Automation and AI

Webhook

Also called HTTP callback, web callback

A message one system sends the moment something happens, so another system can react immediately instead of checking repeatedly.

Quick facts: Webhook

Category
Automation and AI
Also called
HTTP callback, web callback
Level
Intermediate
Affects
Alert speed, automation reliability, duplicate records, integration cost
Where to see it
n8n, Make, Zapier, and the developer or integration settings of most platforms
In this article4
  1. How a webhook works
  2. Why webhooks matter
  3. Where webhooks go wrong
  4. How to act on it

How a webhook works

The receiving system publishes a URL and says: send anything relevant here. The sending system is then told about that URL, and whenever the chosen event happens — an order placed, a form completed, a payment cleared, a subscription cancelled — it immediately sends a small message to it describing what occurred, usually including the record involved.

The receiver’s job is to accept the message and confirm it, which it does with an HTTP success response. If it fails to respond, or responds with an error, most senders will try again for a while and then give up. That retry behaviour is the part people forget: a receiver that is slow or briefly offline may lose messages permanently, and there is rarely any prompt to tell you.

The contrast is with polling, where a system asks another repeatedly whether anything has changed. Polling wastes requests when nothing has happened and still introduces a delay when something has. A webhook reverses the direction — the event announces itself — which is why it is the usual way an API delivers real-time news.

Why webhooks matter

They are what makes an automation feel immediate rather than scheduled. A lead alert that reaches a phone as the form is submitted is a different commercial proposition from one that arrives with the next batch, and the difference is usually a webhook rather than a better tool.

They also lower load and cost. Instead of a workflow waking up constantly to ask whether anything changed, it sleeps until it is told, which matters when your automation platform charges by the run or limits how often you may ask. For anyone stitching together a website, a CRM and a messaging tool, webhooks are the joins that carry the urgent events.

Where webhooks go wrong

They fail quietly, which is their defining hazard. A URL changes after a site migration, a receiving workflow is paused, a certificate expires, and messages simply stop arriving. Nothing errors on your screen, because the failure is on the sender’s side and the symptom is absence — leads that never appear, and a week that felt oddly calm.

Duplicates are the second problem. Retries and network hiccups mean the same event can arrive more than once, so a receiver that creates a record on every message will happily create several. Ordering is the third: messages are not guaranteed to arrive in the sequence the events occurred, so an update can land before the creation it belongs to. And because the URL sits open on the internet, anything that knows it can post to it, which is why senders offer a signing secret and receivers should check it.

How to act on it

Design the receiver to tolerate reality: treat every message as possibly repeated by matching on the event or record ID before acting, respond quickly and do the slow work afterwards, and verify the signature so only genuine messages are accepted. Never put anything secret in the URL itself, since it will end up in logs.

Then monitor for silence. Record when each webhook last delivered something and alert when a source that normally sends daily has sent nothing, because absence is the failure mode you will otherwise discover from a customer. Keep a written note of every endpoint, what sends to it and who owns it — an undocumented webhook is the thing that breaks a site migration. Treat that inventory as part of your workflow automation documentation, alongside every trigger you rely on.

Do and do not

Do

  • Verify the signing secret before trusting a message
  • Match on event or record ID to ignore duplicates
  • Alert yourself when an endpoint goes quiet

Do not

  • Put secrets or personal data in the endpoint URL
  • Assume delivery order matches the order of events
  • Change a site or URL without checking what posts to it

Questions people ask about this

What is the difference between a webhook and an API?

An API is the general interface one system offers others, and you call it when you want something. A webhook is the reverse: the system calls you, unprompted, the moment an event happens. Most platforms offer both, and a typical automation uses a webhook to learn that something occurred and an API call to fetch the details.

Do I need a developer to use webhooks?

Often not. Automation platforms such as n8n, Make and Zapier will generate a receiving URL for you, show the incoming data and let you build the following steps visually. A developer becomes worth involving when you need signature verification, custom error handling, or a receiver that must cope with high volume reliably.

How do I know if a webhook has stopped working?

You usually will not, unless you check, because a webhook that stops sends nothing at all rather than an error. Look at the sender's delivery log for failed attempts, and set up your own alert for silence: if a source that normally delivers every day has delivered nothing, something is broken and needs looking at.

Related terms

Found this useful?

Share it, or ask an AI to summarise it

Back to the glossary

Knowing the term is the easy part

Applying it to your own site and budget is the work. Book a call and I will tell you what actually applies to you.