Data Ingestion

Rockerbox is able to ingest OTT data in two ways:

OTT Device Types

There are different mediums through which OTT ads can be served.

  • Device: These are ads served to users watching on devices such as an Apple TV, Roku or Amazon Fire Stick. This distinction is made by looking at the user agent of the impressions.
  • App: Ads served to a user inside an app on a mobile device (phone or tablet). An example of this is someone being served an ad while using the Hulu app on their iPhone. This classification of App is known by looking at the user agent of the impression as well as ensuring that there is no referrer on the request
  • Web: Ads served to a user on a traditional device (desktop, laptop). This classification is known by looking at the user agent as well as ensuring that there is a referrer and domain.

OTT Methodology

  • Rockerbox uses a hash_ip (SHA256) as our primary identifier to find a match between an OTT marketing event and an on-site session.
    • Shared IPs: Since we use hash_ip as our primary identifier to find matches between OTT Marketing Events and on-site sessions, we need to eliminate the scenario where multiple people within an office or home shared a single hash_ip. If the events served from single hash_ip exceed a threshold, we consider it as a shared hash_ip and perform a more strict rule on it to avoid over-crediting the IP.
  • Rockerbox then looks to ensure that the window between an OTT Marketing Event (i.e. impression) and on-site session, falls between a client and device-specific time window. Assuming the OTT Marketing Event does fall within that window, Rockerbox includes the OTT Marketing Event as a touchpoint on the path to an eventual potential conversion.
    • The Rockerbox approach of looking at the time between an impression and a session does differ from traditional attribution windows that look at the time between an impression and a conversion.
  • Web touchpoints: Rockerbox only considers touchpoints if the referrer is direct (i.e. the user came straight to the website) or via search (i.e. the user came to the website through paid or non-paid search traffic)

Eligibility Time Window

Determination: Rockerbox automatically determines the proper Eligibility Time Window for each client. Rockerbox further determines the proper Eligibility Time Window for each device type for each respective client. These calculations are based off of each client’s specific data (i.e. marketing, funnel, sales cycle).

Customization: Rockerbox enables each customer to customize the Marketing Event to Session window on a per device basis. Changing the Eligibility Time Window will apply only to future OTT Marketing Events – it is not backfilled retroactively for historical OTT Marketing events.

  • The dropdown has default time options based on your data for that specific device
  • You can enter times manually outside of the options presented. The input value is the number of minutes you want as your eligibility time window
  • Note: Time windows are currently limited to 1440 minutes (24 hours). By default, a 300 minute (5 hour) eligibility window is applied.
  • Note: Impressions and sessions must currently occur within the eligibility window. If the session occurs outside of this window, then the session is not eligible and the impression won’t be included on the path to conversion.

Field Definitions

Field Name


Total Events

Total number of marketing events available if using the max time window (600 minutes)

Time Window

The configured eligibility time window

Eligible Events

The number of marketing events filtered based on the eligibility time window

% Classified

The percent of events within the Time Window compared to the total number of marketing events available

Tatari OTT

Tatari OTT attribution methodology is an exception to the above description. This is due to a limitation in the data they share with Rockerbox, as they do not pass the user’s IP address.

Instead, Rockerbox syncs user-level data based on a unique Tatari ID. Tatari then passes attribution log files to Rockerbox log files for OTT events based on those Tatari IDs. Rockerbox then incorporates these events into our attribution.

How did we do?