Utbenat: Data silos

Introduction

Sebastian:
Imagine you're the project manager for a gigantic bridge. You're connecting two cities across a foggy bay. To move fast, you've split the work. Team East is building from one side, Team West from the other. You have the best engineers in the world. You have the most expensive steel and the most advanced tools. Each team works around the clock, with precision down to the millimetre.

Three years have passed. Today is the big day. The two spans are finally going to meet in the middle. You're standing out there on the edge. The fog slowly lifts. You squint towards the other side to see your colleagues on Team West. And then... your heart sinks like a stone. As the fog clears, you see that Team West's anchor point is several decimetres higher than yours. And at least as far off to the left. You didn't miss because you're bad builders. You missed because nobody thought to shout across the ravine and check the coordinates.

Vad is a data silo?

Sebastian:
Today we're going to talk about that missed connection out in the fog. We're going to talk about data silos. A concept that's often described as a technical problem, but that's really about why customer satisfaction takes a hit while the marketing department is celebrating record sales. It's the invisible walls that stop us from meeting in the middle.

Data silos show up as us seemingly speaking different "languages". Because this isn't just about interpreting data differently. It's about why we're looking at different pictures from the very start.

A data silo appears when one part of the company holds data that nobody else has access to, or even knows exists. We often blame the technology: "the systems don't talk to each other". But the root problem is usually that what we want to do with the data shapes what it looks like.

Take production at an industrial company. When they report what they've done, "number of tonnes" might be the only measure that counts. That's their truth. But when that data reaches Eva in logistics... well, then "one tonne" is worthless. She needs to know exactly which items they are and where they should be stored. Our different needs create different dialects of data. And when those differences are then cast in concrete in our IT systems... well, then we're left with two bridgeheads that struggle to meet.

Sub-optimisation: when everyone does the right thing and it still goes wrong

Sebastian:
So what actually happens when we're stuck in our silos? The biggest problem is something we call sub-optimisation. Data silos very easily lead to each department optimising its own little corner of the world to perfection, at the expense of the whole.

Let's take an example many online retailers will recognise. The marketing department has a goal: increase the average order value. They put together a "Cosy Autumn Bundle" at a good price. It contains a soft throw with a good margin and a really stylish floor candlestick that's been trending in the interior design magazines. In their system it looks like the perfect combo. And customers love it. The orders pour in.

But... marketing is sitting in its silo. They don't have access to logistics' data on packaging and shipping costs. When the orders reach the warehouse, the problem appears. The throw is vacuum-packed and compact, but the candlestick is long and fragile. It comes in its own box from the supplier, but the throw doesn't fit in there. So to pack them together, the warehouse has to use a box that fits both, and the one at hand is far too big. So they have to fill it with loads of packing material so the candlestick doesn't rattle around. The result? The parcel becomes expensive to pack and deliver.

Marketing has optimised sales and is bringing in money, but the logistics cost of shipping air and bubble wrap eats up a large part of the profit margin. What looked like a victory parade ends up, on the bottom line, hurting the margins. And it didn't happen because anyone did anything wrong, but because the data on product dimensions and shipping costs wasn't available where the Cosy Autumn Bundle was designed.

The customer service team nobody told

Sebastian:
Then there's the human cost, which often lands on customer service. The head of marketing proudly shows the graph: "We've sold 500 units of the new product!" Everyone applauds. But at the other end of the table, the head of customer service sits looking pale.

Why? Because nobody told customer service the campaign was about to launch. The information stayed in the marketing department. The customer service system hasn't been updated with the new product. They have no manuals, no answers to common questions. So when customers call in wondering how the gadget works, they're met by staff who have no idea. "Hold on, let me check...", they say, and start searching. Waiting times go up. Customers get frustrated. The brand takes a hit. This is the classic sign of data silos: the left hand doesn't know what the right hand is doing, and it's the customer who pays the price.

Duplicate work and the Excel war

Sebastian:
Finally, there's something that quietly eats resources: duplicate work. It happens when the same data is maintained in several places, without anyone really knowing which version is the right one.

Take a product list. Marketing has one version in its campaign tool. The warehouse has a version in its system. Finance has a third in the invoicing system. Everyone updates their own, in good faith. But after a couple of months they've drifted apart. Someone has renamed an item, someone else has changed a price, a third person has removed a product that's still sitting in the warehouse.

At first it doesn't show up in the management meeting. It shows up when a customer gets an incorrect invoice. When the warehouse picks the wrong item. When three people spend half a day a week manually reconciling the lists against each other. It's not particularly dramatic, but it's money leaking out through the hole in the silo wall, every week.

And it's exactly these small, quiet drifts that eventually blow up in the management team. Three versions of "the truth" on the table, and suddenly the meeting is no longer a business meeting. It turns into a classic Excel war. 45 minutes are spent arguing over which number is right. It kills the energy in the room, and it's a direct risk to the company, because decisions are made on a foundation that nobody really trusts. So what do we do about it?

Build windows instead of tearing down walls

Sebastian:
Many people think the only solution to the silo problem is to buy a gigantic system that handles everything for everyone. A central data platform. But that's an enormous investment that often takes a very long time. Instead of trying to tear down every silo at once, we believe it's about starting to build windows and increasing transparency. We shouldn't remove the expertise in the departments. We just need to make sure information can travel between us, and that we understand each other's needs.

Picture a submarine. It's divided into watertight compartments for a vital reason. If there's a leak in the bow, the whole vessel shouldn't sink straight away. That's exactly how your departments work. Silos are really those safe rooms where the experts can focus on their work without being disturbed by all the noise from the rest of the boat.

So the problem isn't the walls themselves. The problem is that there's no intercom system between the rooms.

Imagine the captain up on the bridge orders "full speed ahead". But down in the engine room, the technicians see the gauges starting to glow red. The engine is getting too hot, it's about to overheat under the load. If they have no way to warn the bridge, the captain keeps pushing, right until the engine seizes and the whole submarine is left dead in the water. Sometimes that's a deliberate risk. But very often a different, and better, decision could have been made if the right information had just got through.

So breaking down silos isn't about tearing out the steel walls and turning the submarine into an open-plan space. That would be lethal. It's about building communication channels. Making sure the person setting the speed and the person seeing the heat in the engine can talk to each other. Then you can make the right decisions to go the whole distance.

Three steps to take on Monday

Sebastian:
Here's what you do on Monday. Three simple steps.

1. Pick a department you suspect is holding those puzzle pieces you often find yourself missing.
2. Invite a colleague there for a coffee and ask a simple, sincere question: "What do you see in your data that I'm missing?" And: "Is there anything in my world that would make your day easier?"
3. Find a single data point that increases visibility and helps you see more of the bigger picture, and make sure it's shared and woven into the process. It can be as simple as the previous day's order volume, or whether the system is online.

This is your first window in the silo wall. It doesn't take any big IT investments, just a decision to start syncing your blueprints. Because in the end, this isn't about rebuilding your entire IT architecture. It's about realising that you're actually running the same relay race.

Summary

Sebastian:
To sum up: data silos appear because we have different needs, but they become dangerous when we lose sight of the bigger picture. The solution isn't to tear down the walls, but to create communication between the rooms. Build your communication channels, one coffee at a time.

Thanks for listening to today's episode of Utbenat! If you want to talk more about today's topic, or just talk data, get in touch. You'll find us on LinkedIn. Or contact us through our website, fiwe.com. Talk soon, and take care out there!

You've been listening to Syntra, a podcast from Fiwe.