Product
Live moments capture what a user is doing at a given point in time in the real world, while using an app (e.g. commuting, relaxing, running).
We combine mobile sensor data (e.g. accelerometer, gyroscope, location) using our patented on-device AI software, to create live moments without requiring any IDs or PII.
We predict over 200 live moments.
Behavioral cohorts capture a user’s behavioral attributes (e.g. joggers, regular shoppers) and can be targeted by advertisers without the need for IDs.
As a user uses an app over a period of time, we learn about their habits based on what they do in the real world (e.g. running 3x a week, driving every day) using any combination of our 200+ live moments. These repeat behaviours give us an indication of which cohorts the user might be categorized into (e.g. joggers, vehicle users).
Cohorts traditionally use data from a variety of sources (e.g. mobile apps, IP address, web browsing) and are stored in cloud databases. A shareable identifier is then used to combine these data points, to determine which cohorts a user might belong to. For example: EA7583CD-A667-48BC-B806-42ECB2B48606 → Auto Intender, Car Enthusiast.
With behavioral cohorts there is no centralised data repository, because they are generated and stored within the app itself. There’s no need for a unique identifier to determine a user’s audience memberships, because all behavioral data already resides within the app. For example:
Music App: Joggers, Walkers, Early Risers.
These behavioral cohorts can be activated by advertisers to reach relevant cohorts without requiring any app-level identifiers for targeting, thereby preserving the user’s identity.
Yes. We work closely together with our demand partners and customers, and leverage their experience to bring new ID-less behavioral cohorts and live moments to market.
If the app runs in the background (or has a background service), then we will run as well. Most audio apps always run in the background because users typically listen to music while continuing to use other apps. If the app only runs in the foreground (i.e. when the app is open), then we will only run in the foreground.
A user can a member of multiple cohorts. For example, if they frequently use the app when going for a walk and going for a run, they’ll be members of the Walkers and Joggers cohorts.
Advertisers can show a personalized ad based on 3 different scenarios:
- Based on the cohorts the user belongs to e.g. Walkers, Joggers
- Based on live moments the user is in at the moment e.g. Walking, Post-Exercise
- Based on both the cohorts and the live moments of the user e.g. Joggers in a Post-Exercise context
Our ID-less behavioral solution Euclid works across all channels, but requires a mobile app seed to get activated. The ID-less demographic solution Euclid Lite works on in-app and podcast, podcast is for all channels
iOS, Android, Unity
Around 90%-95% in most countries across the world.
You can request a copy of our taxonomy here.
Yes. You can target users on a cohort level.
NumberEight’s ID-less demographic solution is provided for age and gender ad targeting and should not be used for regulatory or compliance purposes.
Activation
For our ID-less behavioral solution Euclid you can integrate our SDK. For our ID-less demographic solution Euclid Lite you can use an API. Please contact us and we’ll set you up with the documentation.
Yes, you can use our ID-less demographic solution Euclid Lite.
A number of our ID-less behavioral cohorts are mapped to IAB’s Audience Taxonomy 2.0.
The cohorts scale depends on the number of active users and the frequency of their engagement with the publisher throughout the day. Live moments signals will operate at 100% scale from day 1 of the SDK going live whereas behavioral cohorts will begin to start showing up after a few days of app usage.
If you’re an existing NumberEight partner, you can request the demo app here. If you’re new to NumberEight, book a demo with our team.
Please contact sales and we’ll set you up with the documentation.
Yes, you can invite as many as you want. As the admin of your account, you can add users through the ‘Organization Settings’ tab on the Portal.
Technology
On-device computing (or edge computing) refers to bringing the computation of data and its storage closer to where the data is being generated, or to the device itself. For example, in the case of smartphones, instead of sending raw sensor data to the cloud, all of the processing and computation happens entirely on the device, improving response time, reducing bandwidth usage and significantly minimizing chances of data leakage.
The impact is negligible. We work hard to ensure a day long battery life when an app uses our software continuously.
iOS: ~2MB download
Android: ~3MB download
Our-ID less demographic solution uses non-PII data, which is readily available and is otherwise meaningless, to provide accurate prediction regarding the gender and age on an impression level. For podcasting, alternatively we are able to provide a demographic distribution breakdown on a podcast or episode level.
Where possible, our data is updated on a monthly basis but certain data sets such as census data would be updated less frequently.
We provide options for an API integration. HTTPs, Bulk or self hosted. Depending on your preference and technical set up we can assess with you to see what’s most suitable.
No SDK is needed. Euclid Lite (ID-less demographic solution) is lite in touch and integration. A simple API integration is required and NumberEight provides a couple of options to suit your technical and operational requirements.
The data used by our ID-less demographic solution is sourced from diverse channels, including Open RTB bid request parameters, app store metadata, census data, 1st party publisher data, device information, content category, and ad request parameters. Through a comprehensive approach, multiple AI models are constructed using these data sources. The results generated by these models are then intricately combined to provide precise and reliable predictions. This amalgamation of various data inputs ensures a holistic analysis, enhancing the accuracy and effectiveness of our product in delivering insightful predictions.
Commercial
If you are a publisher with direct sales activities, our ID-less behavioral solution Euclid is the solution for you.
If you are an ad provider (SSP/DSP) looking to extend the addressability of your inventory, our ID-less demographic solution Euclid Lite is the solution for you.
If you are an ad provider (SSP/DSP) looking to extend the addressability of your inventory, our ID-less demographic solution Euclid Lite is the solution for you.
No. We are a data platform.
No. We are a data platform.
Our solutions are available globally, although currently, our customers are primarily in the US, UK, EU and Australia.
Yes. For our ID-less behavioral solution Euclid we offer a 2-3 month pilot period to our partners at a discounted rate. For our ID-less demographic solution Euclid Lite we offer a free test.
Any app owner or SDK owner can integrate our software for free (up to 10,000 DAUs) using the Community Software License and get access to our customer dashboard to get insights about their app audience.
Data
You can find all the details about the types of data we collect and process in our privacy policy.
Our demographic data is held to the highest possible standards and frequently validated against 1st party data sets provided by clients. Our models are updated as frequently as possible, and continuously improved upon. We usually see 95%+ accuracy from our probabilistic models when validated against deterministic datasets. The more (non-PII) data clients are able to share, the more accurate our data will be.
With our ID-less behavioral solution, while there is no overarching accuracy number as we predict over 200 live moments across various categories, all of our predictions are probabilistic with high degrees of accuracy.
Our ID-less behavioral solution Euclid is trained on a mix of proprietary datasets (from device manufacturers and academics), public datasets, simulated mobile data, and data collected in-field from staff and volunteers.
Our ID-less demographic solution Euclid Lite is trained with app store metadata, census data, proprietary panel data, content metadata, and other anonymized sources.
NumberEight’s methodology trains multiple AI models, each with a different approach, which provides precise and reliable predictions when combined.
Audience data is currently refreshed every 24 hours.
Our mobile SDK collects and processes mobile sensor data on the device itself, to deliver live moments predictions and behavioral cohorts. Our ID-less solution does not collect information: it processes non-personal data on the fly to predict demographics.
We store our data on Google’s servers located in London, UK.
No. All raw data on our ID-less behavioral solution is only processed on the device and is never sent out.
Privacy
We do our utmost to ensure user data is respected and never mishandled. A few things that we do to deliver a privacy by design solution:
- Doing as much processing on the device and not sending it to the cloud, significantly minimizing chances of any data leakage or misuse;
- Not sending any raw sensor data off the device to prevent fingerprinting;
- Only sending aggregated insights off the device (e.g. walking, joggers, but no identity);
- Only accepting anonymized and desensitized data on our cloud-based products;
- Not associating user data with a unique identifier;
- Not storing data for longer than we need to deliver value.
There is no impact on our offering from the loss of identifiers such as IDFA, as we do not rely on the use of any app or device-level identifiers for any of our live moments or cohort membership predictions.
Yes, we maintain GDPR compliance as well as compliance for all similar frameworks including CCPA and LGPD. See our privacy policy (https://portal.eu.numbereight.ai/policies-license/#software-privacy-notice) for more information.
No. As we don’t send any identifiers there is no way to target a user on an individual level, so retargeting a specific user won’t be possible.
We do not allow Clients to provide us with any information about minors under 13 years old. If we learn that personal information from users less than 13 years of age has been collected, we will take reasonable measures to promptly delete such data from our records.
Our ID-less behavioral solution (SDK) will operate with any user-facing permissions already granted to the app, which may be none at all! Our ID-less demographic solution does not require any permissions.
Location permissions are not necessary, but it will be used when it is available.
Yes. Our TCF Vendor ID is 882.
Payments
For payment-related queries, please contact us.
For payment-related queries, please contact us.
