Your Automation Worked Fine for Six Months. Then It Quietly Broke.

A client of ours once went six weeks without a single new-lead notification landing in their inbox. Nobody noticed, because the leads were still arriving on the website. The automation that was supposed to email the sales team the moment someone filled in the contact form had quietly stopped working after a routing rule changed inside their email provider. The leads sat in a form submissions table nobody checked, going cold, until someone happened to compare website analytics against their CRM and asked why the numbers didn't match.

That's the thing about automation failure: it almost never looks like failure. There's no error popup on anyone's screen, no alarm. The thing that used to run in the background just... stops, and because nobody was watching it closely in the first place, nobody notices it's gone until the cost shows up somewhere else.

Why automations break without anyone doing anything wrong

Most small business automations connect two or more systems that were never designed by the same team, and each one keeps evolving on its own schedule. A payment provider renames a field in its API response. A spreadsheet column gets reordered by someone who didn't realise a Zap was reading from column C specifically. A free trial on a connector tool expires and nobody renews it because the invoice went to an inbox that's no longer checked.

None of these are automation "bugs" in the way a faulty light switch is broken. They're the normal, expected consequence of connecting systems that change independently of each other, and they're exactly why an automation that worked perfectly on the day it launched can stop working eighteen months later with nobody having touched it.

The irony is that the better an automation works, the more dangerous its failure becomes. A manual process that goes wrong gets noticed quickly, because a person is doing it and people complain when something feels off. An automated process that goes wrong keeps running in the shape of doing nothing, and silence doesn't complain.

The warning signs worth actually checking for

A handful of patterns tend to show up before a workflow goes fully dark, if you know to look for them.

Numbers that used to move now sit flat. If your automated follow-up sequence used to generate a predictable trickle of replies and that's dropped to zero, that's worth investigating before you assume interest has dried up. Volumes that drift without explanation are the easiest signal to miss, because a slow decline over several weeks doesn't trigger the same reaction as a sudden stop.

A connected account quietly loses access. Most platforms will send a "please reconnect your account" email when a login token expires or a password changes, and most of those emails get ignored because they look like routine account admin. If anyone in the business changes a password on a tool that feeds an automation, that's the moment to go and check the automation still runs, not assume it will.

Someone changes a form, spreadsheet, or inbox rule without telling the person who built the automation around it. This is the most common cause we see, and it's rarely anyone's fault: a new team member reorganises a spreadsheet to make it easier to read, not realising three Zaps are reading from specific cell positions.

Build the checking-in habit before you need it

You don't need dedicated monitoring software to catch most of this early. Both Zapier and Make keep a run history for every workflow, usually under a name like "Zap History" or the scenario's execution log, and it takes about two minutes to glance at. Zapier can also email you automatically whenever a Zap errors out, which is a setting worth switching on the day you build the automation rather than the day you wish you had.

Put a specific, recurring slot in someone's calendar, not a vague intention to "keep an eye on it." Fifteen minutes on the first Monday of the month, looking at each automation's run history and asking "did this do what it was supposed to do the right number of times this month," catches almost everything before it becomes a real problem. It sounds almost too simple to be useful, and that's exactly why most businesses skip it and end up finding out about a broken workflow from a customer instead.

For anything genuinely important, like lead notifications, order processing, or invoice generation, a failure alert that goes to a real inbox someone checks daily is worth setting up even if it feels like overkill. The cost of an unnecessary email alert is a few seconds of annoyance. The cost of a silent failure on your lead pipeline is measured in actual lost customers.

What to do when you find a broken link in the chain

The instinct when something's clearly stopped working is to dive straight into fixing it, and that's usually fine for small issues. But before you do, it's worth asking how long it's been broken, because that tells you whether you also need to go back and manually recover anything that fell through during the gap. In the lead notification example above, the fix to the email routing rule took ten minutes. Working out which of the six weeks of leads had already gone cold, and which were still worth a call, took considerably longer and wasn't nearly as pleasant.

It's also worth asking why it broke rather than just patching the symptom. If a field rename broke a Zap once, the same field is likely to get renamed again by the same well-meaning colleague next quarter unless someone flags that it's load-bearing. A quick note in the spreadsheet, or a conversation with whoever owns that system, usually prevents the exact same failure from happening twice.

Automation that nobody checks on isn't really a system, it's a bet that nothing upstream will ever change. Most of the time that bet pays off for a while, and then it doesn't. A small, boring monthly habit of actually looking at whether each automation did what it was meant to do is the cheapest insurance you can buy against the kind of silent failure that costs far more to discover late than it ever would have cost to catch early.

If you've got automations running that nobody's checked on in a while, or you're about to build something new and want monitoring built in from the start rather than bolted on after something breaks, that's a conversation worth having early. You can book a free 15-minute consultation with Digital Hand and we'll talk through what's running in your business and whether it's actually still doing its job.