When viewing periodic data at a daily level, the data appears as spikes of metrics on the first day of the week or month. To avoid data spikes, you can map one or more pacing dimensions to the data stream. Pacing equally divides periodic data into multiple chunks of rows and spreads the chunks over the selected time period. Because pacing creates multiple new rows across the time period, using a pacing dimension can increase your row count.
- Mapping Measurements with a Pacing Period
When using a pacing period, it’s important to map the measurements correctly. For example, if you map impressions or clicks, their type is Integer. However, mapping an Integer type can result in the calculation being rounded down. You can avoid miscalculations by using a delivery custom metric. - Map a Pacing Dimension
A pacing dimension divides data into shorter time periods to avoid data spikes. Map a pacing dimension according to the way your data is brought in. For example, if your data stream uses only the start date, map the Start Date to the Day dimension and then set the pacing period.