Scaling content production without it going bland
Increasing volume has become easy. Increasing volume without every piece becoming interchangeable is the actual task — and it is not solved by writing faster, but by dividing the work differently.
In short
- The bottleneck in scaling is never the writing. It is the substance — and that comes from the business, not from a tool.
- Separate five stages: gather substance, fix the topic, draft, enrich, sign off. Only two of them belong to the human.
- Three warning signs of dilution: interchangeable openings, disappearing figures, falling time on page at unchanged traffic.
- Four good pieces a month beat twelve mediocre ones — for visibility too, not just for reputation.
The question usually gets framed as a capacity problem: how do we manage twelve pieces instead of four? That is the wrong question, because a tool can answer it — and the answer is then worthless.
The right question is: where do twelve things a month come from that nobody else could write? Without an answer to that, you are scaling phrasing, not content.
The real bottleneck
A written piece consists of two parts. One is craft: outline, phrasing, transitions, correct language. The other is substance: a figure from a real project, a mistake you made, a position you can defend.
The first part can largely be automated. The second cannot — it arises in the working day and has to be collected there systematically, or it is lost.
The five stages
Stage 1 – Gather substance
One fixed place where whatever comes up in the working day lands: customer questions, surprising figures, mistakes, decisions with their reasoning. Two minutes per entry, no polishing.
Who: everyone in customer contact
When: continuously
Stage 2 – Fix the topic
From the collection plus keyword research comes a topic with a main term, an audience and the question it answers. Before writing, not after.
Who: human
When: monthly, for the month ahead
Stage 3 – Draft
Outline and first version from the topic, the substance entries and a fixed style frame. Fully machine-produced.
Who: model
When: per piece
Stage 4 – Enrich
Your own figures, examples and positions go in. Everything verifiable gets verified. This is where the value of the piece is created.
Who: human with subject knowledge
When: per piece
Stage 5 – Sign off
Fixed checklist: answer at the top, every section understandable on its own, images and alt text present, markup agrees with the content.
Who: human, working the list
When: per piece
Worth knowing
Stage 1 is the only one that cannot be caught up on. Look for your own examples on the day of writing and you will not find any — not because they do not exist, but because they were not captured at the moment they occurred.
A shared document that everyone in customer contact adds two minutes to per week produces, in a quarter, material for more pieces than most small companies manage in a year. That stage costs practically nothing and is still the one most often missing.
The division of work
| Task | Model | Human |
|---|---|---|
| Topic suggestions, clustering | good | selection |
| Drafting an outline | good | check the order |
| Writing the draft | good | – |
| Own figures and cases | not possible | alone |
| Fact-checking | not reliable | alone |
| Cutting and sharpening | good on instruction | decides |
| Translation into more languages | good | spot check |
| Sign-off and accountability | – | alone |
Three warning signs of dilution
1. The openings become interchangeable. "In today's fast-moving business world" is not a style problem but a substance problem: someone with nothing concrete to say begins in generalities.
2. The figures disappear. Compare your last five pieces with the five before them: how many verifiable original details do they contain? If that number falls, the value falls — regardless of word count.
3. Time on page falls at unchanged traffic. People still arrive but stay shorter. That is the first measurable sign, and it comes roughly a quarter before visibility declines.
The most reliable way to increase volume is not more topics — it is reuse. One thorough piece contains material for several smaller formats: a checklist from the section headings, a short piece from the strongest section, a questions part, one or two graphics.
The difference from simply splitting it up matters: every derived format has to stand on its own. A fragment that makes no sense without the original piece helps nobody. One piece, four formats — but each one so that you do not have to chase the other three.
The special case of multiple languages
More languages are the cheapest way to scale, because the substance already exists. Four things have to be right:
- The address gets translated too. A German title under an English address is half a move.
- hreflang for all languages, reciprocally. Every version points to all the others, including itself and a default version.
- Examples get adapted, not just translated. Amounts in the local currency, the legal basis of the relevant country, locally familiar phrasing.
- A spot check by someone who speaks the language. Not every version, but regularly one — especially for technical terms.
Write a draft for a technical piece. I am giving you the substance – do not invent anything on top of it. Topic: [title] Main search term: [term] Audience: [who exactly] Question the piece answers: [question] Substance from our business (use only this): - [figure, case, mistake, decision] - [...] Style frame: - Answer first, reasoning after – in every section too. - Factual, no superlatives, no exclamation marks, no "in today's fast-moving business world". - Every section must be understandable without the rest of the text. - 1,200 to 1,500 words, H2 headings as a question or a clear statement. Rules: 1. Do not use any figure, example or study that is not listed in the substance above. 2. Where you need a detail I have not supplied, write [detail missing: which] instead of inventing something. 3. At the end: five common questions with answers, each answer understandable without the rest of the text.
In closing
Scaling does not mean writing faster. It means collecting substance systematically and handing off the rest. Set up stage 1 and you have moved the bottleneck. Skip it and you produce more text with the same amount of content — and that is dilution, however well phrased.
The rule of thumb stays uncomfortable: four pieces with your own substance beat twelve without. Not only for reputation, but measurably — texts that summarise what exists elsewhere are neither found nor quoted.
Common questions
How do you produce content regularly without a big team?
Through five separate stages: gather substance continuously during the working day, fix topics monthly, produce the draft by machine, enrich it with your own figures and examples, and sign off against a fixed checklist. Only enrichment and sign-off genuinely need subject knowledge — together roughly 90 minutes per piece.
How many pieces a month make sense?
Four thorough ones beat twelve shallow ones — for visibility too, not just reputation. The volume should follow the substance available: planning twelve pieces with material for four spreads the same substance thinner and loses at both ends.
How do you tell quality is slipping?
By three signs that appear before the numbers drop: openings become general and interchangeable, the number of verifiable original details per piece falls, and time on page declines at unchanged traffic. The third is the first measurable one and comes roughly a quarter before visibility declines.
What can a language model take over in content production?
Topic suggestions and clustering, outlines, drafts, cutting on instruction, and translation. What it cannot take over is your own figures and cases, fact-checking, and sign-off — and those three determine whether the piece contains anything that does not already exist elsewhere.
Is translating into more languages worth it?
Yes, it is the cheapest way to scale because the substance already exists. Four things have to be right: translated addresses, reciprocal hreflang for all versions, examples and amounts adapted to the country, and a regular spot check by someone who speaks the language.
Marketing that sets itself up
The Studio Engine beta is live. Claim your spot and help shape it from the start.
Join the beta →