RogueLogic
01How we work

Measure. Model. Build. Compound.

The same four-stage programme runs on every account, whatever the discipline mix. Each stage is delegatable, which is how a small senior team ships at pace, and the whole method is published, including the part where it fails.

The short answer

Rogue Logic runs one four-stage method on every account: Measure the baseline, Model where demand is and what it is worth, Build against that model, and Compound the result month on month. The model is published in full and stays yours, and any channel that cannot be attributed to revenue within three months is stopped.

02The method

Four stages, one programme.

Numbered because the sequence is real. You always know which stage your account is in and what happens next.

01

Measure

Crawl, log files, entity graph, demand model, answer-engine footprint. Two weeks, no assumptions.

02

Model

Where demand exists, what it is worth, and what has to be true for you to own it. Published, not summarised.

03

Build

Senior operators execute. Technical fixes, content, PR, paid. Shipped weekly, not queued behind a sprint.

04

Compound

The model is re-run against reality every month. What works takes budget from what does not.

03Each stage, in full

What actually happens in each one.

01

Measure

Two weeks establishing the baseline, before anyone touches a page. We would rather find the problem than guess at it.

What happens
Technical crawl, render check and server-log review
Entity graph and content baseline against the top-five results per term
Demand model and answer-engine footprint across the assistants that cite sources
You getA findings pack with the numbers, not a slide of adjectives.
02

Model

The demand, mapped to revenue rather than vanity terms, with the assumptions written down so you can argue with them.

What happens
Priorities mapped to commercial value, not search volume
What has to be true for you to own each cluster, stated plainly
The maths published in full, and yours to keep if you leave
You getA plan you own, that survives the person who wrote it moving on.
03

Build

Senior operators execute against the model. Extractable answers, clean schema, entity consistency and third-party corroboration, shipped weekly.

What happens
Technical fixes, content and templates built to be cited
Digital PR for the corroboration the assistants already trust
Paid search and paid AI placement where the model says it pays
You getWork shipped every week, not queued behind a quarterly sprint.
04

Compound

The model is re-run against reality every month. Budget moves from what is not working to what is, and the reporting cannot be reshaped to flatter the result.

What happens
Clean GA4 and Search Console from day one, fixed measures throughout
Monthly re-scoring against the published model
Losing channels flagged and stopped, in writing, on the call
You getA programme that gets sharper each month instead of drifting.
When it fails
If a channel cannot be attributed to revenue within three months, we say so on the call and we stop running it. You do not pay to keep a losing channel alive.

See the method run on your own site.