Editorial Intelligence needs evidence

Editorial Intelligence has reached the point where it needs external evidence more than it needs another framework.

That is a slightly uncomfortable conclusion for somebody who has spent the past few months building frameworks. But it is also the logical result of taking the central argument seriously.

Editorial Intelligence is concerned with how organisations generate and develop evidence, turn it into ideas and narratives, and learn from what happens next. It argues that many organisations do not simply have a content problem. They have an evidence problem: useful knowledge exists, but it is scattered, used once, disconnected from decisions or lost when the immediate project ends.

The methodology has developed substantially from practice. That gives it a useful foundation. It also gives it a limit.

---

The idea came from doing the work

I did not begin with a theory and go looking for examples that fitted it.

The idea emerged from years of working across journalism, content strategy, research-led campaigns, customer stories, events and AI-assisted editorial workflows. I kept seeing versions of the same problem.

An organisation commissioned research but treated the report as the end of the work. A customer conversation contained evidence that could shape strategy, but its value stopped at a single case study. An event produced useful observations, but the learning disappeared into a recap. A content team was asked for more output when the harder question was whether the organisation had anything distinctive and well evidenced to say.

The Editorial Intelligence cycle was developed to describe a better route: generate, capture, connect, synthesise, shape, activate and learn. The Evidence Engine goes deeper into how evidence can accumulate rather than reset with each campaign.

Those models are not inventions detached from work. They are attempts to document patterns I have repeatedly encountered.

But repeated experience is not the same as external evidence.

A coherent methodology can still be wrong

There is a particular risk in building a methodology from your own practice.

Every new observation has somewhere to go. The framework becomes more connected. The language becomes clearer. Examples begin to reinforce one another. Coherence starts to feel like validation.

It is not.

A methodology developed mainly from one practitioner’s experience can become an increasingly convincing account of how that practitioner thinks the world works. The ideas may be useful. The patterns may be real. But the person building the model is also choosing what to notice, how to interpret it and which contradictions deserve attention.

At some point, developing another internal framework produces diminishing value. It can explain the existing hypothesis in more detail without making the hypothesis more reliable.

Editorial Intelligence is at that point now.

From building the hypothesis to testing it

The next phase is an ongoing practitioner research programme.

The primary question is deliberately wider than Editorial Intelligence itself:

How does evidence become an influential idea inside an organisation — and where does that process break down?

I want to understand what happens between research, customer insight or internal expertise being created and somebody deciding to act on it.

That includes questions such as:

• Where does useful evidence come from?
• Who connects evidence across teams, projects and time?
• What makes an idea credible enough to influence a decision?
• Where does valuable knowledge get ignored, forgotten or used only once?
• What happens to good evidence after the immediate project ends?
• How is AI changing the work, and where is human judgement still necessary?

The emphasis is on reconstructing real examples rather than asking people whether they agree with a model. Practitioners do not need to know the Editorial Intelligence language, and the research is not a test of whether they already work in a particular way.

Disagreement is evidence too

This research is not designed to prove that Editorial Intelligence works.

Agreement may show that the methodology describes a recognisable problem. It may also show that I asked a leading question, recruited people too close to my own field or used language that made agreement easier than challenge.

Contradiction is often more useful.

A practitioner may show that knowledge is not being lost where I assumed it was. An established discipline may already explain part of the process better. A framework stage may turn out to be artificial, or two stages may be indistinguishable in real work. The strongest influence may come from relationships, timing or authority rather than the quality of the evidence itself.

Those are not awkward findings to explain away. They are findings that should change the methodology.

The research will therefore record negative examples, competing explanations and disagreement alongside supporting evidence. No single contribution will be treated as proof, and early observations will remain preliminary until they recur across different organisations, roles and ways of contributing.

Three ways to contribute

Different forms of participation reveal different things, so the programme offers three routes.

The questionnaire is a structured way to reconstruct one real example in your own time. It creates comparable responses while leaving room for the parts a fixed model may have missed.

A Compare Notes conversation allows more depth, follow-up and contradiction. It is a practitioner exchange about what actually happens, not a sales call or a disguised assessment.

A written exchange offers the same opportunity asynchronously for people who prefer to think on the page rather than in a form or meeting.

The Compare Notes page explains the research and the current ways to take part. The questionnaire is still being piloted, and the wider programme will develop as the research instruments are tested.

What happens to the evidence

The research needs to do more than produce a report.

Contributions will be analysed for recurring patterns, contradictions and gaps. Useful early observations can become articles or newsletter editions, with their limits made clear. Stronger findings can inform a first Editorial Intelligence report. Most importantly, evidence that challenges the methodology should change the methodology itself.

That creates the learning loop Editorial Intelligence has always argued for: research improves the model; the model shapes new work; the work produces new evidence; and the evidence changes what happens next.

I have built the hypothesis in public. The useful next step is to let other practitioners disprove, refine or strengthen it.

The next version of Editorial Intelligence should be shaped by what practitioners actually experience — not simply by what I think they experience.

Compare notes and choose how you would like to take part →