How does the algorithms analyze data?
  • 01 Oct 2022
  • 1 Minute to read

How does the algorithms analyze data?


Article summary

On most algorithms pages, the Settings tab contains parameters that control the start and end date to aggregate, analyze, and evaluate advertising performance.

rules_settings

Days of recent data to ignore

This parameter controls the end date.

If today is 31-Jan, and you set this value to 2 days, the end date will be 29-Jan.

The presets for most algorithms are 1-2 days.

If you want to change the preset value, consider these first:

  • When a shopper clicks on an ad and purchases a few days later, the sale attribution will only show up later. 
  • Additionally, advertising data can occasionally delay for up to 72 hours
  • Factor sales attribution and data latency to exclude days with a very high % of incomplete data.
  • Most sales happen within a few hours after a shopper clicks on an ad.
  • There is generally enough data after 24 hours to evaluate performance.

Days to analyze (Up to 90 days)

This parameter determines the start date.

If the end date is 29-Jan, and you set this value to 3 days, the start date will be 27-Jan.

When an algorithm runs, it will analyze 3 days of advertising data from 27-Jan to 29-Jan.

Having a shorter period enables better adaptability to changes:

  • Analyzes recent data, factoring recent changes, such as changes in bid, placement, price, conversion, and seasonality.
  • Ideal for high-traffic accounts.

Having a more prolonged period enables decision-making with more historical data:

  • Lesser occurrence of making incorrect optimization due to unsustainable performance (Etc. keywords with 1 click and 100% conversion)
  • Ideal for low-traffic accounts.

Was this article helpful?