Filtering Saves Money By Reducing Storage and Egress
Filtering ensures that metadata or fluxdata values that do not meet certain conditions are not sent to connectors, rather than excluding the entire event. This allows for more controlled data egress and helps customers focus their time and money on only sending the most relevant data to a connector.
Data filtering ensures that you are only getting the relevant data you need. This can not only reduce storage costs but also improve any computations that might need to be carried out afterwards.
Capture Only The Most Salient Data
Any message can be filtered out using any Boolean expression based on data in the message. Here are some examples:
Quality of Experience (QoE)
Refine your data collection to exclude non-fatal errors and performance variations that fall within acceptable ranges. By focusing on critical performance indicators, you can pinpoint genuine issues that impact viewer satisfaction and swiftly take corrective actions.
Enhance data relevance, for example, by excluding pause durations lasting two seconds or less. This level of granularity in engagement data ensures that you're accurately tracking meaningful interactions, providing a clearer understanding of user preferences and content resonance.
Tailor your data collection to exclude outlying device types or irrelevant geographic regions, allowing you to concentrate on interactions that matter to your business. This refined data set enables better device-specific optimizations and targeted user engagement strategies.
Fine-tune your data by filtering out that which is associated with content items that were played back while a certain percentage of the player wasn’t visible, focusing on engagement as you define it. This enables content teams to make informed decisions about content recommendations and production investments.
Optimize your ad performance insights by filtering out ad interactions that don’t contribute to ad revenue, such as ad impressions from certain IP addresses, enabling your ad teams to refine strategies and placements effectively.
With data filtering, companies can choose how much of a specific measurement to deliver, and how often.
Data Filtering In Action: CDN Log Data
Data processing costs resources: compute cycles and time. Although the compute cycles resources required to churn through a large dataset like a CDN log may not be exorbitant, you may not be able to afford the time when trying to narrow down performance impacts on QoE. By taking a representative sample, such as 10% of that CDN log dataset, you may be able to reduce your compute and time costs by exponential amounts. Datazoom’s Sampling Feature enables you to specify the amount of data you need from a specific source.