How do you take over a stale backlog and make it productive?

March 2023

Context

When I joined the StoreKeeper SaaS team, I inherited a backlog with over a year of accumulated tickets. Many were vague, duplicated, or described problems that had already been worked around. The team had no formal prioritization — tickets were picked based on whoever shouted loudest or whichever customer complained most recently. Developers spent more time interpreting tickets than building software.

Decision

I treated the backlog as a codebase: it needed a refactor, not a rewrite. I didn't throw everything out and start fresh — that would have lost legitimate context from customer conversations and past decisions. Instead, I ran a structured triage over several weeks.

Step 1: Categorize every ticket as active (still relevant), stale (problem may no longer exist), or duplicate. Stale tickets got a comment explaining why and were closed — not deleted, so the history survived.

Step 2: Rewrite surviving tickets to a consistent format: what's the problem, who does it affect, and what does done look like. Vague tickets like "fix sync issues" became specific: "Stock sync fails silently when Shopify rate-limits bulk updates — add retry with exponential backoff and surface the failure in the merchant dashboard."

Step 3: Establish a prioritization framework tied to business impact. Customer-facing bugs ranked above internal tooling. Revenue-impacting issues ranked above convenience improvements. This removed the "loudest voice wins" dynamic.

Consequences

Positive:

The team cleared 100+ tickets in 6 months, including the year-old stale backlog
Sprint planning dropped from 90+ minutes of confusion to 30 minutes of focused discussion
Developers reported less frustration — they could pick up a ticket and start working without a 20-minute archaeology session

Negative:

The initial triage took significant time that could have gone toward feature work
Required a substantial number of hours from the already-busy CEO every week to ensure the overhaul would work

Neutral:

Backlog hygiene is not a one-time event. Without ongoing grooming discipline, any backlog returns to chaos within a few months. The triage process I built became a recurring practice, not a project.

Context

Decision

Consequences

Positive:

The team cleared 100+ tickets in 6 months, including the year-old stale backlog

Sprint planning dropped from 90+ minutes of confusion to 30 minutes of focused discussion

Developers reported less frustration — they could pick up a ticket and start working without a 20-minute archaeology session

Negative:

The initial triage took significant time that could have gone toward feature work

Required a substantial number of hours from the already-busy CEO every week to ensure the overhaul would work

Neutral:

Backlog hygiene is not a one-time event. Without ongoing grooming discipline, any backlog returns to chaos within a few months. The triage process I built became a recurring practice, not a project.