Editorial SEO Systems
I have run SEO inside a newsroom publishing hundreds of pieces a week. At that pace nobody is reviewing individual articles, so the only work worth doing is the work that makes the whole desk better by default. That is a process problem, not a writing problem.
Publishing operations I have worked inside
What gets automated, and what does not
This is the pipeline I build, stage by stage. The useful part is not the automation, it is being specific about where it stops. One of these stages has nothing in the automated column and that is deliberate.
Deciding what to write
AutomatedRanking movement, competitor changes and SERP data pulled overnight into a short report that is waiting when the desk starts.
Stays humanWhich of those are worth covering. The desk knows things the data does not, including what is about to happen rather than what already has.
The brief
AutomatedGenerated from SERP and query data. What the piece has to cover, the questions it needs to answer, what it should link to and what should link to it.
Stays humanThe angle. A brief gives you the shape of a piece, never the reason anyone would want to read it.
Writing it
AutomatedNothing.
Stays humanAll of it. Content that nobody thought about reads like content nobody thought about, and both readers and search results have got better at noticing.
Before it publishes
AutomatedOptimisation checks, internal linking suggestions, schema, and a quality pass on whether the page actually answers the question it was written to answer.
Stays humanThe decision. Every check is advisory. The moment a check can block publishing, the desk starts working around it.
After it publishes
AutomatedPerformance tracked against the cluster rather than the page, with decay signals and refresh candidates surfacing on their own.
Stays humanWhat to do about it. Refresh, rewrite, merge into something stronger, or accept it did not work and leave it alone.
What I work on
Editorial workflow design
Mapping how the desk actually publishes, then putting SEO into that flow rather than beside it. If SEO is a review step it loses to the deadline every time, which is a workflow problem rather than a discipline problem.
Briefs at volume
The brief is almost always the bottleneck. Writers wait on it, and a rushed brief produces a piece that needs rewriting. Automating the research part of a brief buys back the time where the thinking happens.
Refresh and decay
Most content operations are set up to publish and have nothing that decides when to go back. Decay signals surface candidates automatically, and then someone has to judge whether it is worth it.
Topical clusters and coverage
Planning coverage as a connected set rather than a list of articles, so pieces support each other instead of competing. Depends on the internal linking underneath it actually working.
Running the team
I have managed five writers and analysts directly, including workload planning, review and the process design around both. Knowing what is realistic to ask of a desk is most of what makes a system get used.
Reporting the desk will open
Dashboards built around the question an editor already has, not the question an SEO finds interesting. Part of my AI assisted workflows.
How I approach a content operation
Five things, and the first one is not optional.
Watch how the desk already works
Before suggesting anything, I want to see a normal publishing day. Almost every editorial SEO system fails because it was designed for a workflow the team does not have.
Fix the bottleneck, not the symptom
Output problems are usually not writing problems. At MoveUp Media the bottleneck was the gap between finding something and shipping it, and closing that took 70% off the cycle without anyone writing faster.
Automate the research, never the judgement
Pulling data, checking structure and surfacing decay are jobs a machine does better and more consistently. Deciding what is worth covering is not, and a system that pretends otherwise produces a lot of content nobody wanted.
Make the default the right one
You cannot review hundreds of pieces a week. What you can do is put the structure into the template and the brief so the easy path is also the correct one.
Close the loop back to planning
Performance has to reach the person deciding what to publish next, or the operation never learns. Where that includes Discover, the feedback is noisier and needs reading over weeks rather than days.
"At this volume a process change beats any single article. Editors will use a system when it answers a question they already had."
Editorial work I can point at
Three examples with the company named. The full write ups are on the homepage.
Editorial SEO inside a 100M pageview newsroom
2,000+ articles optimised, dashboards the desk actually opened, and a new section built from a keyword map.
Content operations automation
Brief generation, optimisation checks and a helpful content quality pass, built with Claude and MCP.
Cutting analysis to implementation by 70%
Rebuilt how a finding reaches the people who can act on it, which was the actual bottleneck.
Frequently Asked Questions
Your team is publishing plenty and it is not working
Tell me how the desk currently decides what to write, and I will tell you where I think the bottleneck is, whether or not we end up working together.