Features
Part of A practical reference for building operational checklists that hold up
Running new operational checklists: the two-week trial
How to create operational checklists that survive real use: a two-week trial, what to watch for, what happens to completed lists, and when to review.
A draft checklist is a guess about how work happens. The first two weeks of use are the test of that guess, and they decide whether the list ends up on a wall and ignored or in daily use.
What to take away
- Write one list, run it for two weeks, and cut it before writing a second.
- Give it to the people who do the work, ask them to run it exactly as written, and watch without helping.
- Handwritten additions are the highest-value finding, because someone has told you what is missing without being asked.
- Decide what happens to a completed list before the list goes into use.
Before the first run
Most checklist projects die the same way. Somebody decides the business needs checklists, writes fourteen of them in a fortnight, and three months later one is on a wall, laminated, out of date, and ticked by nobody.
The way out is to hold the number of lists at one until the first has been through a trial. The draft itself should come from what has gone wrong, as the operational checklists template lays out, rather than from a task inventory.
Check you handed over the right kind of list
The first thing a trial can show is that the wrong kind of list was written. There are two shapes, and the choice determines who runs the list and where it stops.
| Kind | Who it suits | What goes wrong when mismatched |
|---|---|---|
| Read-and-do | Someone new to the task, working step by step | Handed to an expert, it is insulting and gets ignored |
| Do-and-confirm | Someone who already knows the task, pausing to check | Handed to a beginner, it is useless |
Read and do vs do and confirm
Read and do
- When
- Rare or expensive
- Doer
- New to task
- Shape
- Numbered steps
- Run at
- Read then perform
- Confirm
- On each step
Do and confirm
- When
- Risk is missed step
- Doer
- Knows the work
- Shape
- Short list
- Run at
- Named pause point
- Confirm
- At the pause
Getting this wrong is the most common design error. If the person running the list looks bored or lost in the first run, suspect the kind before you suspect the person.
How to run the trial
Keep the trial small and unceremonious. The aim is to see the list used as written, by the people who will use it, in the place it will be used.
- Hand the list to the people who do the work and ask them to run it exactly as written.
- Watch the first run and say nothing, even when you want to help.
- Watch at least one more run on a different day, ideally with a different person.
- Collect every handwritten addition and every item someone skipped or hesitated over.
- Cut the list, then run it again before anything else is written.
The same pattern appears wherever checklists are introduced into real work. The World Health Organization notes that ensuring successful implementation is one of the biggest challenges of its surgical safety checklist, and it launched an implementation manual alongside the checklist to help with introduction.
What to collect during the two weeks
Keep a plain note with one line per observation. Quote what people say in their own words rather than summarizing, because the wording is often the finding.
What to collect in two weeks
- Items ticked without being checked
- Items understood differently twice
- Failures missing from the list
- Actual time against budget
- Anything the operator added by hand
Handwritten additions are the highest-value finding on that list. Someone has told you what is missing without being asked, and the extra item is usually the private check that person has made for years.
Items that fail the trial
An item that people hesitate over, or tick without looking, usually names an intention rather than a state. "Check the door" can be ticked whether or not the door is bolted. "Rear door bolted, from inside" can only be ticked by looking.
Rules for writing observable items
- Name a state, not an intention
- No item containing 'and'
- Name the role on the item
- Put stop condition at the decision
- Say what to do when unconfirmed
- Use the workers' own words
Rewrite such items in the words the people doing the work use out loud. A list written in management vocabulary gets translated in the reader's head each time, which is where the errors return. The worked examples show the difference item by item.
Cutting after the trial
Every list has a real time budget, whether or not anyone has written it down: the time available at that moment in the work. A list that overruns its budget does not get run more slowly. It gets signed.
Ask one question per item: what specifically goes wrong if this is missed? An item with no specific answer comes off. Expect to remove between a third and half of a first draft, and expect that to feel wrong. The people who do the work supply the items, but someone else holds the pen and does the cutting, because practitioners will not remove their own.
Deciding what happens to a completed list
Answer three questions before the list goes into regular use, or the practice will be abandoned within a year.
How long is a completed list kept?
Who reads the completed lists, and when?
What happens when one shows a problem?
Setting a review trigger and stopping
Fix the review on triggers, not dates. A change to the equipment or the system, an incident where the list was involved, a new person struggling with it, or any handwritten addition appearing twice. The reasoning is in operational checklists trends.
Then stop writing lists. The set is finished when the failures you collected at the start are covered, and the remaining tasks sit on a note saying why they have no list. Publish that note, because it prevents the well-meaning expansion that kills the practice.
Common questions
How many lists should a small business have?
Usually two or three: a closing list, a handover, and one for the task that has gone wrong expensively. Beyond that the maintenance cost grows faster than the benefit.
Should a list be laminated?
Only after it has been run for a month unchanged. Lamination makes correction impossible, which is fine for a settled list and fatal for a new one. Layout rules are in operational checklists format.
What if people refuse to run it?
Ask them why, and believe the answer. It is nearly always that the list is too long, badly timed, or that nothing ever comes of it. All three are your problem, not theirs.
Where does this fit with the rest of the checklist guidance?
The operational checklists overview covers what a checklist should contain, and the trial described here comes after that draft exists.
