Advertising, software
and work built to order.

I build accounts around what people actually search for, not around a list of words from a generator. The structure has to be simple enough that you can say which campaign earns its keep, and detailed enough that it can be steered.

Google Ads: scope

• Account audit and campaign rebuild
• Search, Performance Max and remarketing
• Negative keywords and traffic quality control, so budget does not leak into stray queries
• Bidding strategies set against cost per customer, not against click volume
• Local campaigns for multi-branch companies, and export campaigns

Ad relevance and landing page quality get separate attention, because they decide the price per click. A cheaper click usually comes from a better match than from a bigger budget.

An example from practice: the transfer company case study, where a new website and redirected ads won back almost half of sales from intermediaries.

Facebook and Instagram run for sales and enquiries. The biggest threat here is not a poor audience but creative fatigue, so I watch frequency and swap material out before the results start to slide.

Meta Ads: scope

• A funnel split into cold, warm and remarketing traffic, with exclusions between the levels
• Server-side conversions, so measurement does not depend on the browser alone
• Creative production and rotation, including vertical formats
• Local reach campaigns around a physical location

I manage Google Business Profiles on behalf of my clients, including companies with several branches. This is not a one-off listing setup but ongoing care: categories, services, photographs, reviews and posts.

A business profile is often the cheapest channel a company has, and the most neglected. I have seen accounts paying for a treatment in advertising that was missing from the service list on a free profile visited several thousand times a month.

Google Business Profiles: scope

• Setup and clean-up, separately for each location
• Primary and secondary categories chosen against what the company really does
• Services, products and attributes filled in from real queries
• Map rank measured on a geographic grid around the location
• Reviews and questions handled, posts on a schedule
• Name, address and phone consistency across third-party directories

Without trustworthy measurement, everything else is guesswork. I set tracking up so that data from the ad platforms and data from analytics can be placed side by side, and so that any difference between them has an explanation.

Analytics and measurement: scope

• Google Analytics 4 and Google Tag Manager
• Conversions and key events, split into primary and secondary
• Detection of double counting, which inflates results artificially
• A monthly report where every number carries its source

If something cannot be measured reliably, I say so plainly instead of supplying a figure that looks good in a table.

I look after brand profiles on Facebook and Instagram: publication planning, copy, preparation of material, and handling comments and messages. Running a profile is treated as part of the sales system rather than a box to tick, so the publication calendar is built on the same data as the paid campaigns.

Production side

Film work is produced together with NeonCup Studio, a production house in Krakow. That gives access to the whole process, from concept and script through shooting, editing and post-production, without the client having to coordinate several independent suppliers.

The real gain comes from the order of work. Material is created knowing where it will be used: which formats a given campaign supports, where the first second of the message lands, what must stay clear when watched without sound. Production commissioned separately, detached from the media plan, usually needs reworking afterwards or goes unused.

Social media and video: scope

• Presence strategy and publication calendar for Facebook and Instagram
• Copywriting, design work and material prepared for vertical formats
• Comment moderation and handling of private messages
• Brand films, product films and event coverage
• Short-form material for paid campaigns and organic posts
• Product and reportage photography

These are built and working products that can be switched on without waiting for a project from scratch. They run every day on real data, for real clients, and deployment is counted in days rather than quarters.

AI tools: what exists and runs

• Monitoring of accounts and websites that reports a measurement failure the same day
• Local visibility measured on a geographic grid, separately for each location
• Tracking of competitor presence and offers in the channels where customers are won
• Systems that collect market data from many sources and bring it into one database
• Reporting panels built for a specific question rather than for a list of available charts

Building is the smallest part of the cost

Producing a working tool with the help of artificial intelligence is now a matter of days. That is no longer a technological advantage. It is the starting point.

The real cost sits in operation. Suppliers change their programming interfaces, data sources introduce new protections, and formats and edge cases appear that nobody foresaw at the design stage. Each of these requires a response within a reasonable time. Without it the tool either stops working or, worse, continues to run while producing incorrect results.

For that reason I do not run deployments that close with a final handover. A solution left without care loses its credibility within months, usually without any visible signal. My responsibility is that the tool stays useful in the quarters that follow, not only on the day it is delivered.

If a tool only makes sense for one client, it is built for one client. The code and the data stay on their side.

Some problems do not fit any of the categories above: thousands of records to work through, markets to compare, distributors to find in a foreign country, or a check on whether the data somebody bases a decision on is accurate at all.

For work like this I use artificial intelligence as a working tool, not as decoration. Research that would take weeks by hand comes down to a dozen or so hours of machine time, and my role moves to what genuinely needs a person: asking the right question and checking whether the answer can be defended.

What that looks like in practice

• For manufacturers: who sells your brand in a given country, how much traffic they have, whether stock levels add up and whether they hold agreed prices
• Finding and assessing distributors and partners in foreign markets
• Shelf monitoring: tracking product presence at trade partners
• Searching and organising large data sets that cannot be reviewed by hand
• Verifying company identity in official registries before anyone signs a contract
• Automation of repetitive work that eats a team’s time

One rule applies throughout: every number that reaches the conclusions has a source I can point to. Missing data is recorded as missing, not as zero. The result has to stand up to scrutiny, not merely look good.

Sector competence cannot be declared. It is built over years on specific accounts. The areas below are familiar enough that they do not require a learning period paid for by the client. They share one trait: high value per enquiry, a long decision process, and competition that does not make obvious mistakes.

Manufacturers and distributed brands

Companies selling through partners that need to know what happens to their brand across distributor and retailer channels, including abroad. Product presence checks, distributor assessment, entry into a new market.

Automotive

Detailing, coatings, protective films, workshops and accessories. A market where the customer compares for a long time and decides on the basis of a result that has to be shown rather than described.

Property developers and real estate

Residential development sales, where a single enquiry is worth as much as a month of budget in another sector, and the decision cycle is counted in months rather than days.

Health and beauty

Clinics, aesthetic medicine, dentistry, trichology. A category hemmed in by advertising restrictions, where one careless sentence can suspend an account. I know where the line runs.

Travel and tour operators

Tour organisers, transfers, accommodation and travel-adjacent services. This category is unusually demanding operationally: demand is strongly seasonal, prices change mid-campaign, and between the first search and the booking a customer visits a dozen sites and compares offers for weeks. The campaign has to react to availability and exchange rates rather than run on a fixed schedule.

Businesses where the phone closes the sale

Anywhere orders are not placed in a basket but during a conversation: medical, legal and contracting services, individual quotations and project-based sales. This needs a different approach to measurement, because the result happens off the website. I separate genuine calls from accidental ones, take call duration and time of day into account, and draw conclusions from the enquiries that actually turned into work, not from the number of taps on a phone number.

Local services in a crowded market

Cities where a dozen suppliers compete for the same customer and the price per click can run into tens of PLN. Those are won on precision: negative keywords, ad scheduling, radius and map visibility, rather than on raising the budget.

I do not sell guaranteed search positions on chosen keywords, I do not buy links from private networks, and I do not take work where the result depends on a promise that cannot be kept. If I judge that advertising makes no sense in a given case, I will say so in the first conversation.