I was just reading that if you have unfinished items at the end of a Sprint, you should put them back into the backlog and re-sort by value, or, in the worst case, roll back the work. That’s awful advice.

You are creating Lean “inventory”—work done (and money spent) that doesn’t generate revenue. Inventory is a huge source of waste. If you’re not “finishing,” you need to learn to work in smaller increments and narrow the scope of work to something that can be done in a couple of days at most. Anything else is just a band-aid over a bad process. It takes work to make the work small, but not doing that is just laziness. If the work was so unimportant that it’s okay not to finish it, why were you doing it in the first place? The item never should have been on the backlog.

Frankly, I’d dump the backlog entirely. It is not helping you. Instead, do the most valuable, smallest thing. Get feedback. Decide what to do next based on that feedback. The Sprint is a periodic review cycle, not some bag you fill with work. It’s not some tragedy if the bag isn’t empty when the review hits. Sure, review on a regular cycle, but if you’re in the middle of something when the review happens, just keep working after the review. I don’t see the problem.

~ Allen Holub