Operated function · market and competitor watch
It was in a digest, somewhere between an unrelated press release and a job posting. Monitoring that watches everything produces exactly the volume that guarantees the important item is not noticed — and it produces a false sense that the gap has been closed.
It starts well. Somebody sets up alerts for competitor names, product categories, key terms and a handful of regulatory sources, and for the first fortnight everybody reads them.
Volume wins. Most of what arrives is not decision-relevant — a hiring announcement, a conference appearance, an article that mentions a name in passing — and the ratio of noise to signal teaches everybody, correctly, that the digest can be skipped.
So it gets skipped, and it keeps arriving, which is the worst state: the organisation believes it is monitoring, so nobody else watches, and the actual coverage is zero while the reported coverage is complete.
Meanwhile the things that would genuinely change a decision are frequently not in any feed. A competitor’s pricing page changing quietly, a term appearing in their contract, a regulatory consultation opening, a key person moving — these are page-level changes and document-level changes, not news.
And nobody has written down what would actually change a decision. Without that, monitoring defaults to watching names, which is the easiest thing to configure and the least useful thing to know.
Monitoring scope is set by what is easy to watch rather than by what would change a decision, so the output volume exceeds what anybody will read — and the coverage becomes nominal while appearing complete.
The fix begins with a question almost nobody asks first: what change would cause us to do something differently. Not what would be interesting — what would trigger a decision, by a named person, within a defined time. That list is always far shorter than a monitoring configuration, and it is the correct scope.
Once written, the sources follow from it rather than the other way round. A decision that depends on a competitor’s list price is watched by monitoring that page, not by watching their name. A decision that depends on a regulatory change is watched at the consultation register, not in the press.
Routing is the second half and it is where most programmes fail even when the scope is right. An item that reaches a distribution list reaches nobody in particular; an item that reaches the named person who would act on it, with the reason it was sent, is acted on or explicitly dismissed.
And the volume has to stay low enough to be read. That is a design constraint rather than an aspiration: if the output exceeds what the recipients will actually read, the programme has failed regardless of how good the individual items are.
What is never delegated is the judgement about what a change means. Interpreting a competitor’s move, deciding whether to respond, and setting a price are strategic decisions belonging to your people.
Items sent per period — measured by volume against a baseline digest, where a lower number is the objective rather than a shortfall.
Items acted on or explicitly dismissed — measured by disposition rate, against a baseline where a digest had no disposition at all.
Whether decision-relevant changes are actually caught — measured by known changes found by the programme versus found by a customer, a salesperson or the press.
Time from a change occurring to the right person seeing it — measured by elapsed hours to the named owner rather than to a distribution list.
Coverage of page-level and document-level change — measured by monitored sources that are pages and registers rather than news feeds, which is where most decision-relevant change happens.
Relevance rule accuracy — measured by dismissal rate per rule, which is how a rule that is sending noise gets corrected rather than tolerated.
complete coverage of anything. A monitoring programme that claims to catch everything is claiming something unverifiable, and the sources it cannot see — private conversations, unpublished plans, anything behind a login it has no right to — are frequently where the decisive information is. Nothing here interprets a change, recommends a response, or sets a price.
Monitoring reads publicly available sources, and only those. Nothing accesses material behind an authentication boundary it has no right to be behind, and nothing misrepresents itself to obtain access — both are lines that a competitive-intelligence operation is regularly asked to cross and does not.
Routing lands in whatever your decision owners already use, because an item that requires somebody to log into a monitoring tool is an item that waits until they do.
The decision list and the relevance rules are yours: written by your commercial leadership, versioned, owned. They are the entire scope of the programme and they are the artefact that keeps it from expanding back into a digest.
Only public sources. Nothing accesses material behind a login it is not entitled to use, nothing uses credentials belonging to somebody else, and nothing creates an account misrepresenting who it is in order to see a competitor’s pricing or product.
No pretexting. Nothing contacts a competitor, a customer or a former employee under a false pretext to obtain information. This is the request that arrives eventually in every programme of this kind, and the answer is fixed.
Terms of access are respected. Where a source’s terms prohibit automated collection, that source is not collected automatically, and the limitation is reported rather than worked around.
Every item carries its source and the time it was observed, so a decision made on it can be traced back to what was actually seen rather than to a summary.
Operational access is not permission to train. Your decision list — which describes what your organisation is worried about — does not become material improving anything serving another organisation.
Legal will want the collection boundary in writing — public sources only, no credentials, no pretexting, terms respected — because competitive intelligence is the operation most likely to be asked, informally and by somebody senior, to go slightly further than it should.
Commercial leadership owns the decision list, and it is the whole scope. Without it the programme reverts to watching names, which is what produced the unread digest in the first place.
Where an obligation attaches through competition law, a sector regulator or a source’s terms, it is marked applicability-gated rather than presented as standing — in some sectors, systematically collecting competitor pricing carries its own constraints.
A written list of decisions your leadership would take differently on a specific change — no sources configured, nothing collected — followed by an audit of what your current monitoring actually delivers against it.
The first phase collects nothing. It produces the decision list, which is a short document and a genuinely difficult one, and then checks your existing monitoring against it: how many of those decisions would be triggered by anything you currently watch.
The usual answer is very few, and the second usual finding is that the digest volume is dominated by items that map to no decision at all. Both are actionable immediately by switching things off, which requires no operation and no purchase.
If you continue, sources are derived from the decision list, volume is capped at what the named owners will actually read, and every item routes to a person rather than a list.
Then the useful test is the decision list. Write down what your leadership would actually do differently and on what change, then check how many of those would be triggered by what you currently watch. Most existing programmes score poorly on that test and score well on volume, which is the combination that produces an unread digest. It costs a document and no configuration change to find out.
That is what the current digest offers, and the evidence that it does not work is that nobody reads it. Seeing everything is only meaningfully different from seeing nothing if somebody actually reads it, and volume is what determines whether they do. If your team genuinely reads a high-volume digest and acts on it, you do not have this problem.
Often true, and the page says so — a programme claiming complete coverage is claiming something unverifiable, and the decisive information is frequently in a conversation nobody publishes. What is public and consistently missed is page-level change: a price, a term, a consultation opening, a page quietly removed. Those are knowable, they change decisions, and they do not appear in news feeds.
Then it is not collected. Creating an account misrepresenting who you are to see a competitor’s pricing is not something this operation does at any level of request, and a vendor who offers it is offering you a liability that will surface at the worst possible moment. Where a source is genuinely inaccessible, it is reported as a gap in coverage rather than filled by a route you would not want disclosed.