A Two-Person Studio Tries NoOps: Notes From One Quarter

Meet Cardinal & Co. — a composite studio, invented for this piece, built from patterns we keep hearing in reader emails. Two people. A roster of guest-posting clients. A stack made entirely of rented machinery: documents, spreadsheets, a publishing CMS, an email platform, a scheduling tool, a billing app. Every logo on their invoice says “managed.” Neither of them has ever configured a server, and neither wants to start. In January, after one ugly platform outage ate a client launch, they decided to pursue NoOps properly — and to write down what happened. These are those notes, compressed and lightly tidied, because the numbers are more useful than the drama.

The NoOps Problem That Started It

The trigger incident was ordinary, which is what made it scary. Their email platform had a regional outage on the morning a client’s campaign went live. Two hours of silence, a scrambled resend, an apologetic email to a client who was understanding in the way that precedes a rate renegotiation. Afterwards, the founder — call her Rhea — did the honest thing and listed every operational task the studio actually performs: checking deliverability, rotating API keys, reconciling invoices, verifying that scheduled posts actually fired, keeping the client portal tidy, backing up the editorial calendar.

The list came to about thirty-nine hours a month. Her first reaction was relief. Under ten hours a week, split across two people — that is nothing, right?

Her second reaction, one week of research later, was the opposite. That number measures the work. It does not measure the coverage, and coverage is what nearly bankrupted their evenings.

The Coverage Maths Nobody Warns You About

Here is the arithmetic Rhea found, and it is the single most useful thing in these notes. Work hours and coverage hours are different animals. The thirty-nine hours ask: how much time does the task take? The coverage question asks: at how many hours of the week must somebody be ready to act if the thing breaks at that moment? A client campaign does not politely restrict its failures to Tuesday afternoons.

Reliability engineers have formalised this. A service needing thirty-nine hours of monthly attention can demand something like one hundred ninety-five person-hours of monthly coverage, because readiness around the clock costs far more than the chores themselves. And a genuine around-the-clock rotation, run under a sane rule where nobody carries more than a quarter of the on-call burden, takes a roster of about eight based in one place — Google’s single-site minimum, and the reason serious teams respect it instead of improvising around it.

Cardinal & Co. has two people. The gap between two and eight is the entire story of this quarter.

Rhea’s note, week three: “We don’t have an ops problem. We have a coverage problem. Those are different budgets.”

Month One: Writing It All Down

Their first move was the least glamorous one available: a register. One spreadsheet, one row per recurring duty, four columns — what it is, roughly how long it takes, who currently does it, and what breaks if nobody does it for a week. Practical advice if you steal this: keep the descriptions dumb-simple. “Confirm each scheduled post actually published” beats “perform publication verification” every single time.

The register produced two surprises. First, eleven of their thirty-odd duties existed only because a tool was configured badly once and nobody went back to fix the configuration. Second, three duties — key rotation, portal access review, calendar backup — had no owner at all. They simply happened, or didn’t, depending on the month.

Month Two: Buying Coverage Deliberately

Month two was triage. Every duty got sorted into one of three buckets:

  • Buy: hand it to a provider who offers it as a feature — managed backups, platform-level uptime, built-in key rotation. The provider’s engineers become the on-call roster they cannot staff themselves.
  • Automate: script it once so it stops recurring — the publication checks became a five-line scheduled script with an alert.
  • Keep: accept it as studio work — anything involving editorial judgement or client relationships stays human, full stop.

The bucket decisions took two working days. Getting each one verified took most of the month, because “the platform does backups” and “we restored a deleted campaign folder from the platform’s backup” turned out to be different sentences entirely. They ran one restore test per tool. Two passed. One required a support ticket and a nervous Saturday.

Rhea also wrote what she called an exit note for every tool in the stack: where the data lives, how it exports, and what it would cost in lost afternoons to move somewhere else. None of the notes ran longer than half a page. All of them changed a decision at least once — the scheduling tool they adored turned out to hold their calendar in a format with no clean export, which downgraded it from “keep forever” to “keep until something better behaves.”

Month Three: Living With the Result

By the end of the quarter the studio had something quieter than the NoOps of conference talks. Their register shrank from thirty-two duties to nineteen. Every remaining row had a name against it. The pager — such as it is — now routes to providers during platform failures and to Rhea or her partner only for editorial fires, which are the fires they are actually good at putting out. Nothing about their operation disappeared. It moved: upward to vendors, sideways into small scripts, and whatever stayed became deliberate.

Their scarce skill, after all, was never server administration. It is pitching, writing, and placing client work — the same craft our guides on posting rules for Medium and publishing articles on DEV Community exist to support. Every hour reclaimed from operational duty goes back into that craft, which is the only place a two-person studio can afford to spend it. (Their internal wiki, for what it’s worth, is just a well-behaved Notion workspace run by the book.)

The NoOps Playbook Worth Copying

Strip the narrative and you get a compact playbook. It is not exciting, which is exactly why it works:

  1. List every recurring operational duty in one place, with an owner against each row.
  2. Estimate the work hours, then honestly estimate the coverage hours. Expect the second number to frighten you.
  3. Sort duties into buy, automate, or keep — and verify what you buy by testing a restore.
  4. Revisit the register quarterly, because tools drift and new duties arrive wearing friendly faces. And keep the exit notes current: a stale note is almost as expensive as no note, since it sends you digging through old screenshots at the worst moment.

For the wider context — where the term came from, the full coverage discussion, and the maturity ladder reliability teams use — the NoOps page in the Ops Atlas is the reference we would hand Rhea next.

The verdict from the quarter is worth keeping on a sticky note. NoOps, for a small studio, does not mean operations evaporate. It means operations resized until they fit a team of two — purchased where purchase is possible, automated where automation is boring, and kept where it makes the work worth doing. The moment somebody sells you more than that, ask them who is carrying the pager, and count how many of them there are.

Related Posts

Why Logging Every Guest Post the Same Day Saves Hours

A friendly explainer on filing, logging and clearing small bits of content work the day they finish, so a pile of drafts never builds up.

Read More

How Writers Learn From Published Posts: A State-of-Practice Look

More contributors now dissect one finished post instead of copying a favourite author. A level-headed look at what the habit reveals and where its limits lie.

Read More

How to Rewrite a Post in Plain Words, One Step at a Time

A seven-step method for turning a jargon-heavy draft into something strangers can follow, with a few simple swaps and a quick outsider read.

Read More

Content Marketing Within Limits: An FAQ on Creative Constraints

Answers to common questions about planning content when the blank page freezes you, using self-set limits such as one reader, one takeaway and one platform.

Read More

Content Audit Checklist: Fast, Cheap or Good, Which Do You Drop?

A seven-step checklist for auditing your content routine, spotting which of speed, budget or quality you are really sacrificing, and recording that choice.

Read More

Eight Content Patterns to Check Before You Rewrite Any Post

Before you rewrite a flop, count. These eight repeating signals in your headlines, timing and feedback show where the real weakness sits.

Read More

Leave a Reply

Your email address will not be published. Required fields are marked *