An image’s Vizit score indicates how well that image performs against a benchmark, through the lens of a specific audience.
Note: This article explains how to view the score for a single image. You can also download Vizit scores for all images in a folder or project, or use the Compare feature to view multiple scores side-by-side, for up to 10 project images.
To get started with a benchmark comparison, open the project or open the folder that contains the image. If you have assigned tags, you can use tags to segment a set of project images.
Select the AI Task button and select the Generate Vizit Score task from the AI Task menu
Next, select the audience lens to apply to the project to drive the first step of generating the Vizit Score
Then choose the benchmark for comparison
This will then generate the Vizit Score on the project assets
Find the image whose score you want to see, and double-click it.
At the top of the analysis page, your selected audience is listed, which you can change. To the right of the audience is the active benchmark. If the project contains more than one benchmark, you can select the benchmark you want.
Note that although multiple benchmarks can be added to a project, only one benchmark at a time can be used when viewing an image’s Vizit score. But the Compare feature enables you to simultaneously compare one or more project images against one or more benchmarks.
Along the panel on the right is the Predictive Analysis section, where the image's Vizit score is displayed. In this example, the image outperformed 83.5% of the images in the selected benchmark, when analyzed through the lens of the Amazon shoppers audience.
To see where the project image lies within the benchmark, click More detail.
This displays a chart showing the score distribution of all benchmark images, from 0 to 100. In this example, there are high quantities of low-scoring benchmark images and fewer high-scorers. The green dot represents where the project image score falls within the range of benchmark image scores.
You can also take a deeper look into the benchmark visual trends, view the image’s Distinctiveness Rating, as well as image maps, color palette, and metadata.
Changing the audience and/or benchmark will update scores and ratings.