Decoding the Football Information Supply Chain: Why Your Feed Contains Less and Less Football

On the night of February 1, 2026, I stood for the full 90 minutes in Stand B of My Dinh National Stadium, not out of excitement but because there was no seat left. Vietnam beat China 3-1 in the third round of Asian qualifying for the 2026 World Cup — Vietnam's first win at that stage. Ho Tan Tai opened the scoring with a header, Nguyen Tien Linh doubled the lead, and by the 76th minute the whole stand had dissolved into a single block of sound I have not heard at My Dinh since.

On the way home I wrote one line in my notebook: this is the densest raw data Vietnamese football has produced in years. Ninety minutes, twenty-two players, a counter-attacking system executed almost without error. A match like that supplies material for weeks of analysis.

Thirty-six hours later I opened the football feed I follow every day and counted. Of the first two hundred items, the share describing an actual football event — a match, a goal, an injury, a transfer, a referee's decision — did not exceed one third. The rest was private life, off-field remarks, fifteen-second clips, items that would contain nothing about football at all if you deleted the proper nouns.

Then, at 3:12 a.m. on February 3, 2026, my monitoring board lit up with a line tagged with the classification "football." Inside was a story about two personalities from a Mexican reality television programme. No club. No player. No competition. No transfer. No tactics. I sat still for about a minute, then did what most sports editors have done at 3 a.m. at least once: I asked where the system had broken, and since when.

That is the starting point of this piece. It is not about the match at My Dinh. It is about the pipeline.

Context: a three-layer pipeline almost nobody sees

Football sells exactly one thing: attention with an expiry date. Everything else — tickets, broadcast rights, shirts, sponsorship, betting — is a mechanism for converting attention into money. And attention, like water, always flows through the pipe with the greatest capacity, not the cleanest one.

The football information pipeline has three layers. The source layer comprises events and the people who generate them: clubs, players, coaches, agents, competition organisers, press rooms. The classification layer comprises news agencies, content aggregators, feeds, tagging algorithms and distribution platforms. The amplification layer comprises accounts, pages, livestreams and the audience itself — the people who decide which item gets passed on.

For a market like Vietnam this structure has an unusual property. Hourly demand for football content is enormous, while domestic supply is brutally limited. V.League 1 runs with roughly fourteen clubs, and each round produces only seven matches. Those seven matches yield about fourteen halves of football: under fifteen hours of raw data per week. The rest of the week — more than 150 hours — has to be filled with something else.

Across nearly two decades in this trade, I have watched that gap get filled with three materials, in descending order of quality: domestic data mined at greater depth, imported international data, and non-football content dressed in football clothing. The third is the cheapest, the fastest and the most dangerous.

The source layer: news is rarely invented, it slides

A common misconception outside the industry is that false stories are fabricated from nothing. The reality is harsher. Almost every wrong report begins with a true fact, and its death is a slow, layer-by-layer slide in which each step removes a little certainty.

The case I use as a template sits outside football but has the exact structure of a transfer deal. A public conversation, a courtesy invitation, a few occasional messages. From that raw material, the content machine builds a complete story about a relationship. The original story, on its own terms, denies the strongest version of itself. But the strongest version is the version that travels.

The football version of that slide is familiar enough to recite from memory: an agent says his client is "looking for a new challenge." Two days later it becomes "the club is negotiating." Five days later, "personal terms have been agreed." Seven days later, "the unveiling is expected this weekend." On day nine there is no unveiling, but the cumulative engagement of the whole chain is enough to turn a meaningless sentence into a topic inside the week's top discussions.

A report is rarely fabricated from scratch; it slides through the layers, and each layer strips away a degree of certainty. The reader at the end of the pipe receives the weakest version in terms of evidence and the strongest in terms of emotion. That is a design, not an accident.

In this system the source layer has no incentive to extinguish a rumour early. A club that immediately denies a transfer forfeits the attention the rumour generates. An agent who issues one clear denial burns his own negotiating instrument. And a player in contract talks has a very specific incentive to keep things ambiguous. Ambiguity, in the attention economy, is a product with a price.

The classification layer: the wrong tag and the missing entity gate

This is where the non-football story and the football story meet, and where someone in my trade has to stop.

An item about two Mexican television personalities — Gala Montes and Adrian Marcelo — passed through a classifier and received the tag "football." No team, no player, no coach, no competition, no contract appeared anywhere in the content. Handed to a sports editor for thirty seconds, it would have been sent straight back. The classifier did not send it back.

The usual reason is simple. Most content classifiers tag on surface signals: keywords in the headline, matching proper nouns, or data arriving from a feed that was once filed under sport. A television personality mentioned once, months ago, in an article about football is enough to create a weak link. When a new article about that person appears, the algorithm connects the two points and tags out of habit.

What is missing from the whole process is an entity gate. A very cheap rule: if content contains no entity from the football set — club name, player name, competition name, governing body — it may not carry the football tag. That rule can be written in a few lines and run before any more sophisticated classification step.

A classification tag is a contract between platform and reader, and every wrong tag is a breach of contract with no one liable for damages.

When a non-football item enters a football feed, the damage is not the item itself. The damage is that it consumes one slot of the reader's finite attention and dilutes the signal of the genuine items around it. At a one-to-two-percent error rate in a batch of a few thousand, the consequence is effectively zero. Once that rate systematically exceeds two percent, it stops being an isolated bug and becomes a defect of the classifier. And classifier defects do not heal on their own.

I am aware I am extrapolating from a single case. One wrong item does not prove a system is wrong. But my profession is reading structure out of a single passage of play, and the structure here has a familiar shape: contamination enters at a low layer, and every layer above inherits it.

The amplification layer: the hype-to-kill cycle and the price of data

There is a pattern transfer analysts call the hype-to-kill cycle: a weak claim is amplified by audience appetite, peaks, then reverses or is denied, leaving behind a sediment of people who believe they were just cheated.

Its mechanics are measurable. In football, the strongest evidence comes from the period when the world briefly lost its crowds. In 2026, when European leagues returned to empty stadiums, two German economists published a study on Bundesliga data showing the home win rate falling from 43.1% to 31.2%. Those numbers matter more than they look. They prove the crowd is not decoration; the crowd is an input variable, weighted, measurable, and removable from the equation.

An empty stand is a mirror exposing the truth about home advantage. Many clubs believed their home advantage lived in the pitch, the climate, the travel schedule. When the stands went silent, most of it turned out to live in the singing. What remains after you remove the crowd is the team's real capacity.

That same year, money drained out of the international transfer market. FIFA's report recorded total international transfer spending of USD 5.63 billion in 2026, down 23.4% from USD 7.35 billion in 2026 — the first decline after nearly a decade of continuous growth. The notable part is not that money shrank, but that the market collapsed more slowly than revenue did. Clubs responded to lost revenue with cheaper deals, more loans, and more rumours.

Rumour is the cheapest way to keep a market warm. When there is not enough money to create events, people create stories about events. The hype-to-kill cycle is therefore not a media vice; it is the rational response of a system short on capital and long on attention demand.

The crowd inside the pipeline, and a lesson from a night in Kazan

At the amplification layer, the audience is the pipeline's crowd. Remove the audience from the equation and the rest of the system becomes honest to an uncomfortable degree.

I learned this in Kazan. On June 30, 2026, France beat Argentina 4-3 in the round of sixteen at the 2026 World Cup. I was in the stand, and the thing I remember most is not Kylian Mbappe's two goals. I remember writing a very hot piece within ten minutes of the final whistle, arguing that France should build the entire team around Mbappe and that Paul Pogba was overrated. It travelled extremely fast.

By two in the morning I was rewatching the tape and found what I had missed: N'Golo Kante, with a pass completion rate around 87%, was holding that structure upright while the front line exploded. Kazan was not the day Mbappe detonated; it was the day Kante taught modern football. I deleted the old post at dawn and rewrote it from scratch.

The loudest event in a match is almost always the least informative one. The moment Mbappe ran forty metres made the report travel, but structure won the game. A good information pipeline prioritises structure over volume.

This is also why I do not trust the romantic version of football. The story of a small club beating a giant is satisfying, but behind it there is nearly always a financial gap being hidden and a sustainability problem being ignored. A small club can win a match. A small club cannot move a resource structure in one night. Any analysis that skips that part is selling emotion, not information.

Another example sits in tactical structure. In 2026, working as an editor in the domestic football market, I wrote that Wu Lei should not be the attacking centre of the club he played for, on the grounds that he was sacrificing too much for the foreign players around him. The piece drew thousands of angry comments. The next evening I went to the stadium and watched Wu Lei create four key passes in a win. I livestreamed an admission of error while adding the analysis I had left out: the problem was not the player but the system around him. A 4-3-3 does not swallow a Wu Lei in full flight. And a system that cannot absorb what it needs will eventually absorb what it does not — exactly like a classifier absorbing entertainment content into a football feed.

The contrarian angle: the fans are not deceived, they are co-authors

This is where almost every critique of sports media goes in the wrong direction.

The most popular version of the story is: platforms manipulate, audiences are fooled. I think both halves are wrong.

Platforms do not need to manipulate. They only need to optimise for engagement. And engagement in football peaks in exactly two states: when a question is unanswered, and when an answer has just been reversed. A cleanly confirmed fact generates far less engagement than a question left hanging. That is arithmetic, not morality.

On the audience side: years of transfer research show that public opinion does not actually want a clear conclusion. It wants a prolonged ambiguity, because ambiguity lets people participate — to bet, to argue, to predict, to build dream line-ups. Clarity closes the game. Ambiguity keeps it open. Demand for ambiguity is the pipeline's largest fuel source.

If ambiguity is the product and engagement is the metric, then the pipeline is not broken when it produces ambiguity — it is running exactly as designed. The only broken part is the verification gate.

And here is where I may be wrong. Let me be direct about it.

First, if the real tagging-mismatch rate runs below two percent and is concentrated in a handful of stale feeds, then my whole framing inflates the problem. A classifier running at 98 percent accuracy is a good classifier.

Second, I have an occupational bias. I make a living from analysis with depth, so I have a direct interest in declaring the pipeline poisoned. Someone who makes a living from fifteen-second livestreams would look at the same data and conclude the system is healthy.

Third, and this is what I remind myself every week: I have been publicly wrong about Kante, wrong about Wu Lei, and wrong about deleting a post. The people who publish fastest are usually the people who have to correct the most. Correcting at dawn is part of the job, but it does not exempt you from writing wrong less often.

Signals to track, and falsifiable predictions

I will not close this piece with a summary. I will close with three testable predictions, so that anyone can judge whether I was right.

One: within the next ninety days, if you draw a random sample of two hundred items from any Vietnamese-language football feed at a transfer-window peak, the share containing no football entity at all will exceed thirty percent. I count by eye, with a strict definition of entity: club name, active player name, competition name, governing body.

Two: the next domestic transfer rumour to peak in discussion will reverse or fade within twelve days of its peak, with no confirmation from the club involved.

Three: within a year, at least one sports content distribution platform in Vietnam will put an entity filter into production, removing items without football components from football feeds. Not out of professional ethics, but because user churn will be high enough to make it a business decision.

If all three hold, the pipeline will clean itself. If all three fail, the person who was wrong is me, and I will be the first to rewrite — at dawn, as usual. If only the third holds, then the most worrying outcome has already occurred: we fixed the pipeline without fixing our own reading habits.