added and removed via the + and - icons.
If you want to define a more complex filter, for example with nested conditions or using both the AND and OR operators, click Advanced Filter Options.
To add a nested sub-criterion, click the plus sign on the right side. Select the AND/OR operator from the drop-down menu.
In the example figure above, the new segment is created with the following criteria:
- all contacts are from Germany, AND
- they have an email address, AND
- they are female AND have more then 3 children, AND
one of the following criteria is true:
- they are older than 30, OR
- their education is to university-level or unspecified, OR
- they work in retail/wholesale/e-commerce, OR
- their marital status is single or unspecified.
There can be many levels of nested sub-criteria, but we strongly encourage not using more than 4 levels since it makes the criteria on the screen very hard to decipher.
These are based on your contact database and include both system and custom fields. For each field, the appropriate conditions are offered. For this reason it is important that you use the correct field type when creating custom fields.
Please consider the following when using Predict contact criteria:
- Entries within the predict last ... categories and predict last ... products fields are separated by pipes (
- You cannot specify multiple values within a single Predict contact criteria filter. To add multiple values, you need to add each Predict contact criteria filter separately.
- The following examples show you how to use these fields:
predict last purchase categories contains
Men > shirts > slim fit
predict last purchase products contains
- predict last purchase categories contains
- For more information on these fields, see Web behavior fields.
Please note, that using string based contact filters, such as, "Does not contain" or "Begins with", with any value will not include contacts with an empty field for the specified criteria.
To include those contacts, you should use the filter combined with the "Is empty", for example "Is empty Or Does not contain" condition.
For example, when you want to create a segment for all contacts living outside a city (including those without location data) with a purchase history greater than 1000 USD, use the following segments:
Using the filter City - Does not contain - Budapest will not return contacts without location data.
Some date fields are dynamic, meaning they are calculated by Emarsys when the segment is generated. These are mostly date-related fields such as Age.
You can use only segments that contain exclusively contact criteria in Real Time Interaction programs.
To check whether a segment can be used in Real Time Interaction programs, open the segment by clicking the Edit button on the Contacts menu > Segments page and check the message displayed on the bottom of the page. You can select a segment if you see this message: "This segment is ready for use in Real Time Interaction programs."
When you have finished adding the contact criteria, select the AND/OR operator and add the Email and mobile behavior criteria.
Email and mobile behavior criteria
These are based on the contacts’ responses to your email campaigns (or lack of them). For example, you can filter for contacts that have received more than three email campaigns in the last month but have not made any clicks, or for those contacts who have clicked the most links in a particular category.
Behavior segmentation for on-event campaigns is only available if the campaigns are included in an Automation Center program and they have an Email category assigned to them.
You cannot create segments based on your contacts' responses to Triggered emails. However, the Email bounced (bounce notification) email behavior criteria filter contains information related to bounced batch, on-event and triggered emails.
Transactional emails in email behavior segments
By the end of July 2020, transactional (triggered email or on-event) campaigns will be listed in the drop-down when you select specific email campaigns to which contacts did or did not respond. As a result of this improvement:
- You don’t need to include on-event campaigns in an Automation Center program and don’t have to assign an Email category to these anymore when you would like to use them in email behavior segments.
- You can create segments based on your contacts’ responses to Triggered emails.
- Email campaigns used in Interactions programs will be available for email behavior segmentation. You can also use email behavior segments that were created based on Interactions campaigns and optionally, you can exclude such segments from Automation Center programs if those contacts already have received a message from your Interactions program.
These improvements will change the results of your email behavior segments if the segmentation criteria are based on email categories. Previously, transactional campaigns were not available in the following filters:
◦ Email response (open or click)
◦ Email response (click only)
◦ No email response (open or click)
◦ No email response (not clicked)
◦ Received an email
◦ Not received an email
So, if you assigned a category to a transactional campaign but it was not used in an Automation Center program, then this campaign was not taken into account when calculating the results. Due to this improvement, such results may change, therefore, you need to adjust your email categories accordingly if you would like to avoid the results being affected by data related to transactional emails.
Nevertheless, transactional emails were available in the following filters (and this behavior will not change):
◦ Email bounced (bounce notification)
◦ Individual link clicked
◦ Individual link not clicked
◦ Links in category clicked
◦ Links in category not clicked
Changes in email behavior segment results
If you create an email behavior segment, then the related email campaign's launch date will be adjusted to your account's default time zone.
For example, if your account's default time zone is in Australian Eastern Standard Time (UTC + 10) and you run an email behavior segment at 15:00 (in your time zone) that contains those contacts who received the email sent on the current day (between 00:00 and 15:00, so you set the timeframe to Between 0 and 0 days ago), then your segment will return those contacts who received your email between 00:00 and 15:00 on the current day.
Email and mobile behavior criteria filters
The following behavior criteria are available in the pop-up window:
- Email response (open or click)
- Email response (click only)
- Email bounced (bounce notification)
- No email response (open or click)
- No email response (not clicked)
- Individual link clicked
- Individual link not clicked
- Links in category clicked
- Links in category not clicked
- Received an email
- Not received an email
Depending on the criteria you choose, you can further refine your filter with a number of additional options.
- Timeframe – Restrict the response to a specific time period.
- Volume restriction – Restrict contacts by the number of emails they responded to.
- Device (Mobile/Desktop) – Select between desktop or mobile (or both), and if mobile, select from a range of platforms.
The contact behavior here depends on the notifications returned by the target email client. This may not always exactly reflect the behavior of the recipient. Therefore the following points should be considered when filtering for contact behavior:
- An email which is opened without images does not return an open indication.
- Emails which are read in the mail client preview window also do not return an open indication, unless the client is set to display images.
- A contact may click a link several times to open it. Depending on the click speed and browser settings, each click could be counted separately even though the link was opened only once.
Behavior filters like "Responded to a special campaign X" will not work for recurring campaigns, because every recurrence is treated as a separate campaign.
- If you want your segment to include behavior targeting criteria based on contact response to a recurring email, you must give that email a unique category and base the response on the category rather than the email.
- If you change the category of an email campaign that is used in a recurring Automation Center program, those emails that were sent before the change will belong to the previous category.
In most (though not all) cases, Emarsys can identify the city where a response was recorded. In the Geolocation Criteria feature you can select a location, city or country anywhere in the world and create a segment of the contacts who responded within that area during the selected time period.
- Timeframe – Here you select the timeframe for the responses.
- Target group – Emarsys initially treats every contact as a visitor to a given location until their behavior has become regular enough to identify one location as their home. This categorization is dependent entirely on their responses and has nothing to do with any database fields (e.g. City, State or Country). These categories are constantly evaluated and updated by our algorithms after every response.
You can select a country by entering its name in the Location field. Once you are done, the country will be highlighted on the map.
You can select a city either by planting a marker on the map with your mouse, or by entering an address in the Location field.
The map automatically zooms in or out, if necessary, to show the selected location.
This is an integration with Google Maps so any address can be entered. However, only the country or city will be used by Emarsys geolocation. Segments within a city are not possible.
While Emarsys can precisely select countries by their borders, for cities and other locations, you have to define a radius. The radius is measured from the city center or the location marker and creates the area in which responses are identified. You can choose your unit of distance (miles or kilometers) during account setup and modify it on your My Profile page. The minimum radius is 25km; there is no maximum radius.
Click Calculate Contacts to show how many contacts in the selected target group have been identified for your location.
You can only use one set of geolocation criteria per segment, but within that you can define as many locations as you like.
All locations within a segment are linked with the OR operator. If you want to use the AND operator, you should create two segments and combine them.
Home location calculation
Geolocation segmentation calculates the home location of contacts based on their interactions and takes more recent locations into consideration with more weight.
The calculation is made up of the following steps:
1. The platform filters for the daily unique locations of a contact.
Example: The contact has one open and two click events from London and one open event from Luton for a given day. We register one visit to London, and one visit to Luton for that day.
2. The platform aggregates the number of visits on a weekly basis.
Example: The contact has four visits to London and one visit to Luton in a specific week. In this case, London will be set as the contacts home location.
3. Events are weighed based on how long ago they happened. Each event that happened in the past month counts as 1, events that happened between 1-2 months ago count as 0.66 and events that happened 2-3 months ago count as 0,66 * 0,66 etc.
Example: The contact has two visits in London this week and three visits in Luton two months ago. After weighing the number of visits, the contact has two visits from London and 3*0.66 = two visits from Luton. As this would result in a tie, the city where the most recent activity was captured will be set as the home location. In this case, the contacts home location will be set to London.
4. Visits to cities within 25km of each other are aggregated for each location increasing the importance of that area in the calculation.
Example: Two visits to London and one visit to Wembley are registered as three visits to London and three visits to Wembley, as these cities are close to each other.
5. The system selects the country with the largest aggregate number of visits as the home country. If two countries have the same result, the country with the more recent visit is chosen.
6. The system selects the city with the largest aggregate number of visits in the home country as the home city. If two cities have the same result, the city with the more recent visit is chosen.
Smart Insight criteria
If you are using the Emarsys customer intelligence module Smart Insight, you can also add filter criteria based on eRFM activity and web behavior.
When you save a segment in Smart Insight, a segment is created in Emarsys using the filters currently applied on that report.
To save a segment in Smart Insight, proceed as follows:
These segments are initially saved with a single Smart Insight criterion taken from the report. You can further modify this criterion by clicking Edit:
All additional attributes that you add here will be linked by the AND operator (i.e. all must be true for all contacts in the segment).
You can create new criteria by clicking +. In this case, you have the choice between the AND and OR operators.
You can also define the operators between the Smart Insight, Contact and Behavior criteria by selecting Advanced Filter Options.
When combining multiple criteria types, the operator for the Contact and Behavior criteria are first applied, and then the operator for the Smart Insight criteria. In other words, your filter syntax looks like this: ([Contact criteria AND/OR Behavior criteria] AND/OR [Smart Insight criteria]).
Creating a segment including contacts who purchased a specific product on the day before
If you would like to create a segment that includes customers who bought a specific product the previous day, then use the filters Item and Order(s) made X days ago together. In this case, we recommend that you proceed as follows:
- Click the + icon.
- Select the Item and the Order(s) made X days ago filters and set them as required.
- Click OK only when you are ready.
As a result, the two filters will be grouped together automatically. So, this segment will contain only those contacts who purchased the product defined by the Item filter yesterday.
This is how your Smart Insight criteria will look like if the filters are grouped automatically:
If you have selected one filter first, accidentally clicked OK, and added the other filter by clicking the + icon again, then filters will be linked by the or or and operator, and you will not achieve the desired result.
If, for example, the filters are linked by the and operator, then your segment will contain customers who purchased the specific product previously and placed an order yesterday (which can refer to any other product). In this case, the timeframe defined by the Order(s) made X days ago filter applies to any item customers ordered the previous day and it is not limited only to the product specified in the Item filter.
You can modify the filter by clicking the Edit button.
This is how your Smart Insight criteria will look like if the filters are linked by the and operator:
Combining segments with other segments
If you would like to combine segments with other segments, follow the instructions described in Combining segments.