The Error You Never See Coming

And who checks the checkers?

Last week, ten orders didn’t make it into the ERP.

No errors, no warnings, no red flags. Just, gone. No one noticed for five days. And the only reason they did? Someone happened to run a manual reconciliation. The integration didn’t fail. It just never tried.

Welcome to the bit no one talks about: the blind spots between your platforms. The part where your ERP isn’t broken, but your data is quietly, invisibly wrong. Where your systems are integrated, but no one’s checking whether those integrations are doing their job.

Automation Isn’t Accuracy

It’s easy to assume automation means accuracy. That once your orders are flowing from Shopify into NetSuite, or your refunds are syncing back from Loop or Swap, the job’s done. But here’s the thing: integrations fail silently more often than you think. It’s not the errors that get you. It’s the absence of errors.

We see this all the time. Especially with brands running seasonal promos or flash sales. The higher the volume, the harder it is to spot when something goes missing. A 1% failure rate on a quiet Tuesday might be three orders. On launch day, it could be 300. And when your integrations aren’t telling you what they missed, the only way to catch it is the old-fashioned way: downloading two spreadsheets and cross-checking them line by line.

Manual Reconciliation: Still the Default

This isn’t theoretical. It’s still the default for most brands. Someone’s job, often a junior system admin on £25k a year, is to run manual checks across systems. Because there isn’t a dashboard. There isn’t a monitoring tool. There’s just a person with two tabs open, hoping today’s numbers match.

Why Dashboards Rarely Exist

And it’s not because people don’t care. It’s because these dashboards are hard to build. They’re custom, messy, and expensive. Tools like Groww or Mosaic promise out-of-the-box visibility, but at a certain scale, they break. They either can’t handle the data volume or can’t flex around the brand’s unique stack. What works for a smaller, slower-growth brand often buckles under the weight of one processing thousands of orders an hour.

There’s also a fundamental truth here: most brands don’t have in-house dev teams. Their integrations are outsourced, their systems team is one person deep, and their reporting gets built reactively when something goes wrong. It’s not a lack of will. It’s just the reality of where time and money get spent.

What Better Looks Like

So what should this look like instead? For starters, daily delta reports. One view of what Shopify says happened, one view of what NetSuite received, and a simple reconciliation of the two. If something’s missing, flag it. If it failed, retry it. If it still doesn’t go through, escalate it. That alone would prevent 90% of the issues we see.

Why No One’s Built It (Yet)

Ideally, this kind of validation shouldn't sit with the brand at all. But in practice, no integration layer offers this kind of monitoring out of the box. The idea of getting a daily digest of failed, missing, or retried syncs is compelling, but it's just that: an idea. No one's building it, and until someone does, the burden remains with the brand.

The irony, of course, is that most of these checks are simple. Comparing order counts between systems isn’t complex logic. Sending an email alert when something doesn’t match doesn’t require a full data team. But because these things sit in a no man’s land between finance, ops, and tech, they get missed. Everyone assumes someone else is watching.

The Real Risk: Assumed Accountability

And that’s the risk. It’s not about whether your ERP is good, or your 3PL is responsive, or your Shopify store is rock-solid. It’s about whether your ecosystem has anyone checking the checkers.

Because when no one does? Things don’t fail. They just disappear.

Struggling with this problem? Email me: luke@commercethinking.com