
TL;DR
- Notification infrastructure sends event-driven, multichannel messages your application triggers: password resets, OTPs, fraud alerts, onboarding, payment dunning.
- Marketing platforms send campaigns a person schedules against a list or segment: newsletters, promotions, re-engagement series.
- Courier does both jobs from one platform: product notifications from your code, plus journeys and broadcasts that marketers build in the dashboard on the same objects your engineers manage through the API.

Notification infrastructure handles event-driven, multichannel messages triggered by application code. Marketing platforms handle campaign-driven sends scheduled against audience cohorts. Which layer a message belongs on depends on where the trigger lives, who edits the message, and what sender reputation it needs.
The question comes up in vendor conversations: "Are you a Braze or Customer.io alternative?" Often, yes. Courier runs product notifications and lifecycle journeys from one platform, and everything in the UI is available via the API, CLI, and MCP server, so a marketer can build a journey on the canvas while an engineer or an AI agent builds the same journey in code. If your marketing suite mostly carries onboarding, re-engagement, and product-triggered sends, Courier can take all of it over (Courier vs Braze and Courier vs Customer.io have the side-by-side). If you've built years of newsletters and promotions in a marketing suite, run Courier alongside it.
| Axis | Notification infrastructure | Marketing platform |
|---|---|---|
| Trigger source | Event-driven. Application code emits an event. | Campaign-driven. A marketer schedules a send. |
| Build surface | For humans and AI agents: everything in the UI is available via the API, CLI, and MCP server. | Dashboard first, built for marketing teams. |
| Channel choice | Routed per message: user preferences, provider failover, fallback to another channel | Picked per campaign or campaign step |
| Audience model | Per-user identity tied to product events and attributes | Lists, segments, campaign cohorts |
| Sender reputation | Transactional reputation. High deliverability for operational messages. | Marketing reputation. Subject to engagement-based throttling. |
| Cost model | Usually per message or per event. Scales with product activity. | Usually per profile, contact, or MAU. Scales with audience size. |
| Typical use | Password reset, OTP, payment confirmation, fraud alert, multichannel onboarding, support escalation, multi-step approval | Newsletter, promo campaign, re-engagement series, lifecycle marketing |
Vendors in the notification-infrastructure category include Courier, Knock, Novu, and OneSignal. Vendors in the marketing-platform category include Braze, Iterable, Klaviyo, Customer.io, Bloomreach, and Optimove. Courier spans both: notification infrastructure underneath, with journeys, broadcasts, and audiences for marketers on top.
Provider APIs like Twilio, SendGrid, Resend, and FCM are a third layer entirely. They're the carriers. Both notification infrastructure and marketing platforms send through them. For the providers Courier connects to, see the integrations overview.
The signal is the trigger. If a message is triggered by something happening inside your application, notification infrastructure is the right layer. A few concrete examples:
What these share: a product event starts the message, the channels are mixed, and every send needs preference checks, provider failover, and its own delivery log.
For more on the email side, see the transactional vs marketing email post.
Already running a marketing platform but missing the product notification layer? Courier sends product notifications and lifecycle journeys across email, SMS, push, in-app, Slack, and Microsoft Teams, with an API for engineers and a visual builder for marketers. Start free with 10,000 sends a month.
The signal here is also the trigger, from the other direction. If a person schedules one send for a whole group, it's campaign work. Examples:
What these share: a marketer owns the work. Audience selection happens in a UI against precomputed segments. The send is scheduled, not real-time. Sender reputation is managed for marketing-class engagement metrics (opens, clicks, unsubscribes).
Marketing suites like Braze, Iterable, Klaviyo, and Customer.io were built for this work, and they're dashboard first.
Courier covers the same jobs next to your product notifications. Broadcasts send one message to a list or audience, now or on a schedule. Audiences are rule-based segments Courier keeps current as profiles change. Journeys runs the multi-step lifecycle and re-engagement flows. An engineer or an AI agent can create and update the same campaign a marketer built on the canvas.
The full stack at a typical growth-stage SaaS company looks like this:

The two layers don't overlap cleanly. A welcome series is the canonical edge case, and the deciding factor is the trigger. If a welcome message fires the moment a user signs up (an event in your application), it's a journey started by that event, even when a marketer wrote it and edits it. If it's a scheduled cadence sent to everyone on a list, it's campaign work. The customer journey orchestration tools roundup covers the vendor landscape if you're still picking.
Transactional and product notifications grow in volume and complexity over time. If you started with a marketing platform doing both, you'll eventually pull the transactional traffic out, because the platform's sender reputation, latency, and pricing model don't fit it. If you started with direct provider integrations, you'll eventually move them behind notification infrastructure, because maintaining five APIs with five sets of templating, failover, and preference logic stops being feasible.
Notification infrastructure isn't an email, SMS, or push provider. It sits above the providers and orchestrates them.
The difference matters because it sets the cost model, the deliverability model, and the lock-in story. You keep your provider contracts: Twilio for SMS; Resend, SendGrid, or Amazon SES for email; FCM and APNs for push. The infrastructure layer routes each message through the provider you configured, falls back to a backup if the primary returns an error, and logs the result.
In Courier, you bring your own providers and keep your sender reputation and domains. The practical impact:
With the provider relationship in your hands, you control the deliverability of your transactional traffic. Courier charges per message sent, and a notification sent across multiple channels counts as one send, which matches how product traffic scales. See Courier pricing for the current breakdown.
The most common mistake is forcing one tool to do the other's job.
Sending transactional traffic through a marketing platform. You pick Customer.io or Braze for marketing, then route password resets and receipts through the same tool because it's already integrated. Deliverability suffers when transactional sends share a marketing sender reputation, and on a per-profile plan you pay for every profile in your database, including users who only ever get receipts.
Building campaigns on an API-only tool. You pick an API-only notification tool for everything, including marketing campaigns. Every copy change and every new segment now needs an engineer, so marketing eventually buys its own tool and the campaigns move. Pick infrastructure marketers can work in directly. In Courier, marketers build journeys on a visual canvas and design templates in Design Studio, and engineers manage the same objects through the API, so neither side waits on the other.
Treating provider APIs as infrastructure. Direct integrations with Twilio, SendGrid, FCM, and Slack work fine until you need preferences, failover, cross-channel templating, or observability. At that point you either build the infrastructure layer yourself or buy it. The best email API providers writeup goes deeper on where the provider layer ends.
Confusing Resend with a notification platform. Resend is an email API, focused on email delivery. Pair it with notification infrastructure and you get both: Resend delivering email underneath, cross-channel orchestration above. They aren't substitutes for each other.
Buying every layer on day one. With three message types and 10k users, you don't need a marketing suite, notification infrastructure, and a CDP. Start on one platform that covers product notifications and lifecycle sends (Courier's Developer plan is free and includes 10,000 sends a month), and add layers as message types multiply, your teams specialize, and volume grows.
Moving transactional and lifecycle sends off Customer.io? See Courier vs Customer.io and the Customer.io migration guide.
Notification infrastructure and marketing platforms are two layers of the same stack. Ask three questions about each message you send:
If the trigger is a product event and the message needs transactional reputation, route it through notification infrastructure. If it's a scheduled send to a group, it's campaign work: Courier's broadcasts and journeys can handle it, or it stays in the marketing suite you already run. The line between the layers is where the trigger originates.
Comparing notification platforms? See Knock alternatives, Courier vs Novu, and Courier vs SuprSend.
Notification infrastructure is the layer that turns product events into multichannel notifications. Your application emits an event (a payment processed, a fraud signal, a build completed) and the infrastructure handles routing, channel selection, templating, provider failover, preference checks, and delivery logging. Courier, Knock, Novu, and OneSignal operate in this category.
Courier covers both jobs. It's notification infrastructure for the messages your product triggers, and it runs the lifecycle work marketers own: journeys built on a visual canvas, broadcasts to a list or audience, and templates designed in Design Studio. Marketers, engineers, and AI agents work on the same objects. If you already run campaigns in Braze, Iterable, Klaviyo, or Customer.io, you can keep them there and run Courier alongside.
Braze is a marketing suite. A Braze Canvas has to be built and saved in the dashboard before the API can trigger it, and Braze doesn't publish its pricing. Courier runs product notifications and lifecycle journeys on one platform, and an engineer or an AI agent can create the same journey a marketer builds on the canvas. Courier charges per send, and the Developer plan is free and includes 10,000 sends a month.
Not always. If your marketing suite mostly carries onboarding, re-engagement, and product-triggered sends, Courier can run all of it next to your product notifications. Keep a separate marketing suite when you've already built a large library of campaigns in it. Either way, place each message by its trigger: product events go through notification infrastructure, and scheduled sends to a group are campaign work.
Transactional email is triggered by a user action or product event and carries information the recipient is expecting (a receipt, a password reset, a shipping confirmation). Marketing email is sent to a list as part of a campaign and isn't tied to a specific user action.
Resend is an email API, which makes it a provider rather than notification infrastructure or a marketing platform. Courier routes email through Resend the same way it routes SMS through Twilio or push through FCM. The two are complementary: Resend delivers the email, and Courier handles cross-channel orchestration, templating, preferences, and failover above it.
Yes. Start with the transactional and product-event traffic that grew inside your marketing platform, since that's where its sender reputation and pricing model fit worst. On Courier, lifecycle journeys and broadcasts can move too, so you decide whether your campaigns stay in the marketing suite or come over with everything else.

Catch broken lifecycle journeys with synthetic users and AI
Paste one prompt into a coding agent and get a local test of your onboarding journey: synthetic users, real timing, and every send explained. Catch stale reminders before your customers do.

Control how fast Courier sends to your providers
You shouldn't have to think about how fast you send to Courier. Set a send rate limit on any provider, and Courier sends your notifications to it at the rate you choose. Anything over the limit is queued, not dropped.

Know the moment users change notification preferences
Every preference change in Courier now reaches your systems two ways. A `preferences:user:updated` webhook fires the moment a user opts in, opts out, or changes a topic's channels, wherever they made the change. `GET /preferences/logs` returns the history, newest first, with the previous value where Courier knows it. Keep your CRM, warehouse, or marketing tools in step with what users chose.