Listicle articles built from a spreadsheet

What the listicle generator does

You give it a topic, the keyword you want to rank for, and a list of tools you want to write about. It writes the full article. Each tool gets a real summary, pulled live from the company's homepage and pricing page. The finished article lands in a Google Doc, with an images folder ready next to it, and your tracking sheet updates so you know which articles are done.

The whole thing runs from one row in a Google Sheet. You hit "go" once and a 2,000-word listicle is in your editor's inbox about ten minutes later.

Why listicle articles eat a writer's week

If you've ever tried to publish a "best 10 tools for X" article, you know how the time gets eaten. You research each tool, skim its homepage for the one-line pitch, and hunt down the pricing page for the actual numbers.

Then you write the intro that explains why these tools matter, a section for each tool, the comparison and the conclusion, and you find or commission images.

In my experience one article is four to six hours of human work. That's fine if you publish one. It becomes a real problem if your SEO plan calls for ten of them a month.

Most teams handle this in one of three ways. They hire a writer to grind through the queue at a per-article rate. They use an AI listicle generator that produces filler at scale, which rarely holds a ranking. Or they fall behind on their content calendar.

This workflow is the fourth way. You keep editorial control. You set the topic. You set the tone with a styling guide. You review the output before it goes live. But the four to six hours of blank-page work becomes ten minutes of edit-pass work.

One sheet row per listicle

Each article starts as one row in a Google Sheet:

• The topic. Something like "best AI writing tools for startups" or "best customer review platforms for ecommerce".
• The keyword you want to rank for. Usually a phrase a real buyer would type into Google.
• A list of tool homepages, separated by commas.
• A list of tool pricing pages, separated by commas. Each pricing page should belong to one of the homepages.

That's it. Mark the row as "Planned" and the workflow picks it up.

A Google Doc, a Drive folder and an updated tracker

When the workflow finishes, you'll have:

• A Google Doc with the full article, structured as a listicle. The intro hooks on a real problem the reader has. Each tool gets its own section with a real-language summary, who it's actually for, and what it costs. The comparison section helps a reader who's between two tools decide. The conclusion doesn't read like AI wrote it.
• A Google Drive folder named after the article, ready for the images you'll add manually or via a separate workflow.
• Your tracking sheet updated to show which article is done and where the Doc lives.

The output isn't a draft you fact-check from scratch. It's a draft you edit the way you'd edit a strong intern's article. Two or three rounds of tweaks, a fact-check pass, a brand voice pass, and it's ready to ship.

Ten minutes to draft, an hour to edit

About ten minutes from kicking off the workflow to a finished Doc. Your time goes into the edit pass after that.

In practice, my customers spend somewhere between thirty minutes and an hour per article on the edit. That's fact-checks, your team's opinion injected into the comparison, an image swap, maybe a custom intro hook. So end to end: about an hour of human time per article, instead of four to six.

If you publish ten listicles a month, that's a swing from forty to sixty hours of work down to ten.

At that pace new URLs pile up quickly, which is where automated indexing earns its place: it flags which of them Google still hasn't picked up.

Teams with a steady listicle queue

It suits teams where listicles are a steady part of scaling content marketing rather than a one-off, and where these hold:

• You've already mapped which keywords you want to target. The workflow doesn't pick topics. You hand it the topic and the keyword.
• You have a styling guide for your blog. The workflow uses it. If you don't have one, we'll write one as part of setup.
• You publish listicles often enough that doing them by hand is a real cost. If you ship one a quarter, build it by hand. If you ship one a week, this pays back fast.
• You want the output to read like your team wrote it, not like AI threw up on a page. The prompts in this workflow are tuned for that.

The styling guide is the one input that can't be skipped, because every prompt reads from it.

When writing listicles by hand still wins

Three signs it isn't worth setting up yet:

• You want a magic button that picks topics for you and never needs human review. That doesn't exist. Every article still gets your editor's eye before it ships.
• You publish fewer than two listicles a month. The setup time isn't worth it at that volume.
• You're not willing to give it your existing styling guide and tone of voice. Without that, the output reads generic.

For articles that aren't lists, SEO content automation is the sibling workflow: it builds the structure from the SERP instead of from a list of tools.

n8n, GPT, Jina and Google Workspace

n8n runs the orchestration. It's the workflow engine: every step, every loop, every conditional lives there. You own the workflow file. You can change anything in it.

GPT does the writing. The workflow uses GPT for the tool summaries, the article sections, the intro, the conclusion. The prompts are the result of about thirty rounds of editing. I rewrote them, ran them against ten different topics, looked at the output, found what was off, rewrote them again. That's the part that makes the output not read like AI. The prompts are yours at handover.

Jina handles the scraping. When the workflow needs the latest copy from a tool's homepage and pricing page, it fetches them through Jina. That keeps the tool summaries fresh and based on what each company actually says about itself today, not what someone wrote about them eight months ago when the pricing was different.

Google Workspace is where everything ends up. The article goes into a Google Doc. The images folder goes into Google Drive. Your tracking sheet stays in Google Sheets. There's nothing new for your editor to learn: they open a Doc, they edit, they ship.

Prompts and templates, in plain text

When I hand this over to you, you get:

• The n8n workflow file. You can export it, import it, fork it, version it.
• All the prompts, in plain text, documented. You can adjust them as your tone evolves.
• The Google Sheet templates for the input row and the article structure.
• A short Loom walking through how it runs end-to-end, where the levers are, and how to add a new article to the queue.
• A runbook covering the common gotchas. What happens if a tool's homepage 404s. What happens if you fill the row but forget the pricing page. How to rerun a failed article without duplicating work.

If you stop working with me, the workflow keeps running. There's no proprietary middleware to pull out. No vendor lock-in.

Why the prompts matter more than the wiring

I've built this exact shape of system across multiple content teams. The hard part is not wiring the nodes. n8n is forgiving and well-documented. Anyone with two weeks and the n8n docs can wire up a basic listicle generator.

The hard part is the prompts. Generic prompts produce generic output, and generic output reads as AI from sentence one. The prompts in this workflow took me about thirty rounds of edits across real publish-ready articles to get right. They handle keyword variation. They strip the cliches that make AI text obvious. They keep the listicle structure consistent across sections. They avoid the "as an AI language model" tells.

The other hard part is the editorial structure. A listicle that ranks for "best X tools" needs a specific shape: a problem-led intro, a quick consensus on what matters in the category, a per-tool section with the same five facts each time, a comparison section, and a conclusion that doesn't sound like every other conclusion.

The workflow's structure encodes that. The wiring is the easy part to copy; the structure and the prompts are the part you can't easily Google.

Ten to thirty cents an article

The variable cost per article is small. GPT-4.1-mini at current pricing runs about $0.10 to $0.30 per article depending on the size of the listicle and how much scraped content gets summarized. Jina's free tier covers most teams' volume. n8n self-hosted is free. Google Workspace you already have.

You don't buy the generator as a project. It runs as part of execution capacity: AI agents and automations at $2,000 a month flat, which comes alongside a fractional Head of Content retainer. The first two weeks go into tuning the prompts to your tone, setting up the sheet, wiring the workflow, writing the runbook and training your editor on the loop.

Map your listicle backlog on a call

On a 30-minute call we'll map your listicle backlog, look at one of your existing articles to set the tone target, and work out whether this is the highest-payback automation for your team. If something else fits better, I'll say so.