The Priority Dropdown Trap

A Startup Story

Without fail, your Jira Admins will have a solution: behold a Priority field dropdown! (Pro tip enterprise devs: possibly named Priority2 or P-level.)

Curiously, 100% of companies choose between P1, P2, P3, P4 or Low, Med, High, and Critical — apparently there are no other options.

A hard-coded list of four options? Okay. Let’s try it for a few weeks…

2 Days Later

In a shocking-to-no-one turn of events, the organization discovered a ticket with a new, higher priority, necessitating a minor hack: adding P0, or Critical Max+!

3 More Days

Our intrepid boss had some exciting meetings & discoveries at the conference!

Somehow they uncovered an even higher priority than P0!

Since then, the team has been heads-down researching how to label this new Priority.

Maybe -1? No, no. That’s too confusing (P-1 vs. P1). Ok, how about P0.5, no?

In an “inspired” moment, the team invented a higher priority: the double-zero!
Now known as the P00 Priority.

Before the Flood

Before anyone realizes it, your team is somehow drowning in P00 tickets!

How can we avoid this silly game of Engineering Theater?

What if Priority Wasn’t Multiple Choice?

How could we better represent an ever-changing, fluid human concept like Priority?

• In the real world, priorities constantly shift and evolve based on new information, market changes, and organizational goals.
• There’s often a complex interplay between urgency, importance, resource availability, and cost/risk analysis that a simple dropdown can’t capture, especially over time. (Ticket rot.)
• Different stakeholders may have conflicting views on what constitutes a high priority, making a one-size-fits-all approach unsuitable.

So, what next?

There are several alternative approaches worth exploring, from low to high effort:

• To provide more leg and headroom, pick a “neutral” starting value, say 100 or 1,000. You can always increase or decrease the number.
  • Or stick with zero to start, where higher numbers have higher priority.
• Implement a multidimensional prioritization system that considers factors like business value, urgency, and effort required. (Create a composite score to make sorting and filtering easier.)
• Adopt a dynamic prioritization method, such as the MoSCoW technique (Must have, Should have, Could have, Won’t have), allowing regular reassessment. (See also the Kano Model.)

Summary

So much is put on Priority despite its rapid rate of decay. Yesterday’s CRITICAL tickets are not likely to be next quarter’s CRITICAL tickets.

Over time, older high-priority tickets become resistant to cleanup and maintenance. After all, who wants to lower the Priority of something once declared essential? Never mind deleting said irrelevant tickets… (Gasp! Think of the backlog!)

I’ve seen multiple companies mix up Severity & Priority. Severity describes urgency (or time-sensitivity.)

Priority ≠ Severity. It can make sense to define 3-5 severity levels (often used to maintain Service Level Agreements.)

The levels of urgency help communicate zero customer impact to a partial/complete service outage.

A word of caution

Deploying an unbounded Priority field requires a little planning & discipline!

If you are familiar with front-end development, you may have experienced a z-index war.

Essentially, z-index lets designers set any positive integer to ensure their widgets show “above” other lower z-index content.

Even a minor component update could introduce a change to the z-index on their <Dialog /> - making it suddenly invisible. These situations could become messy as our third-party components, feature work, and other team contributions tried to out-z-index each other.

Z-index was once limited to ~32,000. However, I recently saw a snippet with a billion z-index: 1000000000!

Inflation must be hitting z-index hard.

Discuss

• Is this a worthy thought experiment?
• Is the idea of an ever-increasing Priority horrifying? Anxiety inducing?
• Is it inevitable that this approach eventually exceeds 64-bit integer limits?
• Can other fields (beyond Severity or Urgency) add to this conversation?
• How much blame does Jira deserve? Or credit?

We could scream into the internet, “Who’s going to clean all these P00 tix?”

Or, you can get real about your backlog.

• Accept that 90% of your 1,000 tickets will never get done. It’s ok.
• Archive tickets that have been untouched for months. Any initial priority/urgency is no longer applicable. Anyway, archived issues can often be retrieved.
• When an issue comes back, that’s cool; it just increased its priority.
• Anecdotally, I’ve noticed zero harm from discarding older & incomplete tickets.
• Endlessly appending to a backlog-as-database misses the opportunity to focus your team & organization on what matters. (Things ahead of us. Whereas backlogs inherently look backward.)
• A deep backlog ends up like a Bizzaro Trophy Room, celebrating the shit you’ll never ship.