Account performance trend graphs appear on most tabs on the Campaigns page with the graph open. You can always close the graph by clicking the Show graph toggle.
On any graph, you can choose from eleven different metrics: average CPC, clicks, cost per acquisition (CPA), conversions (Conv), click-through rate (CTR), impressions, spend, average position, conversion rate, revenue, and ROAS along with the time comparison of daily, weekly, monthly or quarterly. Peaks and valleys in the graph show where you have had good performance and where you haven’t.
Using the graph, you can compare your recent performance:
On the campaigns page, with the account performance trend graph expanded, select a date range and turn on compare. The darker color line shows the performance trend for the recent period and the lighter color line shows the comparison graph for the previous period, the same period last year, or the moving average, depending on your selection. Weekends are marked by gray vertical bars. You can hover over the graph to see differences in current performance and past performance. You can also see a comparison summary, which shows the total difference between performance for the selected time range and for the comparison range.
You want to see how many clicks you are getting over time for a new ad campaign. Looking at the last three months, you notice that overall the number of clicks looks good, but you see some periodic dips in the graph. You narrow down the date range to last month and notice that clicks drop off significantly on the weekend. To make sure that this isn’t an anomaly, you select the "Show moving average" checkbox. The lighter line is consistent with what you are seeing over the last four weeks.
With that knowledge, you can investigate and take action. For example, are your customers only searching for your product during the week? If so, maybe target Monday through Friday only. Or maybe the competition is stiffer on Saturday and Sunday, and you need to increase your bids on the weekend. To make adjustments like this to your campaign, see How can I get my ads in front of my customers? and How to target my customers by adjusting my bids.
Moving average is calculated by averaging the performance metrics across the previous four days. All days are the same day of the week. For example, if you are looking at the number of clicks on a Wednesday, the moving average will be calculated by taking the average number of clicks for the four previous Wednesdays. Moving average can be used as a baseline because it smooths out random fluctuations. It answers the question: How did my campaign do on this day compared with a typical day? Note that moving average is only available for a time range up to 30 days.