The Producer School01Two days a week went back into building product. Revenue rose 42%.
The Producer School02The whole engagement, in one slide.
Six months, 1 March to 31 August 2026, measured against the six months before. Every figure is pulled from the source system rather than a platform dashboard: Shopify Admin API, Search Console, Google Ads, Meta Graph and Gmail.
The Producer School03Two founders, and one of them was the marketing department.
The expensive part was never the hours. It was that nothing shipped on a schedule. None of this is unusual, which is the point: it is the shape of most product businesses under twenty people.
The Producer School04What a week of content actually cost.
"I lost a few hours just getting to a proper topic. And that was the start, before you even touch the copy or visuals."
| Job | Before | Cost |
|---|---|---|
| Finding the topic | Scanning plugin news and music developments for something worth saying | Most of a day |
| Choosing the angle | Deciding how to treat it, from nothing | Hours |
| Making it | Carousels and blog posts across every channel | The rest |
| Reels | Never consistent enough to count as a channel | Skipped |
Roughly two full days a week, of one founder, in a company of two. One week of content, and reels were dropped because there was nothing left.
The Producer School05What we built.
All five run on always-on agents at a fixed monthly cost, not per seat and not per contact. Two things stayed human and were never candidates for automation: knowing what sounds like the brand, and saying yes.
The Producer School06The workforce behind the five parts.
Not one model writing posts. A staffed team, each specialised in one discipline. Content Director, Copywriter, Social Media Manager, Video Editor and Newsletter Editor are the seats that ran this engagement. The harness assigns the work, runs each agent on a schedule and holds state between them, which is what makes it an operation rather than a prompt.
The roster. Every agent is a named role with its own model, schedule, budget and run history.
The Producer School07One post, end to end.
The same loop runs for a plugin release, a competitor's post that took off, or a customer question.
The Producer School08What the operator actually sees.
This is the operator’s screen, not a diagram of it. What arrives, what is in flight, and what goes out when. Three views run the week.
The Producer School09Rendered from the brand, not designed by hand.
The identity was extracted from the live storefront into tokens, type, components and layout rules. Generation reads it and the renderer applies it, so the margins and the type scale are the same every time. No designer touched a slide in either of these.
"It wasn't right away the quality I'm used to doing by hand. Once we implemented the design system into the system itself, it went really well."
The Producer School, on the first month
How Prospa produces 2026's biggest sound. Seven slides, live on Instagram.
The rise of Brunello. Nine slides, ending in the comment prompt.
The Producer School10YouTube, from topic to thumbnail.
Four published thumbnails. Same layout, same type, different artist and genre each time.
The only step that still needs a person is standing in front of the camera. Research, idea, script and thumbnail all arrive before he sits down.
The Producer School11The hours did not disappear. They moved.
"I save all these other hours to now do plugin development, which has skyrocketed our business."
The Producer School
The Producer School12What two days a week turned into.
| Six months | Before | After | Change |
|---|---|---|---|
| Revenue | €279,203 | €397,801 | +42% |
| Paid orders | 5,498 | 8,877 | +61% |
| Average order value | €50.78 | €44.81 | −12% |
Orders grew faster than revenue, so the growth came from more buyers rather than bigger baskets. Shopify Admin API, paid orders, refunds excluded. 1 Mar to 31 Aug 2026 against 1 Sep 2025 to 28 Feb 2026.
The Producer School13We also spent more. Here is what the extra spend returned.
| Six months | Before | After |
|---|---|---|
| Ad spend, Meta and Google | €68,712 | €112,055 |
| Blended return on spend | 4.06 | 3.55 |
| Return on the additional spend | €2.74 per €1 |
Efficiency fell as spend scaled, which is what scaling does. The number that matters is the marginal one. Blended return uses Shopify revenue, so it is unaffected by the July consent banner. Platform-reported figures are not used here.
The Producer School14Found by people who did not know the name.
| Before | After | Change | |
|---|---|---|---|
| Non-brand clicks | 5,819 | 8,160 | +40% |
| Brand clicks | 4,414 | 4,505 | +2% |
| Total clicks | 26,797 | 33,929 | +27% |
| Then | Now | |
|---|---|---|
| Articles published | 18 | 36 |
| Articles earning at least one click | 51 | 144 |
| Clicks landing on articles | 3,423 | 10,391 |
Brand search is flat at +2%, so the +40% non-brand is new audience, not existing customers searching more. Two different baselines, kept apart on purpose. Search Console anonymises a share of long-tail queries, so brand and non-brand do not sum to the total. Seventeen of the 36 articles went out in a single batch on 28 May, so this is not a cadence claim.
The Producer School15Twice the inbound. Faster replies.
| Customer support | Before | After |
|---|---|---|
| Inbound threads a month | 96 | 202 |
| Median hours to first reply | 13.1 | 5.2 |
| Answered within 24 hours | 64% | 91% |
Volume more than doubled and the response got faster, which is the hard direction. July and August 2026. Latency measured on a sample of answered threads, first customer message to first reply, by Gmail timestamps.
The Producer School16"It systemised something I was never able to keep up with as a two-man team. What I actually notice is the hours it gives back. Those go into plugin development now, and that's what's grown the business."
The Producer School17What we can and cannot claim.
One thing that is not in scope and is worth watching: visits from AI assistants went 86 in June, 118 in July, 350 in August. The channel did not exist before June. That is a separate conversation, not something quietly added to a bill.
The Producer School18What this looks like for your brand.
What does not carry over: their audience, their catalogue and their two-founder shape. This was a content operation inside a wider engagement, and the numbers came from all of it. Ad spend, tooling and platform fees are separate. Content will lose to something with a deadline again next quarter, for the same reason it lost last quarter. The only question worth answering first is how many days a week it is taking you now.