It happens to everyone. The automation that ran flawlessly for six weeks suddenly goes quiet, or an email arrives from Make with the word "error" in it, and your stomach does a little drop. Did I break it? Is everything lost? Am I not cut out for this?
Take a breath. Here is the truth experienced automators know and beginners deserve to hear early: breakage is normal. Automations live in a world of other people's apps, expiring logins, and changing data, and sometimes that world shifts under them. A broken scenario is not a verdict on your abilities. It is a flat tire, annoying, fixable, and no reflection on your driving. This guide is the calm roadside manual.
Where Do You Look First When Something Breaks?
Every fix starts in the same place: the scenario's history. Open the scenario in Make and look at its list of past runs. This is the flight recorder, a log of every time the scenario ran, what data passed through, and how it went.
You are looking for two things:
- Runs with errors. Failed runs are marked distinctly in the list. Click the most recent one.
- The red module. Inside the failed run, the scenario map lights up like a subway diagram, and the module where things went wrong is marked in red, often with a little warning icon. Everything before it ran fine. Everything after it never got the chance.
Pause and appreciate what that red marker gives you. You do not have to wonder which of your seven modules misbehaved. Make is pointing at the exact station on the assembly line where the item fell off the belt. Click the red module and you can read the error message, plus see exactly what data arrived just before the failure. In two clicks you have gone from "it's broken" to "this specific step choked on this specific item," and that is most of the battle.
What Is the Error Message Actually Telling You?
Error messages look intimidating because they are written by programmers, for programmers, and they enjoy words like "invalid" and "unauthorized." But underneath the costume, almost every error a beginner meets is one of three plain-English sentences. Learn to spot which one you are hearing.
"I'm not allowed in anymore" (the expired connection)
Words to look for: unauthorized, authentication, invalid credentials, access denied, token expired.
This is the most common breakage in all of automation, and it is nobody's fault. When you connected Gmail or Google Sheets or your social platform to Make, the app issued a kind of visitor badge. For security, those badges expire, and apps sometimes cancel them early when you change your password or the service updates its security rules. When the badge lapses, your scenario knocks on the door and gets turned away.
Nothing is wrong with your scenario. The key just needs refreshing.
"You promised me data that isn't here" (the missing data problem)
Words to look for: missing, null, empty, invalid value, required field.
Your scenario was built expecting each item to have certain pieces: an email address in this field, a date in that one. Then real life sent through an item with a blank where the scenario expected a value. A form submitted without a phone number, a spreadsheet row where someone deleted a cell, an email with no subject line. The module that needed the missing piece stopped and raised its hand.
This error is actually a compliment to your automation: it processed everything faithfully until reality got weird.
"The place I called isn't answering" (the service problem)
Words to look for: timeout, service unavailable, rate limit, too many requests, server error.
Sometimes the app on the other end is having a bad day. It is down for maintenance, overloaded, or asking Make to slow down because too many requests arrived too fast. Your scenario, your data, and your connections are all fine. The restaurant is just closed right now.
How Do You Fix Each Kind of Problem?
Three diagnoses, three treatments, all beginner-friendly.
- For expired connections: reconnect the app. In Make, open the connections area, find the app with the problem, and choose to reauthorize it. You will sign in again, approve access again, and the badge is reissued. Then rerun the scenario. Total time: about two minutes, most of it spent finding your password.
- For missing data: add a filter, or a fallback. The durable fix is a filter right after the trigger that says "only continue if the important field exists." Items missing the field stop politely instead of crashing a module downstream. Where it makes sense, you can instead provide a default, like "no subject" standing in for a blank subject line. Our guide to filters and routers shows exactly how these gatekeepers work.
- For service problems: wait and retry. Truly. Give it a little while and run the scenario again, and check the service's status page if the outage persists. If a rate limit keeps appearing because your scenario processes big batches, spreading the work with gentler scheduling helps, and it saves operations too, as covered in our guide to Make.com operations.
One more move worth knowing: after fixing the cause, look at the failed run in history. Depending on your scenario settings, Make keeps incomplete executions around so you can retry them with the original data, which means the item that failed does not have to be lost. For scenarios handling anything important, that safety net is worth checking soon after a failure rather than weeks later.
How Do You Catch Breakages Before They Cost You?
The quiet danger with automation is not the error itself, it is the silence afterward. A broken scenario does not wave its arms. It just stops, and if you are not looking, you might discover a month later that no leads were logged since the connection expired.
Three cheap habits prevent that:
- Turn on error notifications. Make can email you when a scenario fails. Say yes to that. An error email is not bad news, it is the system working: you found out immediately instead of eventually.
- Do a one-minute weekly check. Glance at the history of your important scenarios. Green runs, moving data, done. It is the automation equivalent of checking the oil.
- Test after big changes. Changed a password? Renamed a spreadsheet column? Switched form tools? Run the affected scenario once manually and watch it complete. Most "mystery" breakages trace back to a change somebody made without rerunning the flow, and the somebody is usually us.
What If You Are Still Stuck?
Sometimes the error is genuinely cryptic, and that is not a wall, just a longer road. Copy the exact error message and search for it along with the app's name; the odds are excellent that someone in Make's community forum hit the same thing and posted the fix. Simplify to isolate: run the scenario with one test item and watch where it stumbles. And if a module has tangled itself beyond understanding, deleting it and adding a fresh one takes a minute and cures a surprising share of oddities.
Above all, keep the frame: every broken automation you fix teaches you how the whole machine actually works, in a way no tutorial can. The person who has repaired an expired connection, filtered out bad data, and waited out an outage is no longer a beginner. They are an automator with a toolbox, which is exactly what your future, more ambitious scenarios, like the ones from our seven starter automations, will need behind them.
Next step: The best way to get comfortable fixing automations is to have a few running. Visit the automation hub to build your next one, maintenance skills included.
