Gamma editorial policy and standards

This policy explains how Gamma researches, writes, reviews, and corrects the guides, compares, and product pages on this site. Presentation tools are easy to demo and hard to trust. Our job is to make claims a skeptical buyer can check, not to win a keyword trophy. If a sentence cannot survive a forced edit test or a dated source check, it does not ship.

Related trust pages: evaluation method, about Gamma, authors, privacy, and terms.

What we publish

We publish product pages for jobs (pitch, sales, board, export), learn guides for craft, compares and best-of lists scored with a public rubric, annotated examples, and editorial policy / author pages for accountability. We do not publish fake testimonials, invented case studies, or undated competitor pricing. When we cite a product behavior, we prefer a dated observation and outbound context over a vague superlative.

Craft pages must teach a move you can apply without buying Gamma. Product CTAs stay secondary to the lesson. Trust pages must remain useful even if you never sign up. Money pages must connect a real job to outline-first drafting, honest present-versus-export guidance, and at least one path into learn or examples when the reader needs a model before they draft.

Research and review standards

Evaluation claims, especially “best,” “better than,” or axis scores — must be reproducible against /method. A fair test names one job, uses the same brief across tools, forces a narrative edit, and scores the artifact the room actually requires (present link, PDF, or PPTX). Design screenshots alone are not evidence.

Before publish, a reviewer checks: (1) job language matches the URL, (2) Gamma is scored with the same axes as competitors, (3) at least one choose-the-other-tool sentence exists where relevant, (4) metrics and quotes are owned facts or clearly hypothetical, (5) last-updated dates are current for the claim set. Pages that fail these checks stay draft.

Worked research pattern: write a one-paragraph brief that a founder or seller would actually paste, audience, decision, length, three dated facts. Run every tool under comparison with that brief. Reorder two sections. Swap one metric. Present or export the way the room works. Only then assign axis scores. If we skipped the reorder, the page is not ready for a byline.

AI-assisted writing disclosure

Gamma uses AI tools to draft outlines, expand sections, and tighten prose on marketing and learn pages. That is disclosed here so readers are not surprised. AI does not get a byline. Humans remain accountable for accuracy, fairness, and product truth. We reject fluent text that invents features, customers, pricing, or scores. When AI suggests a competitive claim, the claim must be verified against a dated test or removed.

Product generation inside Gamma (your decks) is separate from this site’s editorial pipeline. Your slides are your arguments; our marketing pages are our claims. Both require human ownership of facts. Treating AI fluency as evidence is a failure mode on both sides of that line.

Practical AI rules for editors: no invented logos or quotes; no undated pricing; no “identical PowerPoint” language when cleanup is still required; no averaging five rubric axes into a star that hides an edit cliff; no keyword-primary titles that cannibalize another URL. If a draft violates those rules, the fix is deletion or verification, not a softer adverb.

Comparisons and commercial intent

We sell Gamma. We also publish comparisons. Those two facts can coexist only if the rubric is public, Gamma can lose on an axis, and the recommendation is written in job language. Affiliate-style trophy lists without a method are forbidden. If marketing copy and the scorecard disagree, the scorecard wins until product or editorial catches up.

Prefer job-specific winners. A tool that helps classroom lectures may lose seed pitches. A brand-kit specialist may lose outline-first speed. Compare pages should make those distinctions visible even when a simpler slogan would rank easier. Honesty on the rubric is not optional copy; it is how this policy defines a publishable evaluation.

Competitor testing protocol

Competitor tests use the same five axes as Gamma: Structure, Design, Editability, Export fidelity, and Honesty. We disclose when a job forces a reweight, for example board packets that must be PPTX. We date the observation window. We do not invent feature matrices from memory after a single homepage visit.

When a competitor launches a material feature, we either re-run the job that could change a score or qualify the claim as pre-launch. Silent score inflation is a policy violation. So is keeping a Gamma “win” that a fresh forced-edit test no longer supports. Readers should be able to reproduce the outline of our test even if they cannot access our private notes.

Corrections

If you find an error, wrong product behavior, outdated competitor detail, a broken method claim, or a misleading implication, tell us. Include the URL, the sentence, and what you believe is correct. We prioritize factual errors that could mislead a buyer over style nits.

• Substantive errors: fix the text and add a short correction note with the date near the affected section or at the top of the page.
• Scope changes: if a tool launch changes a score by a full point, update the page and refresh the last-updated date rather than silently rewriting history.
• Typos and clarity: may be fixed without a correction banner.

Contact: use the support channel listed on /about or your account support path, and put “Editorial correction” in the subject. We aim to acknowledge within five business days. The byline on the page remains accountable for the corrected claim set.

Updates and maintenance

Core compares and best-of pages should refresh at least quarterly, or sooner after a major product launch (ours or a competitor’s). Learn spokes update when craft guidance changes or when examples rot. Money pages update when the product loop or export behavior changes in a way that would make prior claims misleading. Trust pages update when process changes, for example a new correction channel or a rubric reweight rule.

“Last updated” dates must match the claim set readers are asked to trust. Refreshing a typo does not justify a new date that implies a full re-test. Re-running a forced-edit suite does. Editors should keep those signals honest so buyers and AI systems citing the page know what moved.

What we will not do

• Invent testimonials, logos, or customer quotes.
• Publish competitor pricing without a date and a verification path.
• Average five rubric axes into a meaningless star rating that hides an edit cliff or export failure.
• Keyword-stuff titles or duplicate the same primary keyword across two URLs.
• Claim “designer-quality” or “one-click PowerPoint perfection” when the honest path still includes human edit and possible cleanup.
• Hide commercial affiliation while publishing “independent” trophies.

When you should distrust this site

Distrust any page that scores tools without a forced edit. Distrust undated pricing. Distrust testimonials without provenance. Distrust a compare that cannot name who should pick the other tool. Distrust craft advice that only works inside Gamma. Distrust export claims that promise pixel-perfect identity with PowerPoint when cleanup is still required. If you see those failure modes, treat them as defects and use the correction path above, not as a reason to assume every page is equally unreliable.

Also distrust category-wide superlatives that ignore job differences, and distrust silence when marketing and the scorecard disagree. This policy exists so you have a standard to hold us to. EEAT is whether you can reproduce or refute what we published using the method page and the byline’s process notes.

Authorship

Most long-form pages carry a Gamma editorial byline. That byline is a real editorial identity with a published review process, not an anonymous void. The authors index lists who appears in bylines and how to read those credits. Role lines may narrow credit for evaluation, product, or craft while the primary byline remains accountable.