While most people are discussing winning combos, creatives, and “where to drive traffic right now,” strong teams have long been playing on an entirely different level.
Today, media buying isn't just about purchasing traffic. It's about:
- operational processes
- decision-making speed
- automation
- and real-time financial control
Operations are exactly that invisible layer that either scales a team or kills it in the long run.
We spoke with Djan from Fellows Team — a company of ~120 people that systematically builds processes, cuts costs, and actively integrates AI into its operations.
Team Structure and Its Evolution
— How many people are currently in the Fellows team, and how much has the team structure changed over the last 2-3 years specifically regarding operational roles?
As of today, Fellows has about 120 employees.
Over the past 2 years, many departments have evolved and become separate business units that try to act exclusively in the best interests of their specific department and team. The C-level/operational manager positions were filled by people we raised in-house, who are deeply immersed in the internal processes of their specific departments.
Operational Expenses and Optimization
— What percentage of the team's total budget currently goes toward operations (support salaries, tools, software, processes, compliance, excluding direct traffic and creatives)?
If we are talking about the overall budget (meaning the infrastructure and ops budget + advertising budget), currently less than 10% of the total budget goes there. And we are constantly trying to reduce operational costs using AI agents.
— By how many dollars or percentage points per year have you managed to reduce operational costs through well-built processes over the last 12-24 months?
Over the last 2 years, we started spending on average 5% less on infrastructure annually. This was exactly when the AI boom hit the entire market. In the SEO department, we managed to cut expenses by 30% in a year, saving 50k, simply by using custom-built solutions and automated scripts.
Automation and Key Processes
— Name the automated process that brought in the most money through operational optimization?
The automation of the finance department. When you can see the numbers live, you always know exactly where to start optimizing expenses. Automating the finance department is all about eliminating as much manual work as possible.
— Which processes have you fully or almost fully automated (Make, Zapier, n8n, scripts, bots)? For each — how many hours/dollars do you save per month?
Fully automated: creatives distribution, caps, UGC content creation, white pages, farming, the finance department, and the HR department. Each department has its own scripts and bots that speed up the workflow. The savings come from needing fewer employees, so it translates to tens of thousands of dollars saved on payroll.
Control, Alerts, and Audits
— Do you have automatic or semi-automatic alerts and protocols that allow a media buyer not to sit at the screen 24/7 without losing accounts?
We have alerts that help the buying team avoid overspending their caps, which is often a very common mistake.
We also have alerts to monitor the uptime of links/domains. The moment traffic throughput drops close to 0, like a zero conversion rate, the alert immediately notifies the Buyer and the tech department that there is an issue.
— How often and through what procedure do you audit "dead" expenses on proxies, antidetect browsers, accounts, bots, etc.?
Domain and proxy audits are done every month. Bots and accounts are audited less frequently, about once every six months. Each department has its own Head who is responsible for this.
Farming, Warm-up, and Anti-ban
— How is the process of managing account farming, warm-ups, cycling, and rotation structured right now to minimize bans and budget losses?
Our farming is manual and only in the Google department. In this specific aspect, warm-ups and the farming process are still done by hand by a real person. This is because AI agents create an obvious action pattern that Google detects very quickly. However, even here, we couldn't do without automation for accepting 3DS payments for warm-ups, auto-logins, and receiving SMS for registrations.
The Creative Process
— Do you have automatic or semi-automatic creative checks before launching a campaign (AI, scripts, checklists in trackers / Notion / Make)?
Checking a creative is ultimately always proven only in practice. We do have an automated tracker that helps the buying and design departments communicate quickly without messengers, which saves everyone a lot of time.
Onboarding and Knowledge Base
— What does the onboarding process for a new media buyer, creative designer, assistant, or farmer look like right now?
We conduct onboarding as quickly as possible, but constructively. New employees have access to a knowledge base and get prompt answers from team leads, so our adaptation period takes no more than a month. Training also happens gradually and natively.
— Do you have an up-to-date knowledge base, playbooks, videos, or templates that are actually used? How do you ensure they don’t become outdated?
We have a welcome book that helps introduce a buyer to the team members right at the start, so they have an understanding of who to contact for specific issues.
There is also a knowledge base tailored to the position of the person joining the team. Mostly, this is in Notion (the main thing is to convey the importance of studying this document to save the time of other employees the new hire might approach).
But the most crucial aspect is the onboarding with the Head of the department, where the most important information and workflows are explained in a personal conversation. Usually, this is recorded so the new person can refer back to it if they suddenly forget something.
Once every six months, we check the relevance of the knowledge base, after which we usually edit or remove unnecessary information.
Contractors and Outsourcing
— Which processes/roles have you fully or partially outsourced to virtual assistants/freelancers (farming, creatives, analytics, support, etc.)?
Outsourcing is partially used for app development (this helps reduce the risk of getting bans for non-unique code). Every other aspect is strictly in-house, as we are always certain that the work will be done with high quality.
— How is the task acceptance and payment system structured with contractors to avoid overpaying or losing money due to errors/unfinished work?
There is a strictly built A-to-Z flow for managers working with external contractors. It’s important that the manager addresses standard objections right at the task assignment stage to save time, and that communication is clearly established with other departments responsible for checking the quality of the completed work. Having multiple verification stages helps avoid mistakes, but it's important to maintain speed so time isn't wasted.
The Hardest Implementations
— Which process took the longest and was the most painful to implement in the team, and why? How much savings did it ultimately bring?
The longest and most painful process to build was the communication between the finance department and the media buying team. The issue was insufficient automation and an excessively large amount of manual work in the finance department. "Not spending more means you saved money." Therefore, you can estimate the savings at tens of thousands of dollars monthly.
Market Mistakes
— What are the 3 most expensive operational mistakes you most often see in mid-to-large media buying teams in 2026?
1 - Appointing people to C-level roles without experience or with insufficient experience.
2 - A slow-moving finance department.
3 - Improper allocation of infrastructure resources to the buying team.
The Future of Operations
— In your opinion, which processes/systems will become a "must-have" in the operations of media buying teams in 12-18 months, even though almost no one is doing them well right now?
The absolute must-have is the correct implementation of AI agents into operational processes. If you don't have this, your team will always be a few steps behind competitors who will do everything faster.
Conclusion
Honestly, after this conversation, you start looking at the entire media buying industry a bit differently.
Everyone is used to discussing creatives, winning combos, and where it’s best to drive traffic. But in reality, that is no longer the main thing. In the long run, the winner is the one who knows how to systematically work with numbers, processes, and people.
You can clearly see this with Fellows: they don't try to make things "perfect," they are just constantly tweaking something — automating, simplifying, accelerating. Where manual work can be eliminated, they eliminate it. Where money can be saved, they calculate and cut.
And at a certain point, this starts to show results. Not a sudden spike, but a compounding effect — when the team simply becomes faster and more stable than the rest.
Perhaps the main takeaway here is simple: if your operations are lagging, no amount of aggressive media buying will cover it up. But if they are well-built, even less-than-perfect decisions can bring in money in the long run.
- You can find more interesting interviews in our Interviews section — HERE
