Intro
Charts offer two essential functions for drawing insights:
Visualization of product inventory performance
Deep dives into interesting discoveries
All charts in Product Insights are interactive, with dynamic responses to clicking and filtering different data points. They present multiple ways to approach interactive data discovery:
Filtering by categories, product types, or brands: Product Attribute Breakdown
Filtering by time period (any given day, week, month, etc): Trend in Time and Product Attribute Trend
Filtering by selecting intervals of data buckets: Performance Breakdown
Note: Products change all the time. Attributes like product price change often while other attributes like product category change rarely.
The charts always consider the latest value and only show products currently falling under the applied filter conditions.
Charts are available for use at the top of the Product Insights menu. To use a chart, select the appropriate tab and calibrate the metrics to your objective. Click the ⚙️ "cog" settings button to change chart options.
Usage
Trend in Time
This chart describes how two metrics develop over time, which helps you learn about their possible correlation(s). However, the real power lies in examining these metrics over time in conjunction with the filters.
Trend in Time applications:
Discover whether increasing spend translates to increasing Product ROAS
Detect any possible urgency to implement a stop loss strategy
Filter poorly performing products, and display their spend and count over time
Zoom out to a longer period of time where it's possible to show the proportion of cost that could have been saved, and to project the cost of not excluding such products from the campaigns in the future
Product Attribute Breakdown
This chart shows your performance through the lens of product attributes such as categories and their subcategories. More importantly, it also capitalizes on the comparison of two available metrics. To switch these metrics, click the "Change sort order" button in the top right. The spend <1 filter may be necessary to get rid of categories without traffic. The chart settings also allow for the number of columns and labels for each data point to be changed.
Product Attribute Breakdown applications:
Identify the highest-spending (sub)categories
Find new opportunities by changing the metric sorting to discover promising categories with low spend but high Product ROAS
Further investigate interesting categories by clicking on them in the graph or filtering them in the Product Insights panel to see insights in the table or their overall performance on other charts
Product Attribute Trend
This chart describes your performance according to how different product attributes develop over time. The chart settings also allow you to increase instances of the selected attribute.
Product Attribute Trend applications:
Immediately see which categories are growing in a selected metric after an important change, such as setting a higher budget
Quickly discover the rising stars or cash burners among categories or brands depending on important KPIs
Follow up on discoveries by diving inside and identifying the products driving the trend, and whether it's changes in traffic, profitability, costs, etc
Performance Breakdown
This chart shows the distribution of your inventory performance. The data intervals of a histogram are useful to explore the nuances of performance according to a range of possible data (compared to a single data point, which may be overly specific for drawing broader conclusions).
Note: By default, buckets do not include products with no data ("null values") as some of the products might be out of stock or simply have no traffic per the bucketed metric. It is possible to display these products as a separate bucket in the settings of the chart to get a sense of their proportion against the rest of the inventory.
Performance Breakdown applications:
See how much spend goes to good and bad products across the whole product inventory
See how much spend (or revenue or any other metric) goes to product bucketed by price
Performance Breakdown strategy example
Performance Breakdown strategy example
Many advertisers who struggle with profitability do not take into consideration the price of the products that they are promoting. While cheap products are often favored by the algorithms and easier to sell, they are not likely to bring enough revenue to keep them profitable.
Selecting average sale price as a bucketed metric reveals that more expensive products are not promoted much. By changing spend to Product ROAS, the profitability of individual price buckets is apparent.
As the average sale price increases, so does the Product ROAS (although this is not always the case). There is a drop at the most expensive end (probably caused by the low spend), which makes Product ROAS uncertain with every sale. Nevertheless, this histogram gives a nice indication to which price buckets are useful to look at when composing a custom segment to boost profitability.