Viewability Predictor Integration
Viewability Server Integration
Demand connection (Viewability Server)
- Connecting Native (fluid) ads to Browsi engine
- Header Bidding support (Body Bidding)
- Header Bidding Demand Integration – Prebid.js
- Appnexus Demand Integration
- Collegare la tua richiesta DFP (Italiano)
- Conectando su demanda DFP (Spanish)
- Connectez votre demande DFP (French)
- DFP demand integration
- AdForm Demand Integration
What is Browsi?Browsi uses AI to solve ad viewability for publishers by automatically optimizing page/user-level ad layouts for viewable impressions and predicting viewability for higher ad yield.
How does Browsi Viewability server work?The way that Browsi viewability server works:
- Page analysis - Browsi analyzes and maps the page structure in real - time
- Ad placement creation - Using machine learning the highest viewable ad layout is suggested.
- Connection with the publisher’s ad server - each Browsi placement will send an ad request, with viewability prediction paired to the af request. The publisher’s ad server will match the ad reuqest and Browsi engine will embed the impression on the page. If there is no impression matched, Browsi engine will not open the placement. After the impression has been embedded, Browsi will track it for viewability.
- User engagement benchmarking - user behaviour metrics like scroll depth and velocity, article bounce rate, time on page and more are benchmarked to guarantee user experience is not compromised.
- Ongoing page optimization - As traffic to page increases, the AI learns and improves its ad layout and viewability prediction.
how does Browsi viewability predictor work?Browsi’s algorithm collects data on your pages, user (non-personally identifiable), ads and many more. Once enough data is collected, Browsi’s algorithm is ready to send viewability predictions, for each ad request sent from the page. The prediction is sent to the publisher ad server, using pair of key-values stating the predicted viewability. For more information about Browsi’s viewability predictor click here
What is the business model?Browsi provides all core products on a SaaS serving fee based on the page view - amounting to a fraction of page RPM. This model enables unlimited impressions on a single page regardless of whether a partner replaces a single unit, seeks a new one, or multiples of both.
How does a test with Browsi look like?The most efficient way to test Browsi is to take a baby step towards upgrading the logic around a single ad placement. To test this, the publisher identifies a low-viewability placement, removes it for 500k to 1M impressions over 5 days, and let’s Browsi dynamically find a more viewable placement with the publisher’s demand connected. The publisher keeps 100% of the revenue during this period. At the end of the test, the publisher will see the delta in viewability between previous performance vs Browsi’s dynamic, real time logic at the page level. During this test, a small percentage of traffic will be allocated for learning and benchmarking, while scanning and detecting all page assets as they load – image density, text size and length, location and frequency of existing ad units, videos and more. With that deep understanding of the page structure in hand and the hard numbers generated from testing a single placement, the publisher can quickly grasp the value of rolling out Browsi to all placements on the page for a holistic improvement to supply, and a host of strategic related levers that subsequently become a available.
How do we proceed after testing?Following 5 days of testing and with a package confirmed, we coordinate with your team and begin rollout across a full site / network with any custom details on demand source(s) – publisher’s direct sold or programmatic/PMP, header bidding, and/or Browsi programmatic partner demand – it’s all possible.
Is Browsi hands on once everything is up and running?Absolutely. Browsi’s Success Management Team keeps a close eye on performance, connecting on ways to continue optimization toward meaningful RPM lift borne from individual pages to full sites and networks.