Buyer-Seller Messaging Service: How to Stay Inside Amazon’s 24-Hour Window

Most sellers treat Buyer-Seller Messaging as a minor compliance chore. That’s backward. A late reply can cost you a negative feedback, an A-to-z claim, and a hit to your Account Health Rating that takes months to recover. The 24-hour clock isn’t a suggestion - it’s one of the metrics Amazon uses to decide whether you keep your Buy Box eligibility.

This page covers what the service does, how it’s changed through 2026, and where it fits against the newer tools - the Request a Review button, the Solicitations API, and the AI content filter that now scans every outgoing message in real time.

Table of Contents

• What the Buyer-Seller Messaging Service Is
• Where You Access It
• The 24-Hour Response Clock (and What Breaks It)
• Critical vs. Non-Critical Messages
• What You Can’t Send
• Request a Review Button vs. Messaging
• Automating Messaging at Scale
• Common Mistakes That Get Accounts Restricted
• Action Plan for This Week
• FAQ

What the Buyer-Seller Messaging Service Is

Treat Buyer-Seller Messaging like an SLA channel, not a marketing lane. Amazon masks both email addresses behind encrypted aliases, keeps the entire thread for dispute resolution, and grades you on whether you respond inside 24 hours.

In 2023 Amazon narrowed the scope to order completion and customer service. Marketing, review solicitation, and external links are out. Since the March 2026 SP-API release, every outgoing message - typed in Seller Central or sent via API - passes through an AI content filter before it reaches the buyer. Failed messages get silently blocked, returned with an error, or queued for human review.

Where You Access It

Four entry points, depending on how you work:

Entry point	Best for
Seller Central → Messages	Desktop, template management, attachments
Amazon Seller App	Mobile responses, push notifications
Forwarded email	Monitoring without logging in (Settings → Notification Preferences → Messaging)
SP-API Messaging v1	Programmatic automation

Inside Seller Account, the Manage Orders page links directly into a message thread scoped to a specific order - faster than hunting through Messages. Click the buyer’s name and any template auto-includes the Order ID and translates to the buyer’s language. The inbox supports Search by Order ID, ASIN, or buyer alias. Note: sellers with inactive listing status still receive messages and are still on the 24-hour clock.

The 24-Hour Response Clock (and What Breaks It)

It’s Monday 8 a.m. You check Messages and there’s a buyer inquiry from Saturday 10 p.m. That message is already 34 hours old. In Amazon’s eyes, you missed.

Amazon’s Contact Response Time Performance Metrics dashboard measures the percentage of buyer-initiated messages you reply to within 24 hours. The clock starts when Amazon delivers the message to your inbox. It runs continuously - weekends, holidays, the day after Prime Day - no pauses. (See Customer Metrics FAQs for the full metric family.)

Here’s the threshold structure most sellers don’t realize they’re being graded on:

CRT score	Status	What happens
≥90% within 24h	Safe	No dashboard warnings
80-89%	Yellow zone	Warning visible on Account Health dashboard
Below 80%	Red zone	Contributes to Account Health Rating decline; combined with other metric issues, can trigger performance notifications

Auto-replies (“Thanks, we’ll be in touch soon”) don’t count as responses - Amazon’s parser wants a substantive human reply on the original message. You can, however, flag a message as “No Response Needed,” which removes it from the CRT denominator. Use it for buyer out-of-office replies, thank-yous, and genuine non-questions. Patterns of over-flagging are themselves reviewable.

The math of a slow response is simple and ugly. At 100 messages a week and an 85% on-time rate, you’re leaving 15 late messages on the table. Industry data suggests late replies correlate with roughly 3× higher rates of negative feedback and A-to-z claims. For a low-volume seller, one successful A-to-z claim can push Order Defect Rate above the 1% ceiling - and that’s the metric that actually suspends accounts.

Critical vs. Non-Critical Messages

Not every message reaches the buyer. Amazon classifies outgoing messages as either critical or non-critical, and the distinction decides whether a buyer’s opt-out preferences apply.

Critical messages are the ones required for the order to be completed: custom product confirmations, delivery scheduling, shipping address clarifications. Those bypass the buyer’s opt-out. Non-critical messages - shipping confirmations, thank-yous, general correspondence - get silently blocked if the buyer opted out of non-essential communication.

You can’t see whether a buyer has opted out, and you can’t override it - the [Important] subject-line tag that worked before 2023 was deprecated precisely because sellers abused it to force-deliver marketing. Amazon now assigns criticality from the template you pick. Use the right template and critical messages go through. Use a generic template and they may disappear into the void.

What You Can’t Send

The prohibited-content list is short and enforced aggressively. In 2026, it’s enforced by AI scanning every outbound message before delivery - not just by buyer reports.

Never send	Why
External URLs of any kind (even inside logos)	Redirects buyers off Amazon
Marketing, promotional text, coupons, upsells	Messaging is for order completion only
Direct review solicitation (“please leave a 5-star review”)	Review manipulation risk
Leading language (“if you’re happy,” “most customers rate us 5 stars”)	Biased review influencing
Personal data requests beyond order needs	Privacy violation
Off-Amazon communication invites (“email us at…”)	Off-platform redirect
Emojis and decorative symbols	Flagged as persuasive content under 2026 scanning
Attachments with prohibited content (marketing PDFs, review cards)	Same rules apply to attachments

The most common mistake is the hidden URL in a logo. Your logo contains your website address, you embed it in every signature, and every message is now a policy violation. Amazon’s AI scanner reads images. Check your templates.

Attachments are permitted - Word, PDF, plain text, and common image formats up to 10 MB - but pass the same content scan as the message body.

Request a Review Button vs. Messaging

If you’re trying to get more reviews, Buyer-Seller Messaging is not the tool - it’s explicitly not allowed to carry review requests. Amazon built the Request a Review button for that, available per-order in Manage Orders from day 5 through day 30 after delivery.

The two tools serve different jobs:

	Request a Review button	Buyer-Seller Messaging
Message type	Fixed Amazon template (no edits)	Free-form
Purpose	Product review + seller feedback request	Order completion, customer service
Timing window	Day 5-30 after delivery	No strict window
Customization	None - auto-translated	Full customization for allowed purposes
Counts toward CRT	No (no-reply address)	Yes - buyer-initiated messages count
Use for reviews?	Yes	No - policy violation

One rule worth tattooing: one review request per order, one channel only. Clicking the Request a Review button and sending a BSM message asking for a review on the same order can trigger a 30-day proactive-messaging restriction. Amazon treats the duplication as manipulation. (See Customer Feedback and Removing Negative Feedback for what happens when reviews go wrong.)

Response rates for the Request a Review button run 5-15% depending on category and price point. Automation tools that click the button via the Solicitations API achieve similar rates - lift comes from timing (7-14 days post-delivery is the sweet spot), not from custom messaging.

Automating Messaging at Scale

Above roughly 100 orders a day, clicking Request a Review manually stops being realistic. The SP-API offers two paths. The Solicitations API exposes createProductReviewAndSellerFeedbackSolicitation - one call per order, same 5-30 day window, no CRT impact. The Messaging API v1 handles typed operations for order communication: delivery details, order details, warranty, digital access keys, negative feedback removal, unexpected problem. Each has a narrow approved use case.

Any tool that claims to “send custom review-request text via BSM automation” is selling you a suspension. Compliant automation (Helium 10, Jungle Scout, FeedbackWhiz, Sellerboard, SageMailer, eDesk) uses the Solicitations API for reviews and reserves the Messaging API for operational communication.

Rate limits sit around 1 request per second per endpoint with a burst of 5 - not enough to matter until you’re processing several thousand orders a day.

Common Mistakes That Get Accounts Restricted

These seven mistakes account for most messaging-related enforcement actions.

The hidden URL in a logo. Your logo contains your website, you added it to every signature, and now every message fails the AI scan. Strip the URL from the image - or remove the logo from signatures.

Leading language. “If you’re happy with your purchase…” or “Most customers rate us 5 stars” biases toward positive reviews. It violates policy even when you never explicitly ask for 5 stars.

Using both Request a Review and a BSM review message on the same order. The easiest way to trigger a 30-day proactive-messaging restriction. One channel per order.

Confirming off-platform channels. When a buyer writes “just email me at gmail,” a polite acknowledgment is itself a violation. Redirect to Amazon messaging only.

The [Important] tag. Deprecated in 2023. Adding it does nothing useful and may flag the message. Pick the correct template instead.

Branded PDFs with a website footer. Attachments pass the same scan as the body. A warranty PDF with your URL at the bottom is an external link under the policy.

Over-flagging “No Response Needed.” The flag is for buyer out-of-office replies and genuine non-questions. Using it on real customer messages to game CRT triggers pattern detection.

Action Plan for This Week

If your CRT is below 90% this week, fix staffing before you fix copy - an audit won’t help if nobody is covering Sunday afternoon. Otherwise, three moves: audit every message template for hidden URLs and leading phrases; set an internal SLA shorter than 24 hours so you have a buffer; and if you’re not already automating Request a Review through the Solicitations API at 7-14 days post-delivery, start this week.

FAQ

What is the Amazon Buyer-Seller Messaging Service? Amazon’s encrypted, alias-based channel between buyers and third-party sellers. Both sides communicate through anonymized addresses, and Amazon retains every thread for dispute resolution. Current scope: order completion and customer service only - not marketing, not review solicitation.

What is the response time SLA for Amazon seller messages? 24 hours, including weekends and holidays. Target Contact Response Time is 90% or higher; below 80% contributes to Account Health Rating decline.

Can I ask for reviews in a Buyer-Seller message? No. Use the Request a Review button in Manage Orders - available from day 5 through day 30 after delivery. Using both channels on the same order can trigger a 30-day messaging restriction.

Do Request a Review clicks count toward Contact Response Time? No. The button sends from a no-reply Amazon address, so no response obligation is generated. Only buyer-initiated messages count toward CRT.

What file types and size limits apply to attachments? Word (.doc/.docx), PDF, text (.txt), and images (.jpg, .gif, .png, .bmp, .tiff). Maximum size 10 MB. Attachments pass the same content scan as the message body.

What’s the penalty for sending a prohibited message? A single clear violation triggers a warning and blocked message. Repeated minor violations lead to a 30-day proactive-messaging restriction. Severe violations - off-platform redirects, review manipulation - can result in immediate account suspension.