Best Practices¶

This article describes best practices for generating reports using the DSP Reporting API.

Time Series Report¶

Size	Normal
Description	`dimensionTypeIds` is an empty array. The number of report rows is determined by the time interval and normal granularity.

{
  "reportOption": {
    "dimensionTypeIds": [],
    "metricTypeIds": [ 44, 1, 2 ],
    ...
  }
}

Group By Report¶

Size	Huge
Description	`dimensionTypeIds` array has 1 or more values. Report row count is the product of cardinality of each dimension and time-axis. This would create a large report.

Top N Report¶

Size	Normal
Description	`dimensionTypeIds` array has only 1 value and there is `limitSpec`. The report size is the specified limit multiplied by size of time-axis. The example generates a report of `Top 10 cities of Ad spending for some time period`.

{
  "dimensionTypeIds": [21],
  "limitSpec": {
    "columns": [
      {
        "metricTypeId": 44
      }
    ],
    "limit": 10,
  }
}

About End User-Related Dimensions¶

Reports contain ad impression events (including clicks). An event has both advertiser-related dimensions and end-user-related dimensions. Restrict the use of end-user related dimensions, which are high cardinality within a time interval. The following table lists the dimensions that are end-user related.

ID	Name	Description	Usage Restrictions
9	Domain Name	The unique Internet address for a website	Yes
27	Age	Demographic dimension: 13–17, 18–20, 21–24, 25–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65+	Yes
25	Browser Type	Safari, Chrome, Firefox, IE, or Unknown	Yes
21	City	Geographic attribute for the city of a visitor	Yes

Note

Use filterOptions with dimensions to avoid large reports. Avoid grouping a report on these dimensions. Instead, request a top-n report. For example, it’s best to report on the ad spend in the top 10 cities rather than ad spend for all cities.

Note

Timeframe and Granularity: Very granular reports that span long time periods, such as a year-to-date daily reports, are not permitted in a single API call. For these types of reports, break up the time period into multiple API calls.