9 November 2026 · Bogusław Kisielewski

Transformation in Media Tech Starts with Economics, Not Technology

In media today, the advantage does not go to whoever has the most content. It goes to whoever can quickly decide which content to fund, which channels to distribute it through, what data to build around it and how much it costs to win an audience member and then keep them. Business change in the media tech sector is therefore first and foremost a new approach to allocating capital, making decisions and measuring value. Technology serves as a tool. The operating model and product economics remain the responsibility of the board.

For many years, media scale rested on a small number of distribution channels, predictable advertising revenue and fairly stable audience habits. That order no longer works. Users hop between VOD, social media, games, live streams and short-form video without committing to any one of them for long. Content is getting more expensive, and margins can melt away even as reach grows. For this reason, transformation cannot be reduced to rolling out a new CRM, a mobile app or a single AI project.

The starting point for transformation in media tech is economics

An owner or board should not begin by asking which technology to introduce. The right question is: what will the company live on in three years' time, and what capabilities does it need to defend that model? In media tech, growing weight is carried by subscription revenue, addressable advertising, one-off transactions, licensing, hybrid models and the exploitation of intellectual property across several formats.

There is no single correct model. A VOD service with a strong premium brand can build value by limiting customer churn and looking after average revenue per user. A news publisher may achieve a better result by combining subscriptions with an offering for the B2B segment. A games company, meanwhile, will assess its portfolio in terms of player acquisition cost, retention, post-launch revenue and distribution potential on Steam or consoles. In every case, though, one thing matters: the model has to be reflected in the financial data, not only in a strategy presentation.

In practice, this means no longer judging products solely by page views, downloads or campaign reach. Reach that is not backed by a quality relationship with the audience can turn out to be a cost rather than an asset. The board should have visibility of user cohorts, retention at day 30, 90 and 180, net revenue per customer, cost to serve and the actual profitability of individual content formats.

Data should change decisions, not just feed reports

Many media companies collect more data than they are able to use. Analytics stops at dashboards that describe what has already happened but have no influence on the schedule plan, the production budget or the commercial offer. It is a costly form of apparent modernity.

A mature organisation brings together product, sales, advertising, editorial and financial data. The point is not to track every move a user makes, but to make sharper portfolio decisions. Which format draws in new audiences? Which one strengthens loyalty? Which one raises the value of the subscription bundle, even if it does not record much viewing time on its own? Which customer segments are profitable once marketing and servicing costs are factored in?

Discipline is needed here too. Data can reveal correlations, but it cannot replace an editorial, creative or investment decision. An algorithm will not decide on its own whether a brand should move into a new content category, acquire a rights library or launch its own FAST channel. It can, however, significantly reduce the risk of such a decision, provided it relies on well-defined metrics.

AI as part of the operating model, not a presentation for the market

Artificial intelligence delivers real value to media, but only when it is built into a process and has a business owner. The most useful deployments concern metadata automation, archive search, content recommendations, language localisation, customer service, support for advertising sales and speeding up the work of production teams.

The most serious mistake is starting with the question of which language model to use. First you need to define the process, the unit cost, the legal risk and the measure of success. If AI is meant to cut the time needed to prepare promotional material from six hours to one, you have to measure the savings, the quality and the impact on conversion. If it is meant to support VOD recommendations, the measure is not the number of descriptions generated, but growth in viewing time, retention and revenue per user.

At the same time, the media industry has to manage content rights, data protection and brand reputation with care. Automation must not lead to publishing unverified material or to losing control over the style of communication. In news, finance or content relating to children, the acceptable level of risk is much lower than when producing internal summaries or tagging an archive.

Focus on a few use cases that can be scaled

A good AI programme is not a patchwork of a dozen or so experiments scattered across different departments. It should start with two or three use cases that have clear economic value, access to data and the potential to be rolled out on a wider scale. Shared architecture, governance rules and in-house capabilities are worth building only once the results have been confirmed.

This approach looks less impressive in communications, but it works better. The board gets proof that technology improves operating results, instead of yet another pilot with no P&L owner. In larger companies it is also especially important to establish which capabilities stay within the organisation and which can be sourced from suppliers. An in-house team is not always essential. What is essential is ownership of the data, the process and the strategic decision.

Multi-channel distribution requires a new organisational architecture

Media tech is not limited to content technology. It means the ability to manage many distribution channels without losing control over the brand, the rights and the relationship with the user. A single piece of content can simultaneously go to the company's own app, a telecoms partner, a social platform, a TV channel and an international distributor. Each of these channels has different economics, different data and different rules of visibility.

A "be everywhere" strategy can turn out to be a trap. Being present on an external platform may lower the cost of reach and help build recognition, but it often means giving away data, margin and the direct relationship with the audience. Focusing solely on one's own ecosystem, on the other hand, gives more control but raises marketing costs and requires a stronger value proposition. The right choice depends on the type of content, the brand's position, funding capacity and the goal: growing reach, acquiring a subscriber or maximising margin.

This challenge forces organisational change. Editorial, product, technology, sales and finance teams cannot operate on separate decision-making rhythms. The content portfolio should be managed like an investment portfolio: with a value hypothesis, a budget limit, performance indicators and a decision on further funding. This matters particularly where a company combines traditional television, VOD, social content and gaming activity.

The board's job: pace, priorities and accountability

Transformation usually stalls not because of technology, but because of a lack of clear priorities. When every department runs its own "digital initiative", the company incurs costs without achieving economies of scale. The board should state unambiguously which decisions are taken centrally: data architecture, AI standards, pricing logic, content rights policy and shared profitability measures.

A rhythm for managing change is also essential. Short test cycles are useful, but they do not exempt anyone from building capabilities for years to come. A VOD platform, a data library or a new advertising model will not mature within a single quarter. On the other hand, a project that shows no measurable value after two or three cycles should be redesigned or shut down. Capital discipline counts just as much as courage in investing.

Leaders of successful transformations also say plainly what the company will not do. Abandoning unprofitable formats, redundant systems or initiatives without an audience takes more maturity than launching yet another innovation programme. It does, however, make it possible to concentrate people and budget on the areas where an advantage can be built.

Advantage in media tech does not come from access to AI alone, nor from the number of distribution channels. It emerges when strategy, product, data and capital pull in the same direction. Well-executed change enables an organisation to learn faster than the market - and that remains the most valuable asset at a time when business models change faster than technologies.

Based on a Polish post at boguslawkisielewski.pl.