First: an SOP and a checklist are not the same document
An SOP explains. A checklist executes. Trying to make one document do both jobs is why procedures get skipped, done out of order, or “done” differently by every person on the team.
Keep in the SOP (reference)
- Why the procedure exists and what it protects
- Roles, responsibilities and approvals
- Policy language, definitions and background
- Exceptions, escalation rules and references
Move to the checklist (execution)
- Actions only, in the exact order they happen
- One action per step, starting with a verb
- The visible result that proves the step is done
- Warnings, immediately before the step they guard
The SOP stays the source of truth. The checklist is generated from it, linked back to it, and regenerated every time the SOP changes. Nobody should ever edit the checklist alone and let the two drift apart.
The 6-step conversion method
Read for actions, not for topics
Go through the SOP once and highlight only sentences that tell someone to do something: verify, open, record, send, inspect, replace. Ignore headings, introductions and explanations for now. In a typical SOP, the real actions are often less than a third of the text — that is normal, and it is exactly why the raw document is hard to work from.
Put the actions in working order
Sort the highlighted actions into the order a person physically performs them — not the order the document mentions them. Watch for hidden prerequisites: a step that says “confirm the backup completed” belongs before any step that changes the system, even if the SOP mentions the backup three sections later.
Split until every step is one action
Break any step containing two actions joined by “and”, a decision, or a wait. “Inspect the unit and log the results” is two steps. Each step should take one person, one tool, one moment. If a step needs a decision, write the decision as its own step with the two outcomes spelled out.
Give every step a checkable outcome
End each step with what the person can see, measure or point to when it is done: a value in range, a photo taken, a name recorded, a light that is green. If nobody can tell whether the step happened, it is not a step yet — rewrite it until the result is observable.
Move warnings next to the step they protect
Pull every warning, caution and “important” note out of the introduction and place it immediately before the step it guards, in one line. A warning read ten minutes before the risky action does not protect anyone. Keep the warning separate from the step itself so the step can still be checked off cleanly.
Test it on a newcomer, then fix the source
Give the checklist to someone who has never run the procedure, with the SOP available but not open. Every question they ask marks a step that is unclear, out of order or missing its outcome. Fix those places in the SOP first, then regenerate the checklist — fixing only the checklist guarantees the problem comes back at the next revision.
A worked example (template, not a case study)
The same short procedure, before and after conversion. This is an illustrative example written for this guide — not a real customer procedure.
Before — as written in the SOP
“Prior to commencing the monthly server-room inspection, personnel should ensure that the backup status has been reviewed and that any anomalies have been duly noted in the appropriate log. The coolant system, power distribution and environmental controls are to be inspected in accordance with the relevant sections, and a failover test should be performed where applicable, after which results are to be documented.”
After — as a checklist
- Open the backup dashboard. Confirm last night's backup shows “Completed” for every server in scope.
- Write any failed or skipped backup in the inspection log, with the server name and time.
- Check the coolant level gauge. It must read inside the green band — record the value.
- Inspect each power distribution unit. Record any breaker that is off or warm to the touch.
- Warning first: the next step briefly switches load — confirm no maintenance window conflict before running it. Then: run the failover test and watch the secondary unit take the load.
- Log all results, sign the inspection record, and file it with this month's records.
Nothing was added and nothing was removed. The actions were found, ordered, split, and given outcomes a second person could verify.
Rules that keep a checklist usable
| Rule | Why it matters |
|---|---|
| Start every step with a verb | “Coolant levels” is a topic. “Check the coolant level” is an instruction. Topics get read; instructions get done. |
| Numbers, names and places, not adjectives | “Check it is secure” cannot be verified. “Pull the latch — it must not move” can. |
| No step longer than two lines | If a step needs a paragraph, it is hiding a second step or a piece of reference material. Split it or link it. |
| One version, one place | Two circulating versions of the same checklist are worse than one imperfect one. Retire old copies when a new version ships. |
| Date and source on the header | Every checklist should name the SOP it came from and the date it was generated, so anyone can check it is current in seconds. |
The conversion checklist (use this on your next SOP)
- Every action in the SOP is highlighted — including the ones hidden inside paragraphs
- Actions are in physical working order, prerequisites first
- Each step is one action, starting with a verb
- Each step ends with an outcome someone else could verify
- Every warning sits immediately before the step it protects
- Background and policy text stayed in the SOP, linked from the header
- The header names the source SOP and the generation date
- A newcomer test is scheduled — and fixes will go into the SOP, not only the checklist
Skip the manual conversion
OniLens does steps 1–5 for you: paste your SOP text — or load a .txt, .md or PDF file — in the free browser app and get an ordered, checkable list back — your documents never leave your device. Prefer not to do it yourself? Assisted Conversion turns up to 10 procedures into finished browser and PDF checklists for $29 one-time, and Team & Operations covers larger sets with branding and shared links for $49 one-time.
Free app: no account, no credit card, up to 50 saved checklists. Pro license $39/year removes the limits. 30-day refund policy on paid plans.
Frequently asked questions
What is the difference between an SOP and a checklist?
An SOP explains a procedure: background, responsibilities, rules and full detail. A checklist is the execution layer of the same procedure: only the actions, in order, each short enough to do and check off. Most teams need both — the SOP stays the reference, the checklist is what the person doing the work follows.
How long should a checklist step be?
One action, in one short sentence, starting with a verb. If a step joins two different actions with “and”, or a second person cannot tell when it is finished, split it. A newcomer should be able to do the step and honestly tick it off without asking anyone.
Should I copy the whole SOP into the checklist?
No. Copy only what the person must do, in the order they must do it. Background, history and policy language belong in the SOP, linked from the checklist header — not between the steps.
Where do warnings and safety notes go?
Immediately before the step they protect, never in a block at the top or the end. Keep each warning to one line, say what happens if it is ignored, and keep the step itself separate so it can still be checked off.
Can software convert an SOP into a checklist automatically?
Software can extract the actions, order them and remove duplicates — that is what the free OniLens app does in your browser. A human still reviews the result before anyone runs it, especially for safety-critical work.
How often should a checklist be updated?
Every time the underlying SOP changes, and every time someone running the checklist finds a step that is unclear, out of order or impossible as written. Fix the source, regenerate the checklist, and retire the old version.