Counting What Counts for Nothing: The Metric Obsession Quietly Reshaping British Television

There is a particular irony embedded in the current state of British broadcasting. The industry has never possessed more information about its audiences — who they are, when they watch, how long they linger before abandoning a programme, which thumbnails provoke a click. And yet, speak to enough producers, commissioners, and writers working across the British television landscape, and a contradictory picture emerges: the more data the industry accumulates, the less confident it seems in what audiences actually want.

This is not merely a creative grievance dressed up in industrial language. It is a structural concern about how quantifiable metrics have come to dominate decisions that were once guided by editorial instinct, cultural awareness, and a willingness to take the kind of creative risks that have historically defined British television at its finest.

The Feedback Loop Nobody Talks About

At the heart of the problem lies a mechanism that is elegant in theory and troubling in practice. Streaming platforms and broadcaster-owned video-on-demand services feed viewing behaviour into recommendation engines, which in turn surface content that resembles what users have previously watched. Commissioning decisions, increasingly informed by these same data sets, begin to favour projects that mirror existing successes. The result is a feedback loop: popular content begets more content like it, which generates data confirming its popularity, which justifies commissioning yet more of the same.

For genres with established audiences — Nordic-inflected crime drama, home renovation formats, true crime documentary — this dynamic is commercially comfortable. But for programming that sits outside familiar patterns, the numbers offer little encouragement, even when the creative case is compelling.

"The algorithm is brilliant at telling you what someone watched last Tuesday," one independent producer, who asked not to be named, told Televisual. "What it cannot tell you is what they might love next month, if only someone had the courage to put it in front of them."

This distinction — between revealed preference and latent appetite — is one that data science struggles to accommodate. Metrics capture behaviour; they are far less equipped to capture aspiration.

When Numbers Replace Judgement

The shift has been gradual but consequential. Commissioning editors who came of age in a culture of editorial conviction now describe a working environment in which data packs accompany every significant decision. Audience Appreciation Index scores, completion rates, and social media sentiment analyses are no longer supplementary tools — for many broadcasters, they have become the primary language of justification.

This is not without logic. Broadcasters facing financial pressure, shrinking advertising revenues, and intensifying competition from American streaming giants have legitimate reasons to seek evidence before committing significant budgets. Accountability demands measurement.

But several senior figures within the industry argue that the pendulum has swung too far. "There is a difference between using data to inform a decision and using it to avoid making one," said one commissioning editor at a major UK broadcaster, speaking on condition of anonymity. "When a number becomes the reason you say no to something original, you have stopped being a broadcaster and started being an actuary."

The consequences for experimental and distinctly British programming are tangible. Projects that defy easy categorisation — work that might sit between genres, or address audiences not yet captured in existing data sets — face a structural disadvantage. They cannot demonstrate prior engagement because nothing quite like them has been commissioned before.

What the Data Cannot See

Consider the trajectory of some of British television's most celebrated programmes. Many of the series now regarded as cultural touchstones — whether in drama, comedy, or documentary — attracted modest initial audiences and defied straightforward demographic profiling. Their longevity was built not through algorithmic recommendation but through word of mouth, critical attention, and the slow accumulation of devoted viewers who encountered them almost by accident.

This mode of discovery — serendipitous, lateral, driven by personal recommendation rather than platform nudge — is precisely what recommendation algorithms are designed to replace. And in replacing it, they may be eliminating the conditions under which genuinely distinctive television finds its audience.

Dr Sarah Whitmore, a media researcher at a London university who has studied platform recommendation systems, puts it plainly: "Algorithms optimise for engagement within a known parameter space. British television's most interesting work has always operated at the edges of that space, or entirely outside it. The system is not designed to find value in what it cannot yet classify."

The concern is particularly acute for voices that have historically been underrepresented in commissioning — writers and directors from working-class backgrounds, from regions outside London, from communities whose cultural reference points do not map neatly onto the demographic profiles that data systems are built to serve. If the algorithm rewards familiarity, it will inevitably disadvantage the unfamiliar.

The Human Element

Not everyone in the industry accepts the pessimistic reading. Some commissioning editors argue that data, properly interrogated, can surface unexpected patterns — audiences for content that might otherwise have been dismissed as niche, or evidence that viewers are more adventurous than conventional wisdom suggests.

"The data does not have an agenda," argued one senior figure at a streaming platform operating in the UK market. "The agenda belongs to the people interpreting it. If you ask the right questions, you can find the appetite for risk in the numbers."

This is a fair point, but it raises its own difficulty. Asking the right questions of a data set requires knowing what you are looking for — which presupposes the very editorial instinct that data-led commissioning is gradually supplanting. The skill lies in holding both capacities simultaneously, using quantitative insight to test and challenge creative judgement rather than to replace it.

There are broadcasters attempting to maintain this balance. Channel 4, whose remit explicitly requires it to take creative risks and serve underrepresented audiences, has spoken publicly about the tension between commercial imperatives and public service obligations. The BBC's commissioning framework similarly acknowledges that reach alone is an insufficient measure of public value. Whether these stated commitments survive the practical pressures of a difficult market is a question the industry will be answering for years to come.

A Reckoning Worth Having

The conversation about data and commissioning is not, at its core, a debate about technology. It is a debate about what British television is for. If its purpose is to serve audiences who already know what they want, then algorithmic commissioning is a reasonable tool. If its purpose is to surprise, challenge, and occasionally transform the cultural landscape — to do what British television has, at its best, always done — then the current trajectory deserves serious scrutiny.

The metrics will keep improving. The data sets will grow richer. The recommendation engines will become more sophisticated in their understanding of existing behaviour. None of that will resolve the fundamental question of whether the industry retains the editorial courage to commission what the numbers cannot yet justify.

British broadcasting built its reputation on precisely that courage. The risk, as things stand, is that it is quietly being data-managed out of existence.