Understand what our SDK and location intelligence solution means for retail brands and publishers.

Accurat UNDERSTAND offers insights in real-world behaviour
from store visits & loyalty to intended behaviour

Our proprietary algorithms automatically appoint users to a diverse set of audiences, based on their real-world behaviour on an ongoing basis. Advertisers interested in specific behaviour audiences can combine specific location visits or even behavioural patterns to define custom segments.

Accurate matching serves as a solid basis

Our high accuracy matching algorithms calculate the propensity of a visit to a location. Accuracy is improved by ground truth measures like a user’s previous behaviour and context data of POIs (e.g. opening hours).


Matching is the basic algorithm on which all our products are built. High accuracy is key and mostly influenced by the data fed to the algorithm. It is the main reason why we use building shapes instead of coordinates or why we skip low accuracy trackings.

Segment users in 200+ audiences across 20+ categories

We automatically tag users in segments. We split heavy users from regular visitors of 20+ categories including supermarkets, DIY, fashion, beauty… We know who’s a sportive, a culturist and know to separate a convenience shopper from a shopaholic.


Determining segments is an ongoing business. By tracking users over time, we are able to determine very exact what their preferences are. Some behaviour takes time to measure, others are easier captured.

Create custom audiences using our audience builder

Your segments are defined by a set of specific metrics? Our segment builder allows to create specific audiences. Combine POI visits, set frequency requirements, define specific brand preferences and more to define your segment of choice.


Create segments to mimic the personas your (advertiser’s) business attract. Define who’s a John, a typical “low frequency weekend buyer” or who’s rather a Marie, the “non loyal shopaholic with kids”.

Ongoing pattern and anomaly detection

On an average day we collect between 70 and 100 data points for every app user that you track using our SDK solution. Once patterns are understood by our AI solution, we constantly monitor deviations to build “predictive segments”.


Building patterns takes days, weeks or even months, depending on the user’s behaviour. Detecting supermarket preference or a weekly running behaviour is easy, defining a user’s “gifting behaviour” potential takes more time.

From churn prediction to predicting the next buy

Our predictive models constantly measure pattern deviations to define the propensity of an event happening. A drop in a regular pattern is categorized as churning behaviour, visits to a new category are considered “next buy” signals.


Our algorithms are smart to separate noise from voice – true preditive signals. Not every drop is a churning signal: context – being abroad for holidays, weather conditions, festivities … – suggests drops can be ignored.

Augment predictive power using your own data

Just like our segment builder, we believe flexibility is required to make the accurat solution tailored to your objectives. Therefore, our API allows you to connect your data with ours or vice versa.


Our API can handle three types of data. User data – whether in-app interactions or static data about your app user, location data – about your venues – or context data – campaign efforts, events you host, … can be shared.