Skip to content
Ashish Kumar AgnihotriOperating Partner

Republic World · Digital news media

Quadrupling newsroom throughput to 400 stories a day

A high-velocity digital newsroom needed to multiply output without multiplying errors — or headcount at the same rate. The constraint was not effort; it was the operating model.

A digital newsroom lives or dies on velocity. This one was busy, the team worked hard, and yet output sat well below what the operation was capable of. Leadership read that gap as a resourcing problem and wanted to solve it by adding people. I read it as an operating-model problem, and the distinction mattered: pouring headcount into a process nobody could see would have multiplied cost and confusion without multiplying stories.

My brief was to lift throughput without sacrificing quality and without scaling cost in lockstep with output. The constraint was never the effort of the journalists. It was the way work moved between them — where it waited, who owned it, and when it was checked. Those are systems questions, and they are the ones I am brought in to answer.

The result was a fourfold increase in throughput, to roughly 400 stories a day, with quality holding because the gains came from how the work was organised rather than from pressure on the people doing it. On the strength of the proven model, the account expanded by two additional news desks.

01The challenge

The newsroom was producing at a fraction of its potential and could not see why. Output was capped not by the size of the team but by how work moved through it: unclear ownership of each stage, queues forming invisibly between desks, and quality checks that happened too late to do anything but reject finished work. Leadership wanted scale, but adding people to an unmeasured process would only have added cost and confusion.

02The intervention

What I actually did.

  1. 01

    Mapped the true production pipeline from assignment to publish, exposing where stories waited rather than where they were worked.

  2. 02

    Redesigned the desk structure and handoffs so ownership was unambiguous at every stage, and built the quality check into the flow rather than bolting it on at the end.

  3. 03

    Installed a daily operating cadence and a small set of throughput and quality metrics the editorial leadership could steer by in real time.

  4. 04

    Scaled the model deliberately, adding two new news desks only once the operating system could absorb them without losing control of quality.

03The outcome

Content throughput rose to roughly 400 stories a day — a fourfold increase — while quality held, because the gains came from the operating model rather than from pressure on the team. The account expanded by two additional news desks, with the structure and cadence to run them already proven.

4×

increase in content throughput

400/day

stories produced at the new run-rate

+2

news desks the account expanded by

In depth

The operating reasoning behind the result.

The symptom was low output; the problem was invisible flow

The newsroom could tell me how many stories it published, but not where the rest of its capacity went. That is the signature of a flow problem rather than an effort problem. Stories were not slow because anyone was idle; they were slow because they sat in queues that nobody could see — waiting between a writer and an editor, between editing and a final check, between a check and publish. Each handoff was a place where a story stopped and waited for attention. Add those waits across a day and they swamp the time actually spent on the work. Before changing anything, I needed to make those invisible queues visible, because you cannot fix a delay you cannot point to.

Mapping the pipeline from assignment to publish

I traced the true production pipeline end to end, from the moment a story was assigned to the moment it went live. The point of the map was not the steps everyone already knew about; it was the gaps between them. I followed where stories accumulated rather than where they were worked, because the bottleneck in any production system is almost never the busy station — it is the queue in front of it. That map turned a vague sense that the newsroom was slow into a precise picture of which handoffs were absorbing the day. Once leadership could see the waiting, the redesign stopped being a matter of opinion and became a matter of removing specific, identified delays.

Redesigning ownership and building the check into the flow

Two changes did most of the work. First, I redesigned the desk structure and the handoffs so that ownership of each story was unambiguous at every stage — no story sat in the gap between two people who each assumed the other had it. Second, I moved the quality check into the flow rather than leaving it bolted on at the end. A check that happens only on finished work can do nothing but reject it, which wastes everything already spent. A check built into the flow catches the problem while it is still cheap to fix. Together these removed both the waiting and the late rework that had been quietly capping output.

A daily cadence and metrics leadership could steer by

A redesign that is not measured drifts back to the old shape within weeks. So I installed a daily operating cadence and a small set of throughput and quality metrics that editorial leadership could read in real time. The cadence gave the day a rhythm — a short, regular point at which the state of the pipeline was visible and decisions about it were made. The metrics were deliberately few, because a newsroom does not need a dashboard it has to study; it needs two or three numbers it trusts and acts on. This is what let leadership steer the operation actively rather than discover at the end of each day how it had gone.

Scaling only once the model could absorb it

The expansion to two further news desks was deliberate in its timing. Adding desks to an operation that has not yet got its flow under control simply reproduces the chaos at a larger scale. I held the expansion until the operating model — clear ownership, in-line quality, a working cadence — was proven on the existing desks and could absorb more without losing control of quality. By then the structure for running new desks already existed; the new desks slotted into a system rather than improvising their own. Scaling after the model is proven, not before, is the difference between growth that compounds and growth that multiplies the problems.

The transferable principle

The lesson generalises well beyond a newsroom. When a high-volume operation underperforms, the instinct is to add capacity, and the instinct is usually wrong. Output is most often capped by how work flows — by the queues, handoffs and late checks between the people, not by the people themselves. Fix the flow first and existing capacity delivers far more; only then does adding people compound rather than complicate. This is the heart of how I approach any throughput problem: make the waiting visible, fix ownership, build quality into the flow, and prove the model before you scale it.

Questions

Common questions.

The work moved in deliberate phases rather than as a single push. The first stage is diagnosis — mapping the real pipeline and making the invisible queues visible — which is observation before change. The redesign of ownership and the in-line quality check follow, and then the daily cadence beds in. Only once the model held on the existing desks did the operation scale to the new ones. The sequencing matters more than the speed: an operator who reorganises before understanding the flow is guessing, and a newsroom that scales before its model is proven simply multiplies its problems.

By the run-rate of stories the newsroom could produce while holding quality, compared against the baseline before the change. The headline figure was a fourfold rise, to roughly 400 stories a day. Crucially, throughput was paired with a quality measure rather than reported alone — a volume gain that quietly sacrifices quality is not a gain, it is a deferred cost. Tracking both on the same daily cadence is what let leadership confirm the improvement was real and sustained, not a temporary surge that the operation would pay for later in corrections and credibility.

No, and that was the point of the design. The throughput gain came from removing waiting and fixing ownership, not from pushing the team to work faster on the same flawed process. Speed bought by pressure erodes quality; speed bought by better flow protects it. Because the quality check was built into the pipeline rather than left to the end, problems were caught while they were still cheap to fix, and finished work needed less rework. Quality held while volume rose precisely because both were treated as properties of the operating model rather than of individual effort.

Because the constraint was not the size of the team. Adding people to a process whose flow nobody could see would have added cost and coordination overhead without removing the queues that were actually capping output. In a flow-constrained operation, more hands often make the bottleneck worse, not better, because the waiting is between stages rather than within them. The disciplined sequence is to fix the flow first so existing capacity delivers what it is capable of, and only then add people — at which point each new hire compounds the gain instead of crowding an unmanaged process.

Yes, because the mechanism is not specific to media. Any operation that moves work through stages — claims, onboarding, billing, client delivery, production — accumulates invisible queues, ambiguous ownership and late quality checks as it grows. The approach is the same in each case: map how work really flows, expose where it waits, fix ownership at every handoff, build the check into the flow, and prove the model before scaling it. A newsroom is simply a high-velocity, highly visible example. The underlying problem, output capped by flow rather than effort, is one of the most common in operations-heavy businesses.

Through the daily cadence and the small set of metrics, which together gave the improvement somewhere to live. Process gains decay when nothing watches them; the redesigned flow would have drifted back toward its old shape within weeks if leadership had stopped looking. By making throughput and quality visible every day and giving the day a regular rhythm in which they were reviewed and acted on, the new way of working became the normal way rather than a temporary project. The expansion to new desks reinforced this, because the proven model — not a set of habits — was what the new desks were built on.

Next case

$20M+

A three-step Makegoods QA framework that protected $20M+ in billings

There is a number like this in your operation.

A diagnostic is the fastest way to find it and put a figure on what fixing it is worth.

Book a consultation