01
Workaround payroll.
The hours your business units burn working around data products that do not fit how they operate. I call it FTE Debt. It appears on no cloud invoice, yet it consumes real salaries every week.
Your data budget grew again this year. Nobody can tell you what it returned.
Most companies know what data costs them. Almost none know what it costs everyone else.
Your finance team rebuilds a report by hand every Friday. Your operations lead double-checks a number before trusting it. A decision waits two days for a manual validation. Those hours are on your payroll, they appear on no invoice, and in one company they added up to €930,000 a year.
I find that number, and hand back the list of what to stop paying for. Your team runs the review after I leave.
2 min · who this is forLior Barak
€930,000
a year in business capacity consumed by data workarounds, in a 120-person company
€611,000
of it leaving the business entirely: hours, unused tooling, and preventable errors
3 weeks
to measure it, with the first cancellations executed before the engagement ends
From client engagements. Anonymized. Measured against the client's own payroll rates and cloud bills.
Built data functions from the inside at
Zalandoidealo
Author of What Data Really CostsStart here
The Data Cost Baseline
Three weeks · €12,000 fixed50% to start, 50% at the readout.
Most companies can tell you what their data team costs. Almost none can tell you what it costs them to run data. In three weeks you get that number, and the list of what to stop.
What’s included:
- →The baseline: total annual cost of running data, split into three layers: infrastructure, the hours absorbed across the business preparing and reconciling data, and the carrying cost of everything already built
- →Cost per product for the top 20 items in the portfolio
- →A ranked keep / kill list with the annual recovery attached to each candidate
- →The template and method, so the team can rebuild the number next quarter without me
- →A 60-minute readout with the CFO and the data lead in the same room
What you provide:
Access to six to eight people for 45 minutes each, cloud and tooling invoices, and a headcount list.
The unvoiced margin leak
Every data product your company ever shipped is still on the payroll.
Compute bills, storage maintenance, engineering support hours, and frontline workarounds: every product your company ever deployed keeps costing something. The ones nobody looks at anymore do not stop billing you. They just stop appearing in active conversations.
This is the Data Zombie Tax: dead products quietly eating budget, live products costing far more than they return, and a core team too busy keeping legacy pipelines alive to build what should actually replace them. It compounds silently, which is why it is rarely named during a budget review.
The tax arrives through three doors:
02
Carrying weight.
The full cost of keeping every data asset alive: infrastructure, incidents, and the engineering hours lost to maintenance, whether the asset moves margins or not.
03
Decisions that wait.
The data exists, but a strategic decision waits days while someone validates the number by hand. Waiting has a payroll cost and a market cost. Neither appears on any invoice.
The calculator below gives you the first rough number for what you're paying.
None of this shows up as a line item. Here is roughly what it is costing you.
Three inputs. Adjust to your situation. Most leaders have never had this number before.
People touching data
Anyone in the business who builds, uses, fixes, or waits for data. Not just the data team.
30
Time lost to friction
Share of their time spent fixing, waiting, reconciling, or working around data rather than using it.
35%
Avg annual cost per person
Fully loaded: salary, benefits, employer taxes, tools. Not just gross salary.
€85k
~10.5 FTEs
locked in friction every month
€893k
estimated annual cost of wasted capacity
Directional estimate. We measure yours precisely in the first engagement.
One engagement, broken down
€930,000 a year. Here is where it sat.
120 employees. Eight of them in the data team. This is the arithmetic, because a number this size is worth nothing if you can't check it.
€611,000 was leaving the business
Stoppable. Some of it within weeks.
- €352,000 · business hours
- Thirty-six people across finance, marketing, operations and product, averaging four hours a week preparing, checking and reconciling numbers before they could use them. 139 hours a week, 6,400 hours a year, at loaded payroll rates.
- €217,000 · stack
- Seats licensed and never activated. Two tools doing the same job. Compute running pipelines that fed products with no users.
- €42,000 · errors
- Mistakes that reached a customer or a report, and the rework that followed.
€319,000 was aimed at the wrong thing
Not recoverable. Redirectable.
38%
of the data team's time
They had hired eight people to build new capability. Thirty-eight percent of that team's time went to maintaining products with no business decision attached, or answering requests nobody acted on. Fewer than two of the eight were building anything new.
That company was not overpaying for its data team. It was under-receiving from it. The declared data budget was €1.31 million, and €394,000 of the €930,000 never appeared in a data line at all, because it was spent by people whose budgets don't mention data. Half of it was invisible. The other half was approved in plain sight, and nobody asked what it returned.
One engagement. Anonymized. Measured against the client's own payroll rates and cloud bills.
See how the same number gets found in your business →Data portfolios are corporate capital. It is time they are managed like it.
You run capital allocation conversations every week: with product, with marketing, with sales. Each one starts with what a bet returns and what it costs to maintain. Then the data budget arrives, and the conversation shifts to uptime, pipelines, and tooling vendors.
The fix is a flipped sequence: value first, team capability second, technology last. The conversation about tooling does not disappear; it gets easier, because by the time it happens, everyone knows exactly what the work is for.
Value first: what decision does this serve and what is it worth.
Capability second: can we actually capture it.
Technology last: the shortest conversation of the three.
You don't need to speak engineering to lead this. You need the correct sequence.
See how the method works →After the numberBuilding or Migrating Running and Tracking Institutional Transformation
Where this usually goes next.
The investigation ends with a number and a recommendation. Sometimes the recommendation is one of these. Sometimes it's "take the plan and run it yourselves": that outcome is written into the engagement.
The Prevention Blueprint
You are about to commit capital to an infrastructure migration, a stack overhaul, or an AI-agent rollout. Before your engineers write a single line of code, we install the financial rules that govern what earns a place in the new architecture.
+ Explore the program →One month to set the rules · then quarterly review · CEO + CFO
What stays when we are done:
- →A clear financial bar every data product must pass before it enters the new stack.
- →Keep-or-kill decisions governed by structural rules, not political capital.
- →An honest assessment of what your team can carry before you add technical weight.
The Hidden Cost Recovery
Your data budget grows every quarter, yet frontline teams are still fixing numbers by hand in spreadsheets. Those manual hours run on your payroll. I locate them, calculate them, and isolate the exact leakage.
+ Explore the program →Follows the investigation · monthly review your team takes over
What stays when we are done:
- →A precise euro figure for the hours your business teams spend compensating for data gaps.
- →A monthly operational review your own people keep running after the engagement ends.
- →An actionable roadmap with the first three months fully mapped for your team to execute.
The Data Capital Program
Your data lead reports in sprint velocity and engineering metrics. Your CFO tracks margin contribution and capital efficiency. Neither can make an allocation decision together. This program installs a single, unified financial language.
+ Explore the program →Six to eight months, phase-gated · Executive board
What stays when we are done:
- →One set of numbers connecting boardroom strategy directly to weekly engineering tasks.
- →A live view of what every data asset costs to carry against its explicit business return.
- →Structured validation cycles that test value before major engineering capital is deployed.
Not sure any of this applies?
30 minutes. No slides. No sales pitch.
Bring the part of your data spend you can't explain. Thirty minutes. We'll work out whether there's a number worth finding, or whether the timing is wrong, which is a real answer and one I give often.
Book an alignment call
Founder statement
"Most data functions were never asked the one question that decides everything else: what should we stop paying for? Engineering was hired to build. Finance was hired to total the bill. Nobody was hired to sit in between and ask what's actually earning its keep."
Fifteen years running data teams inside scaled organizations like Zalando and idealo, and the same gap kept showing up: engineering owns delivery, finance owns the total, and the space in between, where each product either earns its keep or quietly stops, belonged to no one.
Impact Operations is what I built to close it: a method that gives the CEO and CFO the numbers and the rhythm to run data as the capital investment it already is. I sit in your boardroom until your team runs it without me.
The Reality Check
See what the workarounds cost you.
Answer a few questions about how your teams handle numbers today. You get a figure in payroll terms, the hours behind it, and one sentence you can say in your next leadership meeting.
About 4 minutesNo email needed to see your numberAnonymous
Start the Reality Check →Would rather just talk? or message me on LinkedIn.