Sartoro

What I was trying to figure out in 2024: a custom-made menswear brand had real momentum — a genuine product, a founder who could sell it, a store already doing seven figures — but nobody owned the whole thing. Photography, retouching, product setup, SEO, content, tracking: all real work, split across an agency and whoever had time that week. I didn't join to run a department. I ended up running the store's entire product-and-growth line, solo, reporting straight to the founder — and the last two years have mostly been about deciding, one bottleneck at a time, what should stay manual, what should get systemized, and what I should stop doing by hand entirely.

Taking the SEO in-house

The store was paying an agency for SEO. I'd been doing full-stack marketing before this, so the first real call was: bring it in-house and see if I could do it better and cheaper. Organic search traffic roughly doubled, and it has kept compounding since: the Search Console milestone log shows the store at 2,200 Google clicks a month in September 2025 and past 4,000 by 25 September 2026. The agency line item went away — about US$3,000 a month back into the business. That freed budget is a big part of why the AI-pipeline work below was affordable to build instead of buy.

The SEO work also pointed forward, not just up: I started treating GEO — how the store shows up inside ChatGPT and Perplexity answers, not just Google's ten blue links — as a now problem, not a someday one. Sartoro products surfacing inside AI answers earlier than most DTC competitors wasn't luck; it came from writing product content that's structured for a model to quote, which is its own discipline. I wrote the full playbook here: generative engine optimization, end to end.

Growing the line that mattered most

The custom-made luxury collection was a smaller line when I started paying real attention to it. It's now the store's lead line — roughly 60% of revenue. That's a merchandising and content story as much as a product one: the luxury pieces needed pages that matched what they were — real fabric detail, real configuration depth, photography that didn't undersell it — while the rest of the catalog kept moving.

It's also where I hit the wall that shaped everything after. Pushing that collection live by hand — the photography, the metafields, the SEO, the configurator wiring — took the better part of six months. Not because any single step was hard. Because a product page on this store is dozens of interdependent fields, and doing that correctly, hundreds of times, without a system, is a tax you pay in months.

Building the systems instead of just doing the work

Once I could see the actual shape of the bottleneck, the honest move wasn't "work faster." It was: stop treating the recurring stuff as one-off work, and build the thing that does it reliably instead.

That became two separate systems, each its own case study:

• Jarvis — a marketing operating system that pulls the fragmented numbers (Meta, Google, TikTok, Search Console, Shopify) into one place I can actually read, instead of reconstructing the picture from a dozen tabs every week.
• Product Helper — a guard-railed AI agent that fills product pages on the live catalog. The same collection that took six months by hand, redone with the agent, took about a week. That comparison is the clearest evidence I have that the systems approach was the right call: an agent you can trust on a live store.

Neither tool touches price, inventory, or publishes anything without me approving the diff first. The point was never to remove the judgment — it was to stop spending it on work a system could do correctly every time.

Where it stands

Full-time, 2024 to now. I still run the line day to day — photography, retouching, new products, the SEO and content that keeps compounding. The difference between year one and now isn't that there's less work. It's that the recurring 90% runs through systems I built and trust, which means the time I have goes to the part that actually needs a person: deciding what's worth building next. That's the same bet behind marketing engineering as an idea — the leverage is in the system you keep, not the task you finish.

Related writing: The bottleneck moved — on why the hard part of marketing shifted from doing the work to building the system that does it.