ShortPixel tracks every image it manages as a row in the wp_shortpixel_meta table linked to the attachment's wp_posts row. Each ShortPixel row carries a status value such as 0 for unprocessed, 1 for processing, 2 for success, and negative values for various failure modes, alongside a compressionType field for lossy or lossless, a retries counter, and a file_size snapshot. The default bulk optimization screen shows a paginated list with progress bars, which is fine for a fresh import and unwieldy when a site already has tens of thousands of images and a few hundred stuck in retry loops.

SleekView Kanban reads the same wp_shortpixel_meta rows the bulk screen queries. Pick the status field as the grouping column and every attachment becomes a card grouped under Pending, Optimizing, Optimized, or Failed. Card fronts can show the file name, the file size before compression, the compression type, the retry count, and the upload date so a site manager can see exactly which images are stuck and why without leaving the board.

Dragging a card between columns updates the ShortPixel meta row through the plugin's own API. A move from Failed back to Pending re-queues the attachment for the next bulk run by writing status zero and resetting retries. A move from Optimizing back to Pending cancels the in-flight job through the standard ShortPixel cancellation path. Any subscriber to the shortpixel_image_optimised action keeps firing as the rows update, so existing automations that copy files to a CDN do not need to be rewritten.

ShortPixel works brilliantly while it works and turns into a black box the moment something fails. The default bulk screen is built for the first run, when every attachment is moving in the same direction and progress bars are reassuring. Six months later, when the library has tens of thousands of images and a few hundred sit in retry loops, the same screen becomes a paginated list of frustration.

Site managers either ignore the failures, which is what most do, or they spend an afternoon clicking through bulk pages trying to identify the pattern. A kanban view that reads and writes the same ShortPixel meta rows the bulk screen uses keeps the failed cohort visible, prioritized, and actionable. Pending images surface before publication.

Optimizing images carry a real progress reading. Failed images cluster by error code in a single column the developer can clear in a sprint instead of letting them rot for another year.