A 400-item backlog isn't a planning tool — it's an archive that happens to also contain next week's work, somewhere in it. Grooming isn't about getting the backlog to zero or making every item perfect; it's about making sure the part anyone actually looks at is trustworthy.
Groom the top, not the whole list
Trying to groom an entire backlog in one sitting is why grooming sessions get skipped — the task feels endless, so it never starts. The part that actually needs to be reliable is the next one to two sprints' worth of work. Everything below that can stay rough, vague, or even slightly wrong, because nobody is about to act on it yet. Grooming the top 15–20 items regularly produces more value than grooming all 400 once a quarter.
"Ready" means something specific, not "looked at recently"
A groomed item should have a clear description, a rough size, and no open questions that would stop someone from starting it tomorrow. If an item still has "need to check with design" attached, it's not groomed — it's reviewed. Conflating the two is how teams think their backlog is in good shape right up until sprint planning reveals it wasn't.
Old, ungroomed items are information, not guilt
A backlog item that's sat untouched for eight months usually isn't a failure to get to it — it's a signal that it was never actually a priority, and writing it down once didn't make it one. Treating backlog age as data (this hasn't mattered in eight months, so it's probably safe to archive) rather than as a guilt list to eventually plow through is what keeps a backlog from growing forever.
Grooming works best as a standing habit, not an event
A backlog groomed once before a big planning cycle decays within weeks, because new information — customer feedback, a shipped dependency, a changed priority — keeps arriving. A short, recurring grooming habit, even 20 minutes a week, keeps the top of the backlog continuously true instead of accurate for one day a quarter.