Many issues permeate the UA and One D&D feedback process, hindering collection of data that is usable as well as useful.

• Issues deriving from the format or layout:
  • Inconsistency between similar feedback types:
    • Word limit for only the more complicated parts being reviewed. You can say as much as you want about an individual feat, but an entire class is limited to 400 or 600 words between 2-3 text boxes.
    • Feedback on individual class elements is 'optional.' However, use of radio button selection means it can't be removed if you change your mind or mark the wrong row. This is in contrast to required responses which allow an 'I don't know' response.
  • Navigation issues, at least from the view of a user first encountering the form. (Solutions at the end of the form, if any exist, may as well not be there.)
    • Inability to skip forward to provide feedback in an order of the user's choosing.
      • User workaround relies on using 'placeholder' text and remembering to return. Uncorrected placeholder responses result in tainted data.
      • A 'skip for now' button would allow users to provide feedback as it was on their mind rather than in an arbitrary order.
    • Next/Previous Page buttons are slow; this is an obstacle to navigating multiple pages at a time, and is a disincentive for users to update previous feedback on future realizations.
      • Obstacles to providing or correcting feedback can result in users giving up. While users enjoy improved products, here there is no compensation to push them past obstacles in the way of feedback.
  • Consequences of losing connection, whether intentionally or unpredictably, are unclear.
    • Is feedback saved through connection disruptions such as loss of power/internet, Windows update, switching computer/browser, feedback window ending (prematurely)?
      • The answer should be clearly stated to the user at minimum before beginning the survey, ideally during the survey as well.
      • Lacking an answer, users may give less detailed feedback or skip sections they otherwise would have answered. They may also delay too long; some feedback is better than none.
    • The latest feedback window was to end 'on June 7th'. While ending at any point on that day is technically correct, any user-facing ambiguity should be resolved in favor of the user.
      • Other means of removing ambiguity include providing a warning or countdown clock on the feedback form to show the user how much time is remaining.
• Issues deriving from specific content included in the test.
  • Feedback is a collection of equations and unknown variables. Interpreting it is solving the system of equations. Too many variables or relationships makes it harder to get usable data.
    • Users have limited resources to test without knowing the designers' priority.
      • They may focus on lower priority elements while ignoring those designers really wished to test.
      • They may dilute attention so their individual feedback is less valuable, or test elements have insufficient depth of data.
  • Components with conflicting goals or interactions.
    • Testing universal spell lists alongside class specific spells.
    • Complex changes to Warlock's spellcasting: Spell slot count/level, variable spellcasting ability,
      • Special note: Multi-classing with variable ability doesn't require you to choose the ability that is 13. (A bard could have only Cha at 13+, but still select Pact of the Tome.)
  • Underdeveloped components whose predictable feedback is at best unhelpful.
    • Initial Epic Boon Feats
      • Original implementation was a 'freebie' from the DM, they didn't see other options, and the DM chose an appropriate boon.
      • UA version showed players all of the options and told them it was a level 20 class feature - where capstones used to go.
    • Metamagic
      
      • Distant Spell: Undermines existing use of range to balance spells. E.g., Thorn Whip allowing a melee spell attack at extreme range, Summon Greater Demon from a safe distance, or Feeblemind with enough distance to rather easily escape and try again the next day if they save.
      • Subtle Spell: If access to high value components is an issue, that can be corrected for all casters. This bypasses that for a single class, allowing them to save thousands of gold on spell components. Sure, they can't use this for consumed components, but there are still quite a few spells with high value components.
        • Available to Sorcerer: Gate (5kg), Plane Shift (bypasses attunement requirement), Find the Path (object associated with a location), 
        • Requiring multi-class: Imprisonment: It doesn't consume the component, but it is otherwise a complex means of preparing), Summon Fiend (humanoid blood would require questionable morals), Contingency (ends if the component isn't on your person; unclear how subtle interacts with specialized components)
      • In general, these bypass the risk of getting into range or having your gear confiscated. A Sorcerer with both can only be stopped by death, incapacitation, or be held by a mage collar. Otherwise, they can be naked and restrained, and still take a nice jaunt to another plane. With nary a whisper or body movement.