Who’s taking care of your media buying while you’re sleeping?
Opinion
From hackers to human error, the possibility that something goes wrong with your client’s media spend is the stuff of nightmares. The CEO of adnomaly recounts what happened to him.
There’s a particular kind of stress that comes with media buying. Money moves fast, platforms don’t sleep, and those pesky little errors rarely wait until office hours to show themselves.
Most people in adtech know that feeling, whether they talk about it openly or not. The uncomfortable question is simple: when your team is offline, on holiday, or just asleep, who’s actually watching the spend?
The dreaded 10 pm phone call
One evening, while I was still working in paid social, a colleague called me late, saying something was very wrong. We found eight campaigns in a client’s account, each carrying a daily budget of €750,000. Nobody on the team had created them, the creative was trying to sell products, and the whole thing made no sense.
When we looked more closely, the details were all wrong in a way that should have been obvious much earlier. We didn’t use daily budgets like that; we’d never advertised in the US from that account, and the page ID was unknown.
We got the campaigns shut down, but it still took about an hour to reach platform support and stop the damage. That was the moment I started asking myself why there was no system in place to spot something so obviously unusual.
These attacks often happen at night, during holidays, or over long weekends, because that’s when people are least likely to be checking. For a big advertiser, it’s serious enough. But for a smaller business without a strong platform contact, it can be brutal because money can disappear while they wait for a support ticket to be addressed.
The bigger leak is usually much less dramatic
The truth is that hacking is only one part of the story.
The more common problem is much more ordinary, and that’s exactly why it’s so expensive.
I had a colleague once copy €50,000 from a media plan into TikTok, but despite the normal “four-eyes check”, a formatting issue turned that into €5m. It was a Friday afternoon, and by Monday the platform had already spent €1m. It was a human mistake, in a system that gave nobody much chance to catch it in time.
I don’t tell that story to shame anyone. I tell it because it’s still a very human problem, and media buying is a very human job carried out in complicated systems, under pressure, at speed.
People are overworked, often under-supported, and still expected to rely on manual checks, buddy systems, sign-off PDFs, and pure concentration as the last line of defence. In live media, that’s asking for trouble.
The tricky part is that the market doesn’t always see the full scale of it. Agencies can swallow part of the loss, or ask publishers for make-goods, so they’re only passing along some of the pain to clients. This means the brands funding the media don’t always know how much waste has been absorbed behind the scenes. We talk a lot about fraud, as indeed we should, but preventable operational waste is still quietly leaking money every day.
The basics still need guarding
A lot of people want to jump straight to the next shiny thing. They talk about AI, automation, and agentic workflows as if the old operational problems have quietly gone away. They haven’t.
In many cases, the same mistakes will still occur in an agentic workflow, but they’ll happen faster, across more systems, and with less human oversight. That could be a typo, a geo mismatch, a missing end date, or one tool passing the wrong definition to another. The more automated workflows become, the more useful good guardrails become too.
Of course, campaign managers know their own plans better than anyone else, but an automated system can spot something that looks off, pause it, and ask, “Did you mean to do that?”
And importantly, this kind of system helps everyone, not just the person entering the numbers. Junior buyers get a safety net, team leads don’t have to spend their evenings checking every setup by hand, and CFOs get fewer nasty surprises.
One client contact at a UK agency told us it was the first time she’d enjoyed Christmas in 10 years, because she knew there was still something watching the account while she was off.
Media buying will probably never be stress-free, and I don’t think anyone expects it to be. But there’s a big difference between healthy pressure and avoidable panic. I started building this type of technology because I had experienced that panic myself, and I don’t think it should be accepted as normal.
In 2026, QA and warding against hacked ad accounts should be automated: people should be able to focus on strategy, clients, and better work, rather than spending too much time on double and triple checks like it’s 1999, and simply ‘hoping’ that the ad accounts are safe at night. Because no one should lie awake wondering whether a typo or a hacker has been spending their budget since midnight.
Max von Weber is the Co-Founder and CEO of adnomaly