got a lot of good pushback on my last post about the "milestone locked = billable" thing. someone in the comments reframed the whole problem as blast radius instead of size or hours, and it stuck with me. so i sat down to sketch out how i'd build that into the tool and immediately got stuck, before writing any code.
sorting requests into buckets isn't the hard part. cosmetic stuff like copy or colors, just log it and move on. anything touching data, auth, or something already approved, that needs an actual review.
what i can't figure out is who does the sorting.
if it's me tagging every request that comes in, i'm back to manually triaging everything, which is the whole thing i was trying to get away from. if the client tags it themselves, every request turns into "just a small tweak" because nobody rates their own request as the risky one.
only idea i've got so far is giving them a short list to pick from when they submit something, like visual tweak, content change, new behavior, something's broken, and letting that map to a risk level on my end instead of asking them to judge the risk directly. but even that breaks down fast, like a "visual tweak" that happens to touch a component used in three other places.
haven't even gotten to the change budget idea some of you brought up last time either, wanted to sort this part out first.
anyone dealt with this before, where the person doing the classifying has a reason to underplay how risky it is?
Top comments (0)