Early on Tuesday, five websites announced that the San Damián components factory would close before summer and that 480 jobs would disappear. The number looked firm because it was repeated with the same detail: “480 direct jobs.” A neighbour saw it in an aggregator, a worker received it in a messaging group, and local shops began calculating what it could mean for their orders. Yet opening the links brought a surprise. The second article cited the first; the third referred to “reports in the media”; the fourth linked nowhere; the fifth reproduced a note attributed to an investment adviser. None displayed the document, statement or record from which the number came.
Repetition can create a misleading feeling of confirmation. Not because summarising another outlet is automatically wrong, but because every new masthead can add appearance without necessarily adding a route to knowledge. Five similar headlines are not five checks when each depends on the same initial sentence. In employment, industrial or local matters, the distinction has weight: a premature forecast can change choices by families, suppliers, unions and public bodies long before an official communication exists.
The number is not yet the fact#
The first useful move is to separate three things that a headline often compresses. There is the claim: “the factory will close and 480 jobs will be lost.” There is the source being invoked: an advisory note, a conversation, an investor presentation or an earlier publication. And there is the fact that could support the claim: a company communication, a regulatory filing, a union minute, a board decision or an identifiable answer from someone able to confirm it. When those layers merge, a conjecture can acquire the tone of a resolution.
The first useful move is to separate three things that a headline often compresses.
An adviser’s note may deserve attention. It may interpret changes in orders, moves by a parent company or market comments. It is not, however, a primary source merely because it uses technical vocabulary or circulates among people who follow the sector. The question is not meant to discredit the adviser. It asks what they saw, what they heard directly, what they inferred and which part they cannot document. A reasoned forecast remains a forecast. Calling it that prevents it from carrying the weight of a decision nobody has shown.
In San Damián, a careful newsroom would begin by saving the exact version of each publication before it changes. That is not about building a case against other websites; it is about observing the route. Which item used the number first? What words travelled with it: “may,” “plans,” “confirmed”? Was there a link? Did the note refer to a report, a meeting or an unnamed source? This map prevents a common mistake: treating the latest, polished and categorical version as though it were the origin.
Draw the route before counting voices#
A provenance map can be uncomplicated. On a sheet, list nodes: the adviser’s note, a business portal, a local newspaper, an aggregator, social accounts and trade newsletters. Draw arrows according to citations, links and publishing times. If three pieces came through the same portal and that portal rests on the same note, there is one branch of evidence, not three. If a union confirms in writing that it received a closure plan, or a company answers specific questions, another branch appears and should be labelled separately.
Labelling independence is more honest than counting sources by sight. A source may be primary and still have its own interests. Another may be independent of the company while relying on the same leak. A third may supply useful context without confirming the central number. The map does not award permanent credibility medals. It describes each item’s specific relationship to the claim. That description lets readers understand why two different texts may not amount to two different checks.
Silences need recording without being filled in. If the company does not answer before deadline, that means it has not answered; it does not mean it admitted the closure. If the works council says it has received no communication, that reports its position but does not prove that no confidential plan exists. Precise language does not make a story less compelling. It provides a sturdier basis for updating it when new material appears.
Asking for the origin changes the exchange#
A request for the primary source should be exact. Rather than asking vaguely, “Is this true?”, a reporter can ask, “Which document supports the figure of 480?”, “Who issued the estimate?”, “On what date?”, “Does it concern direct staff, temporary workers or contractors?”, “Is there a decision, or an analysed scenario?” Narrow questions reduce the chance of receiving a reassuring but empty answer. They also make clear what will be checked if documents arrive.
When the adviser cannot share material because of confidentiality, the problem does not vanish. A newsroom can report that an estimate has been attributed to that person, explain that it has not been able to verify the estimate’s basis, and avoid turning it into a settled headline. It can choose not to publish the number until it finds additional support. Neither approach hides a story. They are different ways of matching the scope of what is said to what is actually known.
The cost of this caution should not sit only with the writer. If an organisation requires routes to be confirmed, it must protect time to call, read documents and contact people again, rather than measure only who published first. An editing desk can adopt a plain rule: figures with material effects need a visible origin field and an independence classification before distribution. The rule does not replace judgement, but it stops urgency from turning a circular link into collective support.
Correct the route as well#
Imagine that two days later it becomes clear that the 480 jobs were a calculation of plant capacity, not an announcement of redundancies. The correction should not be limited to replacing a number buried at the end of the article. The central claim needs correcting, alongside an explanation of what was published, what has since been checked and where the earlier version travelled. If the outlet sent an alert, posted on a platform or supplied the text to a newsletter, those channels are part of the chain that needs repair.
Correcting the route does not mean humiliating people who shared in good faith. It means making the difference between an estimate and a decision visible. The newsroom can add the correction to the map and contact outlets that relied directly on its text. People who received the information can then understand why it changed, rather than encounter a new version without context while keeping an old screenshot as though it were current.
The lesson reaches beyond one factory. Matters concerning work, health, housing or rights move quickly because uncertainty hurts and numbers appear to offer control. That is exactly why it is worth asking where they come from. The next time five headlines repeat a closure, the job is not to decide which one sounds most certain. It is to trace the route, request the origin, mark which branches are genuinely independent and stop the chain when evidence is insufficient. Responsible reporting can admit that a question remains open; it should not turn an echo into proof.
The practice has a further public benefit: it lets affected people know what they can do while they wait. A workforce can ask its representatives for information without assuming the outcome is decided; a supplier can review scenarios without cancelling agreements over a rumour; a council can prepare support without presenting hypotheses as facts. Traceability does not remove uncertainty, but it keeps uncertainty from dressing as certainty and governing irreversible choices.
