Industrialising content production for your clients

"Industrialising" sounds frightening, because it is associated with falling quality. Done badly, that is true. Done well, industrialising means the opposite: removing the work with no value (formatting, copying, repeating instructions) so human time concentrates on what matters — framing and reviewing. Here is the chain that holds, the roles it assumes, and the three places where it breaks.

The arithmetic that decides everything

An agency selling 4 articles a month to 6 clients has to produce 24 articles monthly. In classic copywriting, that is a full-time writer. Let us break down the real time per piece:

• Writing from scratch: 2 to 4 h per article. Unsustainable at scale.
• ChatGPT plus manual copying: 30 to 60 min (context to paste back, PrestaShop fields to fill one by one). Better, but it breaks on volume.
• Framed generation plus direct publishing: 5 min of brief, 10 min of review, one-click publishing. Around 15 min.

At 15 minutes a unit, 24 articles equals 6 hours a month. The content engagement becomes profitable again — which is the heart of how you bill it.

That arithmetic has a counter-intuitive consequence: the bottleneck moves to reviewing. While you produced 4 articles a month, reviewing was free — it was buried inside writing time. At 40, reviewing becomes the only real constraint in your chain. Everything that follows exists to protect it.

The five links in the chain

1. The brief template. One model per content type (buying guide, comparison, product description): target keyword, outline, products to link, tone. Reusable from one client to the next — this is the standardised link.
2. Generation tied to the catalogue. The AI reads the client's real products and writes with them; nothing generic. That is the condition for avoiding duplication.
3. Targeted review. You do not rewrite, you verify: accuracy, tone, internal links. The passage that needs work is corrected on the fly.
4. Scheduled publishing. You prepare the month in one session and spread publications week by week. The client sees a living blog; you see a batch handled.
5. Sample-based quality control. Across a portfolio, you review everything at the start, then sample once the chain is running smoothly.

What a production session actually does

The theory fits in five links; in practice an agency works in sessions, not as things come in. As-things-come-in is what kills content chains: every piece demands the context, the tone, the store access all over again.

A typical session, over half a day, for a portfolio of six stores:

• Thirty minutes of framing. You look at what measurement says: which pages sit on page 2 two positions off the top 10, which product pages have no meta, which topics are already covered. This is the moment you decide, and it is the only strategic moment of the session.
• An hour of briefs. One target keyword, one outline, products to link, per piece. The templates do most of the work; you only write what is specific to the client.
• Generation, in the background. It occupies nobody: content arrives while you do something else.
• One to two hours of reviewing. Piece by piece, with a short checklist (see below). Passages needing work are corrected on the spot.
• Twenty minutes of planning. You spread it out: two publications a week per store rather than eight on the 1st of the month. A regular publishing profile beats a spike — and on a young domain, a spike shows.

By the end of the session the month is ready for the whole portfolio, and nobody touches it again until the next session. It is the only rhythm that holds at six clients, and it still holds at twenty.

The review checklist, in five points

A review with no checklist lasts forever. A review with one lasts ten minutes. The five questions that catch 95% of problems:

1. Is it accurate? Do the features cited actually exist on the client's products? This is the one non-negotiable point: a factual error about a product commits your client, not the tool.
2. Does it answer the query? Does the content address the search intent it targets, or a neighbouring, easier one?
3. Is there anybody inside it? A concrete example, a judgement call, an owned preference — otherwise it is smooth filler.
4. Do the links go to the right pages? Do they send authority to the product and category pages that sell, or in a loop to other articles?
5. Do the tags hold? Title and meta description in the right formats, no truncation, keyword up front.

What is not on the checklist: hunting turns of phrase, rephrasing to taste, reworking the outline. If you are rewriting, the brief was wrong — fix the template, not the text.

Keeping quality while industrialising

Industrialising does not excuse you from writing: it moves the writing. The risk is the text that is "clean with nobody inside". The remedy is one step — humanising generated content: a concrete example, a verified figure, an angle only the client could give. Five minutes that make the difference between filler and content worth putting your name to.

On the question that always comes up in meetings — "was this written by an AI?" — the position we take is documented in our comparison of AI, copywriter or ChatGPT.

The three places where the chain breaks

They are always the same, and none of them is technical.

• Client approval. This is the first breaking point: your content is ready, the client does not answer, your schedule melts. Two remedies: a contractual approval window (after n days, deemed accepted) and a preview the end client opens with no account to create — one link, they read, they approve or request changes.
• Store access. One expired password per client, and the production session stops for three follow-up emails. That is exactly what publishing through a paired module removes: nobody shares credentials, nothing expires.
• Memory of the context. A client's tone, their flagship products, their forbidden ground — if those live in one person's head, the chain stops while they are on holiday. They have to live in the brief templates.

The next scale up

Once the chain runs smoothly on one client, it replicates. That is where steering several stores from a single account comes into its own: the same chain, applied to the whole portfolio, without changing tool.

Beyond ten or so stores, one last reflex becomes necessary: stop producing at the same rhythm everywhere. Not every store in a portfolio has the same potential. Concentrate volume where measurement shows movement, and keep the others on maintenance. An industrial chain that produces uniformly is a chain that wastes — and across a portfolio, waste is counted in thousands of euros a year.

In short

Industrialising content means removing the 45 minutes of handling per piece to keep only the 15 minutes of value. Quality does not drop — it refocuses on reviewing, which becomes your real constraint: protect it with templates, a short checklist and batched sessions. Want to see the chain running on your portfolio? Ask for a demo.