Take a close look at each element in your existing rubric. The only real art to creating an EssayTagger rubric is to think carefully about the "scope" of each rubric element: sentence-level, paragraph-level, or whole document.
An element like "Evidence" in a traditional rubric is very general and is assessed at a very broad level. For example: "Some of your evidence was weak or insufficient or not entirely relevant." Or, "Included 1-2 pieces of strong evidence but also included some instances of ineffective evidence."
In the EssayTagger world you have the ability to get much more specific than this. We recommend taking an element like "Evidence" and configuring it in an EssayTagger rubric to be a "sentence-level" rubric element. That means you will be evaluating each piece of evidence individually as they appear in the essay. If the student includes three pieces of evidence, you'll evaluate the quality of each one and provide specific, targeted feedback for each one. The grading app will then average together your Evidence evaluations to produce a single Evidence result for data analysis purposes.
An element like "Body Paragraphs" can likewise be broken down to a more specific level. In this case we'd recommend the "paragraph-level" scope. You can then identify and evaluate each individual body paragraph and provide specific feedback for each one. For example, one body paragraph might merit "Strong topic sentence and paragraph stays on-point throughout" while another gets "No discernable focus here; totally unclear what you're trying to accomplish in this paragraph." This sort of feedback is so much more valuable than the generic, high-level version: "Mostly strong paragraph focus but with some paragraphs needing more work."
There will still be rubric elements that call for a high-level, whole document scope. Elements like "Overall organization" or "Voice" can be difficult to isolate down to individual chunks and are best considered against the essay as a whole.