Stop Arguing It, Start Counting It
A deploy left a required picklist with only "-- Clear --" in it and took the whole lead-capture journey down. The response was not another bug report. It was a log with hours in it.
The dropdown had one entry in it. "-- Clear --".
The field was required. So no Lead could be created, which meant the entire lead-capture journey was down, which meant everything queued behind it was down too.
An OmniScript deploy had landed that morning.
Two Bugs Wearing the Same Face
Before the log, the diagnosis — because this part takes seconds and routinely takes hours.
"Required field the user cannot set" is not one defect. It is two defects with different owners, and they look nearly identical on screen.
The first is a picklist whose value set is empty or unavailable in this context. The control renders. You can click it. There is simply nothing in it to choose. That is a configuration problem, and it belongs to whoever owns the field's metadata.
The second is a field locked by Field-Level Security as Read-Only or No Access. The control is greyed out, or it is not on the screen at all. That is a security problem, and it belongs to whoever owns the permission sets.
One question separates them: does the control render? A dropdown that opens and offers nothing points at the value set. A greyed or missing field points at FLS.
Juniors log both as "field is required but I can't fill it", and that ticket then bounces between dev, admin and security for half a day, because nothing in it says which of the two failure modes anyone is looking at. Name the mode in the summary and the hot-potato game does not start.
Ours was the first kind. Rendered, empty, config.
The Part That Was Not a Bug
We could have raised the ticket and moved on. We had done exactly that several times before, for the same underlying reason: a deploy went out untested, QA lost the environment, somebody fixed it, everybody carried on.
So instead of raising one more bug, we started a log.
One row per incident. What was deployed. What broke. How many wall-clock hours QA was down. Whether the deploy was announced beforehand. Whether the fix was ever acknowledged.
That is the whole instrument. Five columns, one row, thirty seconds to fill in.
Why Counting Beats Arguing
A single blocked afternoon is an anecdote. In a room under delivery pressure, an anecdote is waved away as bad luck, and honestly it should be — one bad morning is not evidence of anything.
The same anecdote logged eight times, each with hours attached, is not an anecdote any more. It is a rate. It has a total. It can be put in front of a retrospective or a technical design forum as a number rather than a mood, and numbers do not require the person presenting them to sound aggrieved.
And there is a reason the missing smoke gate never gets built. The cost of not having it lands entirely on QA, silently, in hours that are never totalled anywhere. Nobody is refusing to pay for the fix. Nobody knows the price.
The log is how the price becomes visible.
What This Was Not
It was not a complaint with a spreadsheet attached. The rows are neutral and they are boring on purpose — what was deployed, what broke, how long. If a row shows the deploy was announced and the fix was acknowledged inside an hour, that row goes in too, and it argues against us. That is what makes the rest of the rows worth reading.
It was also not a replacement for the bug. We still raised the defect. The log answers a different question — not "what is broken" but "how often, and at what cost."
The Lesson
When a process failure keeps recurring and keeps getting waved off, stop making a better argument. Start keeping a count.
One lightweight row per incident, with a number attached. Hours down. Deploys announced. Fixes acknowledged. Nothing more elaborate than that, because anything more elaborate stops getting filled in by week three.
A hundred individual bug tickets will never earn the process change, because each one is closed on its own and closing it is the end of it. A tally of eight rows will, because a tally is the only artefact that shows the shape of the thing rather than one instance of it.
The act of counting is the intervention.
What has blocked your team three times this quarter — and could you say how many hours it cost, without guessing?