Your first SOPs should start with one decision

Imagine a founder handing their first support hire a folder labelled “How we work.” It contains fifty pages of background, a few screenshots and a spreadsheet of links. The founder feels organised. The new hire still does not know what to do when a customer asks for an exception to the refund policy.
That is the test for a useful SOP: can someone use it to make the next decision with confidence?
For an early-stage company, writing procedures from scratch can feel like a project to postpone until the team is larger. In reality, the first few procedures are a way to decide how work should happen while the people closest to the work can still sit together and improve it. You do not need to document every possibility. Start with the moments that are frequent, consequential or hard to recover from.
Begin with questions, not templates
Ask the team which situations create repeated interruptions. What do new starters ask again and again? Which customer request gets a different answer depending on who is online? Where does a case wait because nobody knows who owns the next step? What mistake would be expensive, frustrating or difficult to reverse?
Choose three to five workflows from those answers. For a support team, they might be identity verification, billing disputes, a technical escalation, a service outage and a sensitive complaint. For an agency, they might be client onboarding, approval of deliverables and handover when an account manager is away. The right first set depends on your work, not an internet list of “essential SOPs.”
Rank each candidate by frequency, customer impact and ambiguity. A rare but high-risk situation can belong ahead of a common, straightforward task. A useful first SOP removes a real bottleneck; it is not simply easy to write.
Watch the work before describing it
I became interested in procedures through remote support operations, where information was scattered across tools and people. The apparent process and the actual process were not always the same. A document could say “send to the specialist team,” while an experienced colleague knew that a particular issue required a different route. Writing from a manager's memory alone would have captured the neat version, not the route a new starter needed.
Sit with someone doing the task. Ask them to narrate the decisions: What triggers the process? What information must be checked? What changes the route? Where do they record the outcome? When do they stop and escalate? Ask what they do when the ideal sequence fails. If possible, observe a second person too; differences between two capable colleagues often reveal an unclear rule.
Toyota's approach to continuous improvement stresses going to the source to understand the facts. The principle applies to a small service business as well: look at a real case and the system where it is handled before declaring how the process works.
Write for someone in the middle of a task
A useful SOP can fit on one page if the task is simple. Give it a clear title in the language people search for. At the top, state when to use it, who owns the decision and what a successful outcome looks like. Then set out the steps in the order they happen.
Include the decision points that change the route. “If the customer has already been charged, check the payment record before offering a refund” is more useful than “resolve billing enquiries promptly.” Link to the actual form, system view or policy. Explain what evidence to capture and where. Specify the escalation trigger and the person or queue that receives it. If staff have discretion, state its boundary and when approval is required.
Atul Gawande's The Checklist Manifesto offers a helpful reminder: the value of a checklist lies in making critical actions easier to carry out amid complexity. A long document that hides the decision is less helpful than a short guide with the right branch in the right place.
Test the draft against a real case
Hand the draft to a teammate who did not write it. Ask them to use it on a realistic example without the author talking them through it. Watch for the first point where they need to guess, search elsewhere or ask what a phrase means. Revise the page, then test again.
This is also where you can catch a dangerous kind of vagueness: a step that sounds sensible but has no clear action. “Use your judgement” may be appropriate for a genuinely unusual case, but the SOP should still say what information to consider, what authority the colleague has and where to get help.
Research and practice around frontline involvement reinforce this approach. Zeynep Ton's work on good jobs argues for standardising routine work with employee input while empowering people to solve customer problems. Clear standards and human judgement can support each other when the boundary between them is explicit.
Give the page a life after launch
Before calling the SOP finished, assign an owner and a review date. Decide how a teammate can flag a broken link or missing branch in less than a minute. Add the page to onboarding and point to it in the place where work happens. When a tool, policy or role changes, the procedure should be on the change checklist.
The first set of SOPs is not a monument to how the founder once worked. It is a shared starting point that can improve as the company learns. If one new colleague can find the right answer without waiting for the founder to return from a meeting, the first page has already earned its place.
Further reading and source notes
Toyota, human resources development and the Toyota Way — going to the source, improvement and respect for people.
Atul Gawande, The Checklist Manifesto (publisher overview) — checklists and complex work.
Zeynep Ton, “The Case for Good Jobs,” Harvard Business Review — standardise routine work with employee input and empower customer problem solving.
Atlassian, Roles and Responsibilities Play — a practical method for clarifying ownership with the team.




Comments