Automating repetitive tasks
A report assembled by hand every week, the same data retyped between two systems, the same quote prepared from scratch. If a task has a clear input and a clear result, it can usually be automated.
Automation for companies
I automate repetitive work in companies.
If someone at your company copies data between systems, assembles the same reports every week or watches numbers by hand, I can simplify that process and automate it.
19 years of experience in online advertising and working with data. Today I am responsible for programmatic at Grupa Wirtualna Polska.
What I help with
Three things I most often take off people's hands
I build on approaches proven in practice. Every project is built separately, around the client's own process and data.
Watching the numbers nobody has time to watch
A daily check on the result — from the total down to individual items. It finds a problem hidden by a good overall number, tells it apart from ordinary variation and estimates what it costs. It speaks up only when the data genuinely calls for it.
Answers to questions about data
Instead of waiting for the one person who knows where to look: a question asked in plain language, and an answer that comes back with its source and the definition of the metric used.
I reach for AI when something has to be understood from text, judged from data or drafted. When simpler automation is enough, I don't add AI for its own sake.
More examples outside advertising — automating repetitive work →
How we start
We can start in three ways
We agree the scope, the result and the deadline up front. Every way in has a point at which you can see whether it works.
-
1
Process audit usually 5 working days
I take the process apart: what happens today, who does it, how long it takes and what can be automated. It ends with a document containing a recommendation and an estimate of whether it pays off. If automation doesn't pay off, that is what I'll write.
-
2
Pilot usually 2–3 weeks
One chosen part, built to the end and running on real data. Something you can try and assess before anyone commits to a bigger scope.
-
3
Care and development billed monthly
Once the tool is running: maintenance, reacting to changes in the data sources and adding the next pieces. A contract measured in months, not years.
Before we start, I also check that the project does not conflict with my work at Grupa Wirtualna Polska.
Projects
What is already running in production
Three systems I built myself, running every day on real data. They were built for online advertising because that is where I work — the mechanism is the same wherever the numbers come from. No revenue figures: those are not mine to publish.
Monitoring for silent failures in production, daily
Problem
A failure on one site or with one partner disappears into the portfolio total. The overall number looks normal while the money leaks for weeks before anyone notices.
What I built
A daily process that checks revenue and traffic for the whole portfolio, then drills down to individual sites, partners and ad placements. It tells a failure apart from an ordinary drop in traffic, pinpoints where the problem is and estimates the loss. It raises an alert only when the data genuinely calls for one.
monitoring · failure detection · loss estimate
Company data agent in production, on demand
Problem
Answering a simple question about a result could take hours or days, because you had to know which system to look in and which metric was the right one.
What I built
A documented map of company data joined to a tool that answers questions asked in plain language. Before calculating a result it checks the live system to confirm the data is where it should be. The answer comes back in minutes, with its source.
knowledge base · databases · validation
Automated A/B test analysis panel in production, nightly
Problem
Evaluating a single test meant hours of assembling spreadsheets by hand — so tests were reviewed infrequently, often late, and some not at all.
What I built
An overnight process that collects results from every revenue source, calculates statistical significance and shows the potential financial loss. Dozens of tests are checked every day, and the report is ready in the morning.
A/B testing · statistics · automated report
Other projects
Finding B2B customers — a process that collects companies and decision-makers from a chosen market, verifies the contact details, assesses how well each one matches the offer and writes everything into a structured spreadsheet. I use it to prepare contacts ahead of industry events across several European countries.
Experience
Nineteen years in online advertising — from running campaigns to leading programmatic monetization
I came up from the operational end of this business: getting campaigns live on sites, checking creatives and supporting publishers. So when I propose a change to someone's process today, I know what it looks like from the desk of the person who runs it.
19 yearsin online advertising, since 2007
9 yearsat Grupa Wirtualna Polska
5 rolessuccessively at Wirtualna Polska, from Senior Traffic Manager
4 marketsPoland, Czechia, Germany, Romania
Grupa Wirtualna Polska
2017 → present
Warsaw · independent contractor
- Programmatic DirectorMar 2025 → present
I own programmatic end to end: monetization strategy, yield, header bidding and ad revenue across the whole portfolio of sites. I have led the analysis and optimization team continuously since 2023.
- Head of Programmatic Analysis and Optimization TeamApr 2023 – Feb 2025
Leading the analysis and optimization team: data, reporting and the decisions that move auction outcomes.
- Programmatic CoordinatorJan 2019 – Mar 2023
- Programmatic SpecialistJul 2017 – Dec 2018
- Senior Traffic ManagerJan 2017 – Jun 2017
Business Ad Network
2010 – 2016
7 years · Warsaw
- Head of TrafficJan 2010 – Dec 2016
Running the traffic department of a network of premium business content sites: campaign management in SmartAdServer, DSM and DFP (now Google Ad Manager), creative QA, publisher support, development of new ad formats and implementation of custom ad server solutions.
Sabela Media Polska
2008 – 2009
Warsaw
- Traffic Manager · Adnet Team LeaderFeb 2008 – Dec 2009
The Adnet traffic department — several hundred websites across many categories and several hundred campaigns a month. Implementation of XClaim (contextual advertising) and nugg.ad (behavioral targeting), client support on Eyeblaster ACM, and designing custom ad server tooling.
Adnet
2007 – 2008
Warsaw
- Junior Traffic ManagerAug 2007 – Jan 2008
Campaign delivery in the AdOcean ad server and Jet.Mail for email marketing, creative QA, publisher support and web analytics in gemiusTraffic.
I studied at Warsaw University of Technology. Since 2012 I've also run my own practice — Adekspert.
Contact
Describe the problem in two sentences
Tell me what keeps repeating, or what stopped adding up. I'll tell you whether I can help and what I'd propose as a first step. If it isn't a topic for me, I'll say that too.
I have run Adekspert since 2012. I work on a project basis and remotely, alongside my work with Wirtualna Polska. I work best on a specific problem that can be resolved with an audit, a rollout or a finished tool.
- Email[email protected]
- LinkedIn/in/pawelscieszko
- Based inWarsaw, remote across Europe
Adekspert Paweł Ścieszko
ul. Eukaliptusowa 4/2
02-765 Warsaw, Poland
02-765 Warsaw, Poland
- VAT ID
- PL7742848812
- REGON
- 146112770
- Since
- 2012