Getting Started FAQs
A. Yes. Permutive’s identity framework allows you to identify users across different domains.
Permutive uses a third-party cookie, the AppNexus ID, in our identity framework. This means that you can identify a user as the same Permutive user when they visit different website domains owned and operated by you.
For more information around how AppNexus collects, uses and shares AppNexus IDs, click here.
A. If the property type is a 'List of Strings' then it is case-sensitive
A property value that pulls into Permutive as a 'List of Strings' must be entered into the segment builder exactly as it is being ingested, including case-sensitivity.
A. There are two ways to check if a property is a 'List of Strings'.
1. Log in to your Permutive project. Under the Events' tab choose the event you want to exlpore and click 'Quick View' to see the list of properties attached to that event:
Once in this view, check the final column and you can see what 'type' the property is, including whether it is a 'List of Strings'
2. When you add a property into the segment builder, the operation that succeeds it can help you determine if the property is a 'List of Strings' or not.
If the operation is "list contains", rather than just "contains", the property is a 'List of Strings' and the value must be entered exactly as it is ingested into Permutive.
A.You can use the Permutive Browser Extension to check the values that are being pulled into Permutive for each property.
A. The user may have closed the page before segmentation occurred.This article explains the difference between the predicted audience size and the live audience size of a segment. The predicted audience size counts raw Pageview events whereas the live audience size counts Segment Entry Events. The Pageview event is the first event to fire on the page. If a user closes the page after the Pageview event fires but before the Segment Entry event fires, this could result in the predicted audience size being higher than the live audience size. This could also happen in AMP as, while events are tracked for all users, segmentation only happens on the second pageview.
Note: The predicted audience size will be larger than the live audience size if the segment is less than 30 days old.
A. Yes. Write a list of the values separated by a comma and copy and paste it into the segment builderInstead of manually entering each value, you can write a list of all the values separated by a comma. Copy and paste this list into the segment builder and press 'Enter'.
A. Permutive is currently integrated with over 200 data partnersClick here to view the full list of partners we can enable for your project. We can also ingest data from providers not on the list via LiveRamp. Please reach out to support if you have a preferred data partner not on this list.
A. ITP stands for Intelligent Tracking Prevention and refers to Apple’s anti-tracking feature in Safari.ITP was added to Safari in 2017 to limit sites’ ability to monitor users’ activity on the Safari browser. This effectively blocks third-party cookies in Safari and on Apple devices. With ITP versions 2.1 and 2.2, Safari also limited the ability of companies to track users with first party cookies via link decoration by deleting all cookies after 7 days with 2.1 and then after 24 hours with 2.2, further limiting companies' tracking ability. With the most recent version of ITP (2.3) Safari now prevents cross-site tracking as well. This does not limit Permutive’s ability to track Safari users as the platform does not rely on third-party cookies, link decoration, or cross-site tracking.
A. No, not by default.Permutive is not generally considered an 'ad-tech' tool because our platform doesn't serve any ads. However if a user configures their ad-blocker to also block tracking scripts (e.g. Mixpanel, Google Analytics) then Permutive will also be blocked.
A. We support with our real time dashboards which show articles trending.
Permutive's real-time Data Warehouse can be coupled with BI tools to create real-time editorial dashboards. These can be used to understand what content is trending and to optimize content production.To enable the Data Warehouse product, please get in touch with your Customer Success Manager.
A. You can change your password in the settings tab in your Permutive Project1. Log in to your project and click on settings 2. Click Team 3. Find your user name and click on 'Change Password'
Note: Passwords must be a minimum of 8 characters in length and contain at least one letter and one number. If you need to reset your password please reach out to email@example.com
A. Most up to date browsers are supported including Google Chrome, Apple Safari, and Mozilla Firefox.If you have issues with a certain browser please contact firstname.lastname@example.org
A. Permutive uses identity management to sync user_ids and can be used to deterministically associate browsers to a user.
If a user has an account with you and is logged in on multiple devices, we are able to identify that user as the same user on multiple devices. If they are not logged in, the user will be recognized as multiple unique users.
A. Asynchronous load and compression mean that Permutive doesn't impact page load time.Permutive is designed and optimized to not have any impact on page load time. Our script typically takes less than 100ms to download. This will differ marginally based upon the number of segments a user falls into, though this will be tens of milliseconds difference.
Permutive goes through two stages of compression: minification and then gzipping. All methods execute in web workers where available, meaning it doesn't block JS on the page. All methods register callbacks, rather than executing blocking code.
A. Our segments, reactions and analytics update in milliseconds.
Permutive processes segments within milliseconds of receiving new data. This means that our segments are constantly updating before the page loads, and ongoing as the user browses on your site. Every time a new event fires to Permutive we update the user's segments within milliseconds.
Whenever a user enters or exits a segment we fire a 'Segment Entry' or 'Segment Exit' event and trigger any corresponding 'On Entry' or 'On Exit' activations.
Activations that are set up as 'Every Time', trigger every time a user fires a new pageview event. So the activations are constantly up to date!
Similarly, our data warehouse is real-time, so our analytics reports update with sub-second frequency.
A. Permutive uses the standard HTML property document.referrer.
The client.referrer property returns the URI of the page the user was on to get to your site. You can use this information to target users based on their referrer.
A. Permutive will collect Slot events for all ad units that are accessible by the Google Publisher Tag googletag object on the publisher's page.
Depending on where the creative is hosted, we may or may not be able to collect metadata on it. This means that we will record SlotClicked/SlotViewable events for all ads set by GPT, but some of these events will be missing metadata.
A. In the case of Slot events, metadata would be the data about an ad that was shown. This includes campaign ID, order ID and line item ID.
A. Permutive is able to collect creative metadata for creatives that are hosted within GAM. This applies to both display and video creatives. Usually, this set up corresponds to direct deals.
A. If a creative is hosted outside of the DFP, Permutive will not be able to collect metadata, as the metadata comes with the creative, and the creative is coming from a different AdServer. In most cases this will mean that no metadata is available, although in some cases metadata may come through, it will refer to IDs of the other AdServer, rather than the DFP.
A. A session lasts 30 minutes. If we don't see any events for a user for 30 minutes, their session expires.
Example: If a user records 180 seconds of PageviewEngagement events and then disappears for over 30 mins, their session is 3 minutes.
A. Permutive can infer device type from the client's user agent. Every event is enriched with the client's user agent.
A. Permutive has an infinite look-back period.
We keep your users' entire data history for you to use when creating audiences through segmentation.
A. Yes, Permutive's integration with GAM allows for ad viewability tracking in real-time via the SlotViewable event in the dashboard. This method of tracking is MRC certified.
A. Currently, we are unable to see users to see who edited segments in the Permutive platform.
In some cases, we are able to see recent activity logs. For urgent requests, or if you have any further questions, please get in touch with the customer operations team and we'd be happy to help.
A. We store your users' entire history so you can use it when building segments. There is no maximum look-back period.
A. Yes, all 'passer by' traffic is targetable with Permutive.
Permutive identifies and segments users before the page loads so they can be offered targeted advertising from their first page view. This means all traffic - including passers by - is targetable.
A. Yes, segments remain in Permutive until you choose to remove them.
We store your users' entire history in Permutive so you can use it when building segments. So if, for example, you had a summer segment of 'Wimbledon Watchers' in July 2018 you could target that same group in July 2019.
A. When you create a Permutive segment it will instantly go live, however, it can take up to 5 minutes to propagate globally across our CDN.
A. In short, Predicted is an estimate of unique users and Live is the actual number of unique users who have entered the segment. While most of the time the numbers will agree with each other, in some specific cases differences in numbers between these can arise, as the logic between them differs
A. This is usually the result of not having enough data to calculate correctly.
If you have a segment that is created with an 'in session' clause, Permutive will be unable to calculate the audience size. However, once you have set up the segment, you can see its growth in the analytics reports. After 30 days you will have enough data to report for an RFP response. To receive an up-to-date estimate, the estimator will need to be re-calculated by manually pressing the 'Calculate' button at the bottom of your segment builder.
A. Two segments cannot have the same name. Try adding 'v2' or differentiating the segment name in another way.
Integrations and Activations FAQs
A. Permutive integrates with all DSPs
A. Campaign Insights can be used on Programmatic Direct and Preferred Deals only. Insights are not available for OMP deals.
Permutive will collect Slot events for all ad units that are accessible by the Google Publisher Tag googletag object on the page.
Permutive provides insights on direct and preferred advertising campaigns with two caveats:
1.The campaign cannot contain video creative as this is not tracked using Permutive slot events.
2.The campaign cannot contain third-party hosted creatives. If a creative is hosted outside of Google Ad Manager, Permutive will not be able to collect metadata, as the metadata comes with the creative, and the creative is coming from a different AdServer.
A. All segments in your project can be used as a base segment for index reporting
You can choose 'Everyone' or select a specific segment to be your base segment.
Example: If you have global audiences you might want to run an index report by GEO location, comparing your UK and US visitors. To do this, you would go into insights, click the index and overlap tab, run an audience report on your UK visitors segment and filter your base segment to select the US visitors segment.
A. Yes, you can switch between 3 different providers: Eyeota, Bombora and Experian (US).
You can switch between these providers in the Third-party Index tab, both in the Audience and Campaign sections, under Insights. To do so, use the 'Provider' filter:
A. An index of 140 could be read as either 1.4x more likely or 40% more likely.
A. For every report that uses an index, 100 is the base and demonstrates an even likelihood.
If the index is above 100, this indicates more of a likelihood. If the index is below 100, this indicates less likelihood.
There are two ways to interpret a positive index number, either as a decimal or as a percentage:
1. An index of 125 would either be 1.25x more likely or 25% more likely.
2. An index of 245 would either be 2.45x more likely or 145% more likely.
A. The demographic data comes from a third-party provider that is enabled in your project. By default, the data will come from Eyeota.
You also have the ability to change providers, with the choice of Eyeota, Bombora, and Experian as shown below:
A. In most cases, a campaign is not appearing in Campaign Insights because one of the line items of that campaign contains video creative.
The reports are pulled from the SlotViewable and SlotClicked events, which do not apply to video creative and cannot recognize campaigns containing video.
You can look in GAM to see if a campaign contains video by following these steps:
1. Login to your Google Ad Manager and search for the campaign ID
2. Select each line item from the order
3. Confirm that there are none of the creatives contain video
4. If there are no video creatives, please contact email@example.com and we will investigate further.
A. Yes, we can pass your segments into 3rd party analytics platforms such as Google Analytics and Omniture for reporting within those systems.
All our data is also stored and updated in real-time within the Permutive Data Warehouse for querying and joining with other data sets. Please get in touch with your Customer Success Manager if you'd like to know more.
Permutive doesn't currently pull data out of these systems, however, our open API supports this type of ingestion.
A. We store your user's entire history of data and use that for segmentation and reporting. So you can run any report over any time frame.
Note: Please contact support to run reports over less time than the standard date range allows (eg. the last 5 minutes).
A. Coordinated Universal Time (UTC).
A. If you know that your target seed segment is only related to a few segments, and the rest of them will create noise, then you can use 'Fixed Set' to select only the relevant segments required for training.
A. Run a comparative test against the seed segment/ a third-party segment and measure performance following the steps here.
A. Use the weights produced by the model. For example, if a model with a Sports Lover seed segment has a Rugby segment with a big negative weight then you could speculate that something is working incorrectly. In general, we expect someone who likes Rugby to also be a Sports Lover and vice versa
A. You should avoid having too many similar segments in your project, eg Sports Lovers and Sports Enthusiasts. This is because similar segments aren't useful as they essentially capture the same data. If your seed segment has a strong relationship with another highly similar segment, the model may learn this relationship and ignore the rest.
A. No, you can only choose one seed segment to train a model with.
You are able to select which of your first-party segments you want the LAL model to be trained on.
Example: If you are building a lookalike model on top of your declared female users to identify other users who are likely to be 'Female' you may want to train this model excluding your male declared users.
For more information on setting up a Lookalike Model, please click here.
A. Yes. LAL segments can be used like normal segments.
A. No. However, the seed segment properties can be edited to define the users that will be in the LAL model segment. The model will then take 24 hours to retrain.
A. Yes, LAL segments will behave as any other segment.
A. All lookalike segments created from a seed segment will not be automatically deleted and will still work.
However, since models are trained daily, using that day's data, when a seed segment has been deleted, users will no longer be entering the seed segment and the model will become less accurate.
A. Yes, lookalike segments behave like normal segments for all intents and purposes. Think of them as segments with automatically generated conditions.
A. We suggest a segment have at least over 1k active users in the past day to create an accurate lookalike model.
A. There is no minimum, but we recommend at least 10.The more segments you have (and the more users your segments have in them), the more accurate the lookalike model will be.
A. Model creation takes up to 24 hours, but in many cases will finish within less than an hour, depending on the amount of available data.
We schedule model trainings at a specific time of day, hence in a worst case scenario, the model will be available 24 hours after you select 'Build lookalike model...
A. No. Lookalike models can only be built from segments using first or second-party data.
A. A red dot represents that a lookalike segment has been created with the given similarity and reach.
A. The goal of lookalike (LAL) segments is to expand a segment (the "seed segment") by finding similar users.
In order to create a lookalike segment, Permutive uses machine learning to generate a lookalike model. The model learns the common features (in this case, what segments a user is in) and their relative importance.
Example: If you want to expand a segment of "Tennis Lovers", our lookalike model may learn that many users in the segment are also in a "Sports Lover" segment. This means that if we see a user that is in "Sports Lover", within some given probability, they are "almost" a "Tennis Lover" as well. Conversely, the model may learn that users who are "Tennis Lovers" are not "Basketball Lovers", and a negative importance can be given to the "Basketball Lovers" segment.
A. No, there's no extra cost for this. Permutive does not charge to activate lookalike segments for targeting.
A. An API key is a unique code that is used by computer applications to authenticate a user, developer or project with the API.
Application programming keys are normally used to assist in tracking and controlling how the interface is being utilized. Often, it does this to prevent abuse or malicious use of the API.
For more information on when and why to use an API key, please click here.
A. To add a new Workspace to your Permutive platform, please contact firstname.lastname@example.org
A. Login to your Permutive Project and head to Settings > Team.
A. If you need to remove a user's access to the Permutive dashboard, please contact email@example.com
Privacy and Compliance FAQs
A. The Interactive Advertising Bureau (IAB) is developing a Framework to help publishers and technology providers comply with CCPA.
Initial technical specifications for the framework are available here.
When a user clicks that “Do Not Sell My Personal Information” link, a signal is sent to technology companies they do business with via a technical mechanism that is being developed by the IAB Tech Lab.
A. Once a publisher receives a request from a user to opt out of the sale of their data on its site, they should contact Permutive. We will stop capturing information on that user and prevent their data being sent to an advertiser.
Further information on how to remove Permutive data collection for a user, as defined by GDPR, is documented here.
A. Catalog your list of users by the user ID in a list or spreadsheet and email it to firstname.lastname@example.org. We will action this within 45 days of receipt and confirm when the request is complete.
A. For CCPA, Permutive will be classified as a ‘service provider’ to our publishers. We collect user information on behalf of these businesses, and only on their instruction.
Permutive takes consumer privacy seriously and does not sell publishers’ user data to third parties. However, we allow publishers to do so using the Permutive platform.
Permutive will ensure compliance with CCPA by providing an option for publishers to access or delete data from our systems, when requested by a user. We have experience offering this capability for GDPR in Europe.
If a consumer opts out of the collection or ‘sale’ of their personal information, we will stop Permutive segmentation and targeting for those users, when requested by a publisher.
We encourage publishers to leverage first-party data. Permutive’s Edge DMP is designed to support you in segmenting and targeting users without using third-party data.
A. The California Consumer Privacy Act (CCPA) is a state law restricting the use of personal data for California State residents.
It has been enforced since January 2020. Under CCPA, consumers have been given the right to: access their data, erase their data, opt-out of the sale of their personal information.
A. A range of sensitive data points must not be collected by Permutive.
These include: credit card numbers, passport numbers, social security numbers, genetic data, physical health information, mental health data, sex life or sexual orientation data, religious or philosophical beliefs, political beliefs, trade union membership.
A. Catalogue your list of SAR users by the user ID in a list or spreadsheet and email it to email@example.com. We will action this within 21 days of receipt and confirm when the request is complete.
Process is documented here.
A. This process is documented here.
A. No, third parties do not have access to use the data we process for our publishers unless the publisher decides, as a controller, to enable this.
A. Permutive is a data processor.
Permutive processes personal data on behalf of a controller and has been audited with a ‘Technical evaluation of data privacy for the product “Permutive, version 2.0”’ by ePrivacyseal, who have classified Permutive as a processor.
A. As a user interacts with the site, a permutive.track call is encountered and a user-event is recorded.
This event is sent to the Permutive API for processing and storage, and internally, to the segmentation engine in the SDK for decisions and actioning.
At the Permutive API level, the event sent by the user is validated against the event schema setup by the publisher, and is rejected if it does not conform. The event is also enriched with Geo information based on the IP address of the user.
The Permutive API initiates a data flow that aims to store the event for posterity and analytics and trigger any server side integration for SegmentEntry and SegmentExit events.
At the segmentation engine in the SDK, the user-event causes an update to every segment. For every segment defined, when an event is fired, the user’s segment state is updated to either true or false, denoting the user having membership or not of the segment, respectively.
A. Permutive collects user data on behalf of the publisher. As the data controller, the publisher configures Permutive to decide which data points they want to send to Permutive for processing. Standard events collected can be seen here.
A. Permutive uses an AppNexus ID to match users to third-party audiences.
While Permutive doesn't rely on third-party cookies, you are still able to build segments using third-party audiences. Furthermore, the Insights module uses third-party data to power the 'Third-Party Index' tab, which allows you to see third-party insights into your audiences.
If a user has an AppNexus ID in their browser when they visit a site where Permutive is deployed, Permutive uses this ID in order to match users to these third-party data audiences.
For more information around how AppNexus collects, uses and shares AppNexus IDs, click here.