What does data chaos cost? Usually more than any single line of your budget shows, because it is not a line of its own. It is spread across hours of manual work, decisions made on numbers that don't match, projects that come apart halfway through, and the valuation of the company. So before you approve another tool, another hire or another AI project, start your 2027 budget by putting a number on it.
This article walks through five places where data chaos hides and shows how to estimate each one from your own figures, without borrowing anyone else's statistics. By the end you will have a single sheet of paper you can put on the table during planning and ask: what are we actually paying for twice?
Your budget knows what Excel costs. It doesn't know what misalignment costs
A company grows. Person number fifteen joins, then person number forty. Each builds their own spreadsheet, their own dashboard, their own small tool. At some point everyone has different data and different numbers.
Picture a typical planning meeting in a company like that. This is an illustrative scenario, not a description of a specific project. Sales says September revenue was one figure. Finance has another. Marketing has a third, because it counts by a different date. The first fifteen minutes go on working out which one is real. Then someone says: "Let's check and come back to this next week." The decision is postponed, or it is made on whichever number suits someone best.
That quarter of an hour appears in no budget. Neither does the week of waiting, or the hours someone spent at night stitching a report together. All of it is Data Debt.
Data Debt is the sum of shortcuts, manual workarounds and inconsistent definitions a company takes on in its data in order to move faster, and which charge interest every month in people's time and in the quality of decisions.
You pay that interest out of the payroll budget and out of the leadership team's attention. Nobody books it as a data cost, so at planning time it doesn't exist. More on the concept itself: The debt you can't see on the balance sheet — until It's too late.
Why the 2027 budget starts with a tool instead of a diagnosis
Most planning looks the same. Someone submits a request: a new reporting system, a data warehouse, an AI project. The request has a price, a timeline and an expected outcome. It is easy to approve or reject, because it is concrete.
The problem is that a tool solves what it is shown. If sales and finance define "revenue" differently, a new system will display both definitions faster and more attractively. The misalignment stays; it just looks better. That is why a dashboard alone doesn't mean you control your data: You Have a Dashboard? That Doesn't Mean You Have Control Over Your Data.
The same goes for AI. A model fed inconsistent data returns inconsistent answers, only in a more confident tone. It won't fix the mess at the source. At best it will speed it up.
The justification for such a request usually sounds like this: "We'll save the team time," or "We'll get better visibility into the data." It sounds reasonable, but a year later you can't check it. How much time, whose, on what work? Without a baseline from today's situation, any result can be declared a success or a failure, depending on who is asking. That is why the starting number matters more than the request itself.
Code is a cost. The asset is the solution to a business problem, and to name it you first need to know where the problem really sits and what it costs. The order should be: business, decision, and only then the tool.
The goal is simple. The leadership team should look at one version of the numbers, a Single Source of Truth: one place, or one agreed set of definitions, that everyone draws the same figure from. Until that exists, every data investment decision starts with guessing. Why a board that receives three different numbers picks the convenient one: Single Source of Truth: Why a Board That Gets Three Different Numbers Picks the One It Likes.
Five places where data chaos costs you
Below are five places worth checking before you close the budget. For each one I show how to estimate it from your own data. Don't look for an exact number. A range is enough.
1. Hours of manual work
Who in the company stitches a report together from several sources every week or every month? Start with a list of people, not systems. Next to each name, write down how many hours a month that work takes, and multiply by the internal hourly rate your finance team already knows.
This is the first component of EBITDA Leak, the leakage of margin: a cost with no line of its own in the P&L, because it sits inside the salaries of people doing work that a Digital Worker, meaning an automated workflow, could do. The more expensive the specialist, the more expensive the leak. More: EBITDA Leak: definition and the 3 places it leaks most often and The cost of manual work: why "we'll manage a bit longer" is the most expensive sentence in the boardroom.
2. Decisions made on numbers that don't match
This is the hardest cost to capture, so estimate it indirectly. Go through the last quarter and answer three questions. How many decisions were postponed because "the numbers need checking"? How many were changed after it turned out the first number differed from the second? How many were made on a number chosen simply because someone had it at hand?
Don't turn the answers into currency amounts unless you have grounds to. The count of such situations, and what they concerned, is enough: advertising budget, hiring, pricing. One look at that list usually tells you more than a polished chart.
3. Dependence on one person
Ask one question: what happens if the person who knows how this metric is really calculated disappears for a week? Or leaves?
If the answer is "the report won't get done" or "nobody will know where this number comes from," you have a risk you can describe: how many reports, how many decisions, how many days of delay. More: [LINK: Why one analyst leaving can stop a company for a week — PL only for now].
4. Projects that come apart halfway
Go through the projects of the last twelve months, especially implementations and integrations. In how many of them was the scope changed because, midway through, it turned out the source data looked different from what had been assumed? Each such case is a cost that looked like a project cost but was in fact the cost of not knowing your own data.
Count them and write the number next to the projects planned for 2027. If the pattern repeats, the same scenario will hit the projects you are only now approving.
5. Valuation and due diligence
If a funding round, a sale of the company or an incoming investor is on the horizon, the buyer and their advisers will check one thing above all: do the company's reports reconcile, and can anyone explain why or why not? Data that cannot be explained becomes a bargaining chip against the price.
Even if you don't plan a transaction, ask yourself this once: if someone from outside walked in tomorrow and asked for a consistent picture of revenue and costs, how many days would it take to assemble it without errors? More on this channel: Data Debt and company valuation: how messy data lowers your price in M&A.
A budget with no line for data chaos will pay for it anyway: in other people's hours and in worse decisions.
How to estimate it before you approve the budget
The whole exercise fits on one sheet of paper and takes a few hours of one person's time from finance or operations.
- Draw a table with the five rows above.
- In each row, enter an estimate as a range: hours per month, number of decisions or projects, days of downtime.
- Next to each estimate, mark where it comes from: "measured," "estimated" or "don't know."
- Sort the rows from most expensive to least.
It helps if at least two people from different departments fill in the sheet together, for example someone from finance and someone from sales or marketing. If their estimates for the same row differ widely, that is already your first diagnosis: the company has no shared picture of its own costs. Don't settle the difference straight away. Write it down, because it shows where the numbers diverge.
Pay attention to the rows marked "don't know." They are not a missing result. They are the result: they show where the company can't see its own costs. Often that is where the most is waiting to be recovered, simply because nobody has looked yet.
What the sheet of paper won't show
This kind of exercise has limits. People who work in chaos stop noticing it. Each department knows its own piece, and nobody sees the whole map. Each also, quite naturally, defends its own version of the numbers.
That is why, where the stakes are higher, an external, fixed-scope data audit makes sense. Its purpose is not to build anything. It is a diagnosis: a map of the places where numbers diverge, a target picture of how information should flow, and a roadmap ordered by what costs you most. It ends with a report and a presentation for the leadership team. What happens next is the leadership's decision, and the roadmap can be implemented by your own team or by someone from outside.
This order has one advantage for the budget: the 2027 lines get a justification that comes from the company's own data, not from whatever happens to be fashionable.
What to do differently when planning 2027
- Add one line to the budget: the estimated cost of data misalignment. A range, not a single number, with a confidence label. It becomes the reference point for every investment request.
- Change the order. Diagnosis first, then the decision, and the tool last. Requests for systems, automation and AI should answer the question: which of the five places do they reduce?
- Name one owner of the numbers. Not for every report, but for the five to ten metrics the leadership team actually decides on. One definition, one source, one person accountable for it.
Data is not a cost to be cut, nor a fashionable project for the new year. It is the foundation on which every other budget decision is made, so it is worth knowing what its current state costs before you spend on anything else.
Want to go through this sheet together?
If you'd like to see what such a table would look like in your company, book a 15-minute call about your situation.