1. Newsletters: an owned channel with an editor’s voice
A newsletter reaches a named inbox on a schedule the publisher controls, which makes it the most durable digital format a desk can operate. It also has unusual editorial properties: the reader has opted in, the sequence is fixed, and there is no ranking layer between the editor’s judgement and the reader’s screen. That argues for genuine selection rather than an automated digest of everything published. The constraints are equally specific. A newsletter cannot be silently amended once sent, so verification has to be complete before dispatch, and the correction route is a follow-up note in the next edition or a corrected item on the site linked from it.
2. Audio: structure carried by the ear
Spoken journalism cannot rely on a reader’s ability to re-read a sentence. Attribution has to arrive before the claim rather than after it, numbers have to be rounded to something the ear can hold, and nested clauses have to be unpicked into separate sentences. Documentary and interview formats introduce their own duties: consent to record, clarity about what will be broadcast, and editing that preserves the meaning of an answer rather than merely its words. Transcripts should be published alongside episodes, both because they make audio accessible and because they make it searchable, quotable and correctable.
3. Vertical video and the grammar of the small screen
A vertical clip is usually watched once, on a handheld screen, often without sound, in a feed that offers an endless alternative. The practical consequences are concrete: establish the subject immediately, burn in legible captions, keep the essential information away from the edges where interface elements sit, and design the first frame as a headline rather than a title card. The editorial risk is compression. A clip short enough to hold attention may be too short to carry the qualification a claim needs, and the honest answer is sometimes that the story does not fit the format and should be published another way.
4. Live blogs and rolling correction
A live blog is a published draft. Its value is speed and its hazard is that early entries remain visible and quotable after they have been overtaken. Handle this structurally rather than by hoping readers scroll: timestamp every entry, mark superseded entries visibly rather than deleting them, keep a running summary at the top that reflects the current state of knowledge, and state explicitly when something is unconfirmed. At the end, publish a settled article and link the live blog to it, because an abandoned live blog left as the permanent record of an event preserves every intermediate error in searchable form.
5. Interactive explainers
Interactivity earns its cost when the reader’s own situation changes what the story means — a map they can locate themselves on, a table they can filter to their county, a calculator that turns an abstraction into a quantity. It is wasted when it merely animates a static finding. Interactive work also brings obligations that plain articles avoid: it must degrade gracefully without scripting, it must be operable by keyboard and legible to a screen reader, its data source and calculation must be documented, and it must be maintained, because an interactive feature quietly breaking is worse than a static graphic that never moves.
6. Synthetic voice and generated assets
Machine-read audio versions of articles and generated illustrations are now ordinary production options, and both raise the same question: what is the reader or listener being told? A synthetic read-out should be identified as one, both in the item and in the audio itself, so that it is not mistaken for a reporter’s voice when shared as a clip. Cloning the voice of a real person requires their agreement and a clear reason. Generated illustration should never depict an event as if it were photographed; where an image stands for something rather than recording it, the caption must say so.
7. Accessibility is part of the format, not an extra
Each format carries a specific accessibility duty. Audio needs transcripts. Video needs accurate captions and, where visual information carries meaning, a spoken or written description. Charts need a text alternative that states the finding rather than describing the shape. Live blogs need a heading structure that lets a screen reader move between entries. Automatic captioning is a starting point and not a finished product: it reliably mishandles names, places and technical terms, which are precisely the words a news item cannot afford to get wrong. Budget the correction time when the item is commissioned.
8. Choosing the format from the story
The reliable sequence is story first, then audience, then format. Ask what the central evidence is: a document, a place, a voice, a dataset, an unfolding sequence. Documents suit text with excerpts. Places suit photography and mapping. Voices suit audio. Datasets suit charts and filters. Unfolding events suit live formats followed by a settled account. Then ask what the format costs to produce and maintain, and whether the desk can sustain it beyond the first three editions. A format adopted because it is fashionable and abandoned six weeks later costs more credibility than it ever earned in reach.
Formats compared
| Format | Effort per item | Shelf life | Correction route | Core accessibility duty |
|---|---|---|---|---|
| Newsletter | Medium | Short | Note in the next edition | Plain-text alternative |
| Podcast episode | High | Long | Corrected note plus updated transcript | Full transcript |
| Vertical video | Medium | Very short | Replace the asset and state why | Burned-in and platform captions |
| Live blog | High while running | Very short | Marked in place, superseded entries retained | Heading structure per entry |
| Interactive explainer | Very high | Long | Amended data note with a date | Keyboard operation and text alternative |
| Standard article | Low to medium | Long | Amended in place with a correction note | Semantic headings and image alt text |
Effort figures are relative and desk-specific. The column that most often decides the question is maintenance, which is rarely estimated at commissioning stage.
Format selection and sign-off
- The central evidence of the story has been named before the format is chosen.
- The format can carry the necessary qualification, not only the headline claim.
- Production and maintenance effort has been estimated for at least ten editions.
- A correction route exists for the format and has been written down.
- Transcripts, captions or text alternatives are budgeted at commissioning.
- Interactive elements degrade gracefully without scripting and work by keyboard.
- Any synthetic voice or generated asset is labelled in the item and in the asset.
- Live coverage has a planned end point and a settled article to replace it.
If a format cannot pass this list, publishing the story in a simpler form is usually the better editorial decision.
Questions readers ask
Is a newsletter worth the effort for a small desk?
Often yes, because it is the cheapest owned channel to start and the least exposed to someone else’s ranking changes. It also imposes useful discipline: selection, sequence and a fixed deadline.
Do short videos have to be superficial?
No, but they have a hard limit on how much qualification they can carry. When the qualification is essential to the accuracy of the claim, the story belongs in a longer form or should link clearly to one.
Are automatic captions good enough to publish?
As a first pass only. They routinely mishandle proper nouns, place names and technical vocabulary, so a person should correct them before the item is published.