When a worker job fails, you hear it in 2 minutes. Not from a silent feed.

The 7 AM post that never went out

A scheduled post fails at 7 AM. You find out at 6 PM from a client who noticed the feed went cold. That is 11 hours of silence on a channel you told yourself was handled. The post didn't publish, and nobody told you. The client did.

That gap is the real problem. Not the failed post. The 11 hours. You can fix a post in 90 seconds. You can't undo the client's impression that you're not on top of it, especially if that client pays you a retainer to be on top of it.

The same thing happens with your newsletter. You set it for Thursday at 9 AM, you get busy, and at 4 PM someone replies to last week's issue asking if they missed one. The ai email marketing and newsletter tool for small teams should have flagged it. If it didn't, you're the alert system, and you're already doing three other jobs.

Most schedulers treat a failed job as a log entry. You have to go looking for it. You won't. Nobody does at 9 PM after a full day. So the failure sits there until a customer finds it for you.

The fix is a check that takes 2 minutes at night. Not a dashboard deep dive. A short, repeatable routine that tells you three things about every scheduled run: did it publish, did it land on the right channel, did the captions come through. If all three are yes, you close the laptop. If not, you know exactly which channel to look at.

Two minutes at night beats an apology email in the morning. That's the whole trade.

Check one: did it publish at all

Open the dashboard. Look at the runs scheduled for today. Each one has a status. Published, scheduled, or failed. You're looking for failed. You're also looking for scheduled on something that should have gone out 3 hours ago. That second one is the quiet killer, because a job stuck in scheduled looks fine at a glance.

The dashboard flags any run that didn't go out and tells you which channel to look at. That's the point of the flag. It doesn't just say error. It says the Instagram run at 7 AM didn't publish. You click it, you see the status, you move on.

If it failed once, re-run it. That's it. One click. Most failures are transient: a token expired, an API hiccuped, a file didn't upload. Re-running fixes the majority of them, and you get the post out by 8 PM instead of finding out tomorrow.

If it failed twice, stop. Don't touch the post. Re-running a broken connection just burns another cycle. The problem isn't the post, it's the account link. Go to the channel connection settings, check that the account is still authorized, and reconnect if you see a warning. Then re-run the job.

That's the rule. Failed once, re-run. Failed twice, check the account connection before you touch the post. It saves you from editing a perfectly good post while the real problem sits in the connection.

When you fix it at 9 PM, the post goes out and the feed doesn't show a gap. Your retainer client sees the normal rhythm. You see the same thing. That's the win: no apology email, no explanation, no quiet doubt about whether you're worth the monthly fee.

Check two: did it land on the right channel

A post can publish successfully to the wrong place. It happens more than people admit. You scheduled an Instagram post and it went to Facebook because the account connection mapped them together. The job says published. The channel says otherwise.

So the second check is channel confirmation. Open the run, look at which account it published to. If you run 3 channels, this takes 20 seconds. If you run 6, it takes a minute. You're not reading the post. You're reading the destination.

This matters most when you're cross-posting. One piece of content, 3 destinations. If one destination is wrong, you've now got a post on a page that shouldn't have it. That's harder to undo than a missed post. You can always publish late. You can't always unpublish quietly.

If the channel is wrong, fix the mapping first. Don't delete the post and re-do it. Go into the agent settings and correct the channel assignment, then re-run. Otherwise the next scheduled post repeats the same mistake and you're doing this check every night for the same reason.

The business analytics dashboard software view helps here. You can see runs by channel and spot the one that never gets activity. If your Facebook page hasn't received a post in 6 days but your schedule says it should have, that's your signal. No customer had to tell you.

The customer win is simple. You catch a wrong-channel post before a customer does, not after. That's the difference between a 2-minute fix and a 10-minute apology.

Check three: did the captions come through

The post published. It's on the right channel. But the caption is empty. Or it's the caption from the previous post. Or the alt text didn't attach and the image is a blank square on someone's phone.

This is the third check, and it's the one most people skip. Published and right channel doesn't mean complete. A post without its caption is a silent post. It looks broken to anyone who sees it.

Open the published version. Read the first line. If it matches what you approved, you're done. If it's blank or wrong, delete and re-run. This takes 30 seconds per post, and most nights you have one or two posts to check.

The caption problem usually traces back to how the job was set up. If you approved a draft and the caption field was empty at approval, the job publishes empty. The check catches it, but the better fix is to make sure the draft has the caption before you approve. That's a 5-second glance at approval time that saves a 10-minute fix at night.

For newsletters, the equivalent is the subject line and the preview text. If the subject line didn't come through, the email sends with a blank subject. People delete those without opening. You won't see a bounce. You'll see a lower open rate a week later and wonder why.

The ai email marketing and newsletter tool for small teams should confirm the send the way it confirms a social post. Published, right list, subject line intact. Three checks, same routine, whether it's a tweet or a newsletter.

If the caption didn't come through, re-run once. If it fails the same way twice, the draft is the problem, not the connection. Rewrite the caption, save it, re-run. That's the whole fix.

What to do when a job fails twice

Twice means the problem is structural. The post is fine. The connection is broken. Check the account connection before you touch the post. This is the rule that saves the most time because it stops you from editing content that was never the issue.

Go to the channel settings. Look for the account that failed. You're checking for an expired token, a revoked permission, or a password change on the other end. Any of those will fail every job on that channel until you reconnect. Reconnect, then re-run the failed job.

If the connection looks fine, check the file. A video that's too large or an image format the channel rejects will fail every time. Swap the file, re-run. If that works, the original file was the problem, and you've learned something about your export settings.

If it still fails after a reconnect and a file swap, stop re-running. You're now past the 2-minute check. This is a 10-minute problem, and it's worth doing in the morning with fresh eyes. Note which channel failed, leave the job paused, and pick it up tomorrow. Paused is better than deleted. You keep the draft.

This is where most solo owners get stuck. They re-run 4 times, edit the post each time, and lose 40 minutes to something a reconnect would have fixed. The rule exists to stop that. Twice is the limit. After twice, you check the connection, not the content.

The same rule applies to invoices and email. If an invoice reminder fails twice, check the payment connection, not the reminder text. If a newsletter send fails twice, check the list connection, not the subject line. The pattern is the same across every agent. Twice means connection. Once means retry.

The 2 minutes that keeps the feed from going quiet

The check is 2 minutes at night. Open the dashboard, look at today's runs, confirm 3 things per run: published, right channel, captions in. If everything passed, close the laptop. If one failed once, re-run. If one failed twice, check the connection.

That's the whole routine. It fits in the time it takes to make tea. It doesn't require you to understand the API or read a log file. It requires you to look at 3 statuses and act on one of them.

The customer win is the 11 hours. The failed post gets caught at 9 PM instead of 6 PM the next day. The feed doesn't go quiet. The client doesn't notice. The retainer doesn't quietly disappear because you looked like you stopped caring.

This is also where a crm with email marketing automation earns its place. The same dashboard that shows a failed post shows a failed reminder. One check covers both. You're not opening 4 tools to confirm 4 things. You're looking at one screen and confirming the runs that matter today.

If you add team members to your crm account later, the check scales. Each person sees the runs for their channels. The nightly check becomes a shared 2-minute habit instead of one person carrying it. That's how a solo operation grows without the owner becoming the alert system for everything.

Start tonight. Open the dashboard, find today's runs, check the 3 things. It takes 2 minutes. The alternative is finding out from a customer, and that conversation takes longer than 2 minutes every time.