Editorial standards
ScrollInsights covers business and competition, banking and markets, cloud and FinOps, data and analytics, and cybersecurity for people who build and buy this technology. This page describes how we work. It is a description of controls we actually operate, not a statement of aspiration.
What we publish
We publish original reporting, original analysis, and original datasets. We do not republish other publications' articles, and we do not rewrite them. Where another outlet reported something first, we say so and link to them.
Sourcing and licensing
Every source we monitor carries an explicit licence classification recorded before anything is stored, and that classification is enforced by our database rather than by editorial habit. Sources fall into three groups:
- Fully licensed — public-domain government material, Creative Commons publishers, and company newsrooms. We may quote and build on this text directly.
- Summary-licensed — publishers whose terms grant display of the summary they syndicate. We reproduce that summary verbatim, attributed and linked, and never the article body.
- Link-only — publishers who grant nothing. We record the headline, the URL and the timestamp so we know the story exists, and we link to them. Their article text is discarded at the point of collection and never stored.
Because the discard happens at collection rather than at publication, there is no copy of a link-only publisher's article anywhere in our systems for a writer or an automated process to draw on. We also honour robots.txt, text-and-data-mining reservations and equivalent signals, and we keep a per-fetch record of what those signals said at the time.
Facts, not paraphrase
Where we write about something another outlet also covered, we work from the underlying facts — what was announced, by whom, on what date, at what price — and write independently. We do not follow another publication's structure, framing or narrative order.
Quotations
We do not manufacture quotations. Anything presented as a person's or an organisation's words comes from primary material and is checked against it. No automated process in our pipeline is permitted to produce attributed speech, and drafts that contain any are rejected before they reach an editor.
Human accountability
Every published article carries the name of a real, identified editor who reviewed it and is accountable for it. This is enforced as a database constraint: a record cannot enter the published state without an identified editor and a review timestamp attached. There is no administrative override, and no article has ever been published without one.
We never publish under invented identities. Every byline is a real person with a profile page. See how we use AI for what is and is not automated.
Corrections
We correct errors promptly and visibly, and we say what changed. Ourcorrections policy explains how to report one and what happens next.
Independence and disclosure
- Coverage is not for sale. No advertiser, sponsor or commercial partner receives or is offered favourable coverage, and none reviews articles before publication.
- Sponsored or affiliate placements are labelled as such wherever they appear, and carry the machine-readable markup search engines expect. They are produced separately from editorial.
- Where we write about a company we have a commercial relationship with, we disclose it in the article itself.
Data and methodology
When we publish a dataset or a measurement, we publish how it was collected, over what period, and what its limits are. Where a figure can be independently reproduced, we say how.
Contact
Corrections, questions and complaints:see the corrections page for how to reach the editor.