In analysis mode you can instantly view the predictive attention heat map or opacity map for any piece of content, identify specific regions of interest and view quantifiable attention scores.
Views allow you to create and save multiple analysis configurations for a single piece of content. For example, when analysing a landing page design you may wish to create one view which shows the top performing areas in opacity mode and a second view which shows the attention scores for key regions of interest like the call to action, hero images and key messaging.
2. Heat mode
The heat map that you see is the result of over 10 years of cutting-edge research in the field. Dragonfly AI heat-maps use heat to instantly visualise which regions are more or less likely to draw the attention of an audience.
Red indicates a region that is highly attention grabbing (high saliency). Blue indicates that a region is less attention grabbing (low saliency).
3. Opacity mode
Opacity mode uses opacity to visualise which regions are more or less likely to draw the attention of an audience. Attention grabbing regions are moved into the light and low salient regions are pushed into the shadows.
4. View top performing regions
In top performing mode, a grid is overlaid on top of the content and the cells with the 5 highest scores (most attention grabbing) are highlighted with the associated score displayed in the centre of each cell. In addition, the cells representing the median average for that piece of content are also highlighted with the associated scores displayed in the centre of each cell.
All cells are scored between 0-100 with 100 being the highest saliency score and cell scores are calculated as an average for that cell. The cell size can be set to small, medium or large. We recommend using the smallest grid size available to ensure the highest level of precision.
Tip: When working in top performing mode, in addition to identifying the most and least attention grabbing areas, you should also take note of the distance between the highest scoring cells and the median averages for the content as a whole. Larger distances between the two values indicate a greater level of clarity in the design as high performing regions are drawing more attention relative to the rest of the design.
In region mode you can create a region of interest by clicking and dragging on the grid that sits on top of the content below it. Once a region has been drawn you can drag to change its position, use the resize handles to modify its size and share and toggle individual cells on and off to set whether they should be included when calculating the average attention score for that region.
Tip: To toggle multiple cells within a region on or off at one time click and drag across the cells you wish to toggle. Starting the click and drag action on a cell that is toggled on will cause any cells that are turned on to be turned off when dragged over. Starting the click and drag action on a cell that is toggled off will cause any cells that are turned off to be turned on when dragging over.
Zoom in or out of the content to help with you analysis
7. A/B Test
Create a new A/B test in a single click!
To create a new A/B test:
- Upload version A for the content you want to test
- Click on Create A/B to split the screen
- Upload version B*
In A/B mode, any changes you make in edit mode (e.g. adding a region or switching to opacity view) will be reflected in both versions to make it easy to compare.
To remove an A/B test simply click on the Remove A/B button and the screen will revert back to the normal mode without a split screen.
The new regions panel makes it easy to view, rename, reorder and delete regions.
9. Capture analysis notes
While analysing content with Dragonfly AI, it's good practice to capture your thoughts and comments as you go along so that you can share these with your team later. You can do this by adding notes via the analysis panel.
Next, learn how to export your analysis as an image.
Need more information? Please contact [email protected] or any member of the Dragonfly AI team.