The Hidden Cost of Internal Data Silos

A dark, gritty, close-up photograph of a messy wooden desk with trophies, stained papers, a dirty coffee mug, and frayed cables scattered around, emphasizing failure and disorganization.

Rex Manning Day: The Danger of Marketing for Applause Instead of Results

June 17, 2026

July 22, 2026

Tags

A man in a rugged jacket inspects a broken piece of the Berlin Wall at night.

The only piece of the Berlin Wall I have ever seen is in Missouri.

I have wanted to stand at the real thing for years. So far the closest I have managed is a run of it on a college lawn in Fulton, on the spot where Churchill named the Iron Curtain in 1946. Someone bolted the sections upright and turned them into a sculpture. A wall kept intact for the express purpose of remembering the day it stopped working.

What gets me about that wall was never that it fell. It is how much work it took to build.

Rerouted trains. Bricked-up stations. Streets that had been one street, split down the middle and patrolled around the clock. A vast, deliberate, expensive effort to hold apart things that had every reason to be together. Nobody divides a city by accident. It takes budget, planning, and a great many people agreeing not to ask why.

I think about that more than I would like to admit when I look at how most companies run their marketing.

Everyone in my corner of the internet is worked up about the walls somebody else built. Meta owns your audience. Google decides who gets seen. Now the model answers the question before anyone reaches your site at all, and the new anxiety is whether you get quoted before you get forgotten. Fair enough. Most of it is true, and I have written my share of it.

But the wall that is quietly costing you money is rarely the one Meta built. It is the one running straight down the middle of your own operation.

Here is the shape of it. Your email platform knows who opened, clicked, bought, and went quiet. Your site analytics knows what those same people read before they drifted off. Your ad account knows what it paid to bring them in. Three systems, three views of the same human being, and in most companies they are not on speaking terms. Different logins. Different owners. Different meetings. A checkpoint at every crossing, staffed by people who never compare notes.

This is not a soft problem. It has a price, and people have gone to the trouble of measuring it. Gartner puts the cost of poor, disconnected data at roughly thirteen million dollars a year for the average company. The research firm IDC has estimated that silos and the inefficiency around them can drain a fifth to nearly a third of a company’s annual revenue. You can argue with any single figure. The direction is not in dispute, and it points the same way every time.

Part of what makes it dangerous is that the people in charge cannot see it. In one survey of North American CMOs, fewer than a third said they had real confidence in their own data, and two-thirds named siloed data as their single biggest obstacle. These are the people signing off on the budget. They are steering with instruments they do not trust, and most of them know it.

The clearest place to watch the wall cost you money is paid acquisition.

Meta will happily keep serving acquisition ads to someone who already bought from you, unless you tell it not to. And the usual way you tell it, an uploaded list of your own customers, is only ever as current as your last export. Your customer base grows every day. That list gets refreshed, if you are diligent, once a week. Picture the gap. Someone buys from you at ten on a Friday morning. Your exclusion list does not update until Monday. For the whole weekend you are paying, across every channel, to chase a sale you already closed. Multiply that by every customer and every product line and the leak stops looking like a rounding error.

The spend goes out under the word “prospecting,” which is the tell. You are prospecting for people you already own.

The maddening part is that it hides. Platform reporting cannot tell the difference between an existing customer who saw your acquisition ad and shrugged, and a genuine stranger who did the same. Both land in the same column. The waste is invisible unless you go looking for it on purpose. There is a second cost that never shows up in the media report at all: the loyal customer who gets served a new-customer discount they were never offered, and quietly notes that the brand rewards strangers better than regulars.

When agencies do go looking, the numbers are not small. Account audits routinely find that a large share of “new customer” budget, in some cases close to half, is flowing to people the brand had already earned. Tighten the exclusions and connect the purchase data properly, and the recovered spend tends to land somewhere between fifteen and twenty-five percent of the paid budget, redeployed toward actual strangers. That is not a growth hack. That is a company that stopped paying twice to reach the same person, because it finally let two of its own systems talk.

Nobody built that wall on purpose. That is the thing worth sitting with.

It builds up a piece at a time. You buy a tool one year to solve a real problem. You hire a team the next, and they choose their own tool for their own real problem. You bring in an agency to own the thing neither of them wanted. Every one of those decisions was reasonable in the room where it was made. The wall is just what is left standing in the gaps between them. When Salesforce surveyed IT leaders, roughly seven in ten described their own systems as overly interdependent, and eight in ten said their silos were actively getting in the way of what they were trying to do. These are not careless people. They made a hundred sensible local choices that added up to a border.

So the honest version is uncomfortable. You did not get walled in. You walled yourself in, one reasonable purchase order at a time.

Which is also, oddly, the good news.

The wall in Berlin stood for twenty-eight years. It came down in a night. Nobody dismantled it slowly and carefully; the logic of the thing simply collapsed the moment enough people treated it as optional and walked through. State-backed walls, with an army behind them, can end that fast. The wall between your email data and your ad account has no army behind it. It is held up by nothing but org charts and the fact that no one has questioned it lately.

Walls like that do not need a demolition project. They need someone with the standing to say the checkpoint is closing, and the willingness to be the one who says it.

The section in Fulton is just concrete on a lawn now. At some point someone decided it was worth more as a thing people walk through than a thing that keeps them apart, and shipped it across an ocean to make the point. The walls inside your own business are usually worth more down than up too. The only hard part is being the person who notices you built them, and then does something other than route around them for another year.

If your data lives in three rooms that never speak, you already know the feeling. If you have watched this cost real money somewhere, I would be curious where it broke.

I write these most weeks, if you want to stay close to the thinking.

Sources