Brand suitability tiers, explained for Amazon DSP

Brand safety and brand suitability are two different jobs

Brand safety is the floor: content that is not appropriate for any advertiser, whatever the product. Think illegal activity, hate, terrorism or explicit material. Every reputable platform filters it out before a bid is placed, and an advertiser cannot opt back in.

Brand suitability sits above that floor and is personal to each brand. A news report about a war is a perfectly legitimate place for a newspaper subscription ad and a poor place for a children's snack. Neither placement is unsafe; one is unsuitable for one of those two brands. Suitability tiers exist so each advertiser can set its own line instead of accepting a single rule for everyone.

The practical difference matters for reach. A blanket "avoid anything sensitive" rule also blocks large amounts of ordinary news, sport and entertainment, which is where many shoppers spend their time. Tiers let you avoid what genuinely hurts your brand while keeping that inventory.

Where the tiers come from: the GARM framework

The usual vocabulary comes from the Global Alliance for Responsible Media (GARM), an initiative of the World Federation of Advertisers. Its Brand Safety Floor + Suitability Framework, first published in September 2020 and later extended to cover misinformation, defines the content categories advertisers care about and grades each one.

The categories are:

• Adult and explicit sexual content
• Arms and ammunition
• Crime and harmful acts to individuals and society
• Death, injury or military conflict
• Online piracy
• Hate speech and acts of aggression
• Obscenity and profanity
• Illegal drugs, tobacco, e-cigarettes, vaping and alcohol
• Spam or harmful content
• Terrorism
• Debated sensitive social issues
• Misinformation (the later addition)

The WFA discontinued GARM on 9 August 2024. The vocabulary outlived the organisation: verification vendors, platforms and agencies still describe suitability in its terms of floor, high, medium and low risk, so it is the most useful common language when you brief a team or read a report.

The four levels, in plain terms

For every category, the framework separates a floor from three risk tiers. The deciding factor is the treatment of the subject, not the subject alone.

• Floor: never monetised. Content that promotes or glorifies harm, such as instructions for violence or the illegal sale of weapons.
• High risk: graphic, explicit or sensational treatment, for example gratuitous violence shown in detail.
• Medium risk: dramatic depiction in an entertainment or news context, such as a crime drama or a report on a conflict.
• Low risk: educational, scientific or informational treatment, such as a documentary or a historical explainer.

A brand chooses, category by category, which tiers it accepts. A beer brand may accept medium-risk alcohol content and refuse all medium-risk violence; a streaming service may accept medium-risk violence because its own catalogue contains it. The tiers are the menu; the brand's policy is the order.

How brand suitability works on Amazon DSP

Amazon DSP applies baseline brand safety filters that every supply source has to meet. On top of that, in November 2025 Amazon announced an inventory-level brand suitability control with three settings: expanded, standard and restrictive. One setting covers Twitch, third-party websites and third-party mobile apps, so you no longer manage a separate rule for each supply source.

The setting can be made for the whole advertiser and adjusted per ad group, in the DSP console, through the API or in bulk sheets, and there is reporting by inventory tier and by content exclusion category so you can see where ads actually ran. Amazon describes the control as best effort: classification errors, language coverage and content that changes after it is classified can still produce the occasional misplaced ad.

Third-party verification remains available alongside it. Integral Ad Science lists brand safety and suitability among its solutions in Amazon DSP, together with viewability and invalid-traffic measurement, and DoubleVerify offers its own suitability controls for third-party supply. Teams that already hold a verification contract often keep it so that one policy applies across every platform they buy.

Amazon's own streaming inventory, such as Prime Video, is professionally produced and curated, and the November 2025 announcement does not list it among the inventory the setting covers.

Choosing a tier without giving away your reach

A narrower setting leaves fewer impressions eligible, so it costs scale and usually raises the price of each impression. The right setting is the loosest one your brand can genuinely live with.

• Restrictive suits brands with strict rules from legal or a regulator, products aimed at children, and launches where a single bad screenshot would do real damage.
• Standard is the sensible starting point for most consumer brands: it keeps mainstream news, sport and entertainment while avoiding the riskier treatments.
• Expanded suits retargeting and performance campaigns where reach drives results, as long as the brand has checked that nothing in the wider pool conflicts with its policy.

Set the advertiser-level default first, then tighten individual ad groups where the creative or audience calls for it. Changing the setting for one sensitive product is better than restricting the whole account.

A checklist before you launch

1. Write down the categories your brand must never appear next to, and the ones it is comfortable with. Use the category list above so legal, the brand team and the media buyer read the same words.
2. Map each category to a tier you accept, then pick the Amazon DSP setting that matches most of the account.
3. Decide whether a third-party verification vendor is required, and whether its segments should run pre-bid, post-bid or both.
4. Keep domain and app block lists short and specific. Long block lists copied from other accounts quietly remove good inventory.
5. Check the brand suitability report after the first week, then monthly, and after any major news event.
6. Record the policy and the date you set it, so the next person can tell a deliberate choice from a default.

reMKTR runs Amazon DSP as a managed service and sets these controls as part of the media plan, and it is part of the Full Circle group, with $500M+ in managed revenue across 100+ brands.