AI or Automation? The Real Cost of Admin Work in Marketing

Monday, 9:00 AM. I open Adform, Google Analytics, PowerBI, Spreadsheets, a Tracker, my suppliers' dedicated dashboards, and backoffice. And I start checking manually. How is programmatic doing? How do the markets and brands look? Any red flags? Does the marketing data match what's happening in backoffice? Do we have a "tracking issue" again? On top of that, I'm building weekly reports, manually copying scattered data from multiple platforms. Manual calculations, comments written from memory. Some meetings in between. Then copying into spreadsheets, presentations, sending out reports, discussing results… and the strategy part, what could actually be improved, gets the last half hour of the day.

It wasn't a bad day. It was an ordinary Monday, exactly the kind I had myself, and the kind I watched other marketing teams go through, in a company with many brands and many markets. Nothing dramatic. Just boring, annoying, tedious tasks, repeating every single Monday.

These days, when I talk to marketing teams about AI, I hear a similar version of the same story, just with a new tool inside it: instead of writing a comment for a chart by hand, they paste data into ChatGPT. Instead of formatting a slide manually, they ask AI for a suggestion. I'm not sure if that makes the process faster, but it definitely adds one more tool you need to log into.

AI vs. Automation

AI doesn't run on its own: you copy data, paste it into the model, wait for the output, fix the prompt because the result is too generic or doesn't sound like you, wait again. Every week you do this from scratch, the model doesn't remember last week's context, doesn't know your business rules, doesn't know iGaming, uses weird naming. And somewhere at the back of your head there's always the same question: can I even send this to ChatGPT? Sometimes another voice shows up too: why do I still have to do this by hand in the 21st century? Is my career going nowhere? Will I be copy-pasting until retirement? Can I even keep this up? (the longer the process, the more dramatic these thoughts get, just admit you've been mentally in that place too, no judgment).

AI-powered automation is a different thing.

You map the process once: every step, every data source, every decision rule ("if the budget changed by more than 20%, flag it as an anomaly"). The system connects to the platforms, pulls the right data, normalizes it into one format, and produces a ready output (a report, a comment, a presentation, a message to the team). Before you've had your coffee. It runs in the background, like air conditioning during a Malta heatwave, doing its job so you don't have to think about the temperature. This isn't a cosmetic difference. It's the difference between work you constantly have to manage, and work that is simply done.

What This Actually Costs. Measure It Yourself

Let's assume, conservatively, 3 hours a week collecting data and building reports. At an annual salary of around €50,000 for a digital marketing specialist in iGaming in Malta, with a standard working year, the hourly rate comes out to about €25.5. Three hours a week over 52 weeks is 156 hours a year, roughly €4,000 per person.

But this conservative assumption rarely reflects reality. I personally had 2-3 reporting meetings a week: weekly marketing meeting, bi-weekly per geo, monthly per geo, sometimes dedicated per-brand reporting, because every stakeholder group wanted its own format and its own period. It adds up: the realistic weekly cost is closer to 4.5 hours, not 3, almost €6,000 a year per person.

Multiply that by a team of ten people doing a similar or comparable process, and the conservative assumption already comes to almost €40,000 a year. The realistic assumption, based on a multi-stakeholder structure, comes to close to €60,000. That's the equivalent of one full-time position, spent entirely on collecting data and formatting presentations.

And it's still an underestimate. From my own experience and from watching my team, tedious and repetitive tasks always take longer than anyone is willing to admit. Not because people lie, but because boring work naturally invites distraction: did I put on the right music, do I have my coffee, let me just check one email, coming back to the same chart the next day because "I'll just fix it a bit more." Nobody reports that as part of reporting time, because admitting it sounds like admitting to being inefficient, not like admitting the task itself is demotivating by nature.

Before you decide whether automation is worth it, measure it yourself. Don't estimate. Use a stopwatch on your next report and write down the real time, including breaks and returning to the document at different points in the day. The result usually surprises people.

Costs You Won't Find in Any Spreadsheet

Hours multiplied by rate is only the starting point. Some costs are harder to calculate.

The first is lost revenue. Every hour spent assembling data is an hour not spent on something that actually generates money: testing a new creative, negotiating with a publisher, digging deeper into why one market converts worse than another. If just one hour a week, recovered and redirected toward optimization, lifts a channel's performance by even a few quality NDCs, the yearly effect often outweighs the cost of the hours saved on reporting.

The second is the new-initiative tax. A team fully absorbed in maintaining existing reporting pays this tax before it even starts working on anything new. Want to enter a new market? Someone has to build a separate report for it. Want to test a new channel? Someone has to research suppliers, review offers, get through the procurement process. Good ideas don't get rejected because they're bad, they get postponed because nobody has the bandwidth for them.

The third, the most human one, is the personal growth ceiling. A manager who spends 30-40% of their time on administrative work doesn't develop the skills they were actually hired for, neither their own nor their team's. The company pays a manager's salary for assistant-level work, and good people who see no growth path leave. In a market as small and competitive as Malta, that's a real, expensive risk.

The Pace of Company Growth

All of this comes down to one question: how fast can you grow your company if every new market, every new brand, every new initiative demands not creativity, but more administrative work?

A company without automated processes scales by adding people, which is expensive, slow, and linear. A company that has automated its admin work scales its existing team instead. The same person who used to handle one market can now oversee five, because the system does the tedious work for them. So in theory, you have more to do. In practice, you have more strategic tasks to figure out, because the tedious part just happens in the background.

Summary

Copy-paste is not work that inspires anyone, and nobody will miss it. Companies that still treat it as "just part of the job" aren't only losing hours, they're losing people who leave for places that use their time better. They're also losing momentum they won't get back by adding one more headcount to the same manual puzzle.

AI changed the rules of the game. AI-first companies don't win because they have better access to models. They win because they stopped paying people to assemble data, and started paying them to think.

I'm rooting for the companies that have already understood the potential hiding inside AI automation. This is genuinely disruptive technology, and used the right way, it can let smaller companies sweep the old, fat cats off the board, the ones that have been winning on size for years, not on their efficiency. And as someone who has always rooted for the underdog, this is going to be a fascinating thing to watch.