Where Guides Get Their Clients: Channel Share

Broadcast rules require a station to announce, at the time of transmission, that matter is sponsored, paid for or furnished, and by whom or on whose behalf the consideration was supplied. Then they add a duty to go and find out: reasonable diligence to obtain from employees and from persons dealt with directly the information needed to make that announcement.

Two obligations, and the second is the interesting one. Knowing where something came from is treated as work somebody has to do rather than as information that arrives by itself. Which is the whole problem with channel share in a guiding business: nobody is doing that work, so the answer everybody gives is the answer the client volunteered, and the client is guessing. No channel share figure is asserted anywhere on this page. Everything alongside it is indexed at the guide industry data hub.

What does the sponsorship rule require?

Disclosure at the time, and diligence to find out beforehand.

Section 73.1212(a) of Title 47 requires a broadcast station transmitting matter for which money, service or other valuable consideration is directly or indirectly paid or promised to announce, at the time of the broadcast, that the matter is sponsored, paid for or furnished in whole or in part, and by whom or on whose behalf the consideration was supplied.

It provides that service or other valuable consideration does not include property furnished without or at nominal charge unless furnished in consideration for an identification beyond one reasonably related to the use of that property on the broadcast.

Paragraph (b) requires the licensee to exercise reasonable diligence to obtain from its employees, and from other persons with whom it deals directly in connection with any matter for broadcast, the information needed to make the announcement.

Which treats provenance as something to be established rather than assumed, and places the burden on the party doing the publishing.

The section itself sits at ecfr.gov, and the rules there bind licensed broadcast stations rather than anybody in a boat; confirm the current position with a lawyer before applying any of it to your own publishing.

No aggregate answer exists for this trade, for reasons unpacked in the methodology piece.

Risen Tide Sportfishing, CA

Why is the client's answer unreliable?

Because they are reporting the last step of a sequence they did not track.

A person who heard your name at a fly shop in March, followed you for four months, and searched for you in July will say they found you on a search engine, because that is the last thing they remember doing.

Which is not dishonesty and it is not a failure of memory; it is what happens when somebody is asked to reconstruct a process they were never attending to.

The consequence is that a channel figure built from client reports systematically credits the final step and undercounts everything upstream.

Which produces a familiar and wrong conclusion, that search brings the clients and everything else does not, in operations where the reverse is closer to true.

The error is not correctable by asking more carefully, since the information the question requires was never encoded.

How the question should be asked instead is dealt with below.

The same reconstruction problem wrecks surveys, which the methodology piece examines.

Wings and Walleyes, MN

What is the diligence version?

Two questions instead of one, asked at different moments.

The first is asked at booking and is the ordinary one, which produces the last step and is worth recording as exactly that.

The second is asked in conversation on the water, when the client is relaxed and there is nothing at stake, and it is a different question: how long had you known about me before you got in touch.

Which produces the upstream information that the booking question cannot, and it produces it in a form people can actually answer.

Answers cluster into three groups: found and booked in the same week, knew about you for months, and were recommended by somebody specific.

Recording which of the three, alongside the reported last step, gives a picture no single question produces.

It takes ten seconds and it happens in a conversation you were having anyway.

Where the answer gets recorded is set out in the debrief piece.

Why does the aggregate question have no answer?

Because the mixture differs so completely between operations.

A destination lodge fills through a small number of travel agents and repeat groups, and search barely matters.

A day-trip operation in a tourist town fills through walk-past, hotel recommendations and platforms, and its website may be almost irrelevant.

A specialist technical guide fills through reputation within a small community, and could stop advertising entirely without noticing for two years.

An operation running paid search in a competitive market has a channel mix nobody else in the trade would recognise.

Which means an average across them describes none of them, and would be actively misleading to any operator using it to allocate effort.

The mixture is a consequence of the business model rather than a fact about the trade.

The same structural point about rates is set out in the state day rates piece.

Does the mixture change over time?

Steadily, and the direction is predictable.

A new operation acquires almost entirely through whatever is purchasable, because nobody knows its name and there is no word of mouth to draw on.

An established operation acquires increasingly through referral and repeat, which is the same shift every service business makes and it happens over years rather than seasons.

Which means a channel mixture read at a point in time is also a statement about the age of the business, and comparing a five-year operation to a fifteen-year one measures mostly that.

The practical consequence is that the marketing effort which built the business is not the effort that sustains it, and the transition is rarely made deliberately.

Operations frequently keep paying for acquisition long after referral has taken over, and occasionally stop paying just as growth would have compounded.

Watching your own mixture move across three or four seasons is the only way to see which situation you are in.

The retention half of that shift is set out in the repeat rate piece.

What about the channels that produce enquiries but no bookings?

The most misread part, and it needs separating.

A channel producing many enquiries that never convert is frequently blamed on the channel and is frequently a mismatch between what it promises and what you sell.

Which shows up as a high enquiry count from people asking about half days when you run full days, or about a species you do not target.

Recording enquiries as well as bookings, by source, separates volume from suitability and turns a vague dissatisfaction into a specific one.

The fix is almost always upstream, being what the listing or the advertisement says rather than the channel itself.

Which is a cheap fix and it only becomes visible when the two counts are kept apart.

An operation counting only bookings has no way to distinguish a channel that reaches nobody from one that reaches the wrong people.

What the enquiry should be answered with is set out in the enquiry replies piece.

Is any single channel enough?

Only until it changes, and it always changes.

An operation filling entirely from one platform, one lodge relationship or one search position is a decision away from having no business.

Which is a concentration risk rather than a marketing question, and it is invisible while the single channel is working.

The recognisable version is an operation that has not thought about acquisition in five years because a relationship has been carrying it.

Building a second route while the first is working is unglamorous, feels unnecessary, and is the only moment it can be done calmly.

The specific test is whether losing your largest source tomorrow would end the season, and for a great many operations the answer is yes.

Which is worth knowing deliberately rather than discovering.

Where a second route usually comes from is set out in the direct bookings piece.

What can be measured properly?

Anything with a mechanism, which is a short list.

A booking arriving through a platform is unambiguous, because the platform is the mechanism and there is nothing to infer.

A booking arriving on a phone number used nowhere else is unambiguous for the same reason, which is why a distinct number on one channel is worth the small cost.

A booking arriving through a form on a page that only one campaign links to is unambiguous while that remains true.

Everything else is inference, and the inference is worth making explicitly rather than pretending otherwise.

Which suggests the practical approach: build mechanisms where you can, and treat everything without one as unattributed rather than assigning it to whatever the client said.

An honest report with a large unattributed bucket is far more useful than a complete-looking one built on guesses.

A report built that way is described by the marketing report piece.

Does the referral answer mean anything?

More than any other answer, and it should be pursued.

Where a client names a specific person who recommended you, that is the one genuinely reliable attribution available, because it is a memory of a conversation rather than a reconstruction of a process.

Which makes it worth capturing precisely, including who, since the pattern across a season identifies the small number of people generating disproportionate business.

Most operations discover the same three or four names recurring, and most have never thanked any of them.

Doing so is the highest-return action available from this entire exercise, and it costs a message.

The same record also reveals whether referrals come from clients, from other guides, or from shops and lodges, which are three different relationships requiring three different responses.

None of that is visible without recording the name rather than the category.

How to handle those relationships is set out in the tradition piece.

What about platform bookings?

Measured perfectly and interpreted badly.

A platform booking is unambiguously a platform booking, which makes it the best-measured channel and therefore the one that dominates any report.

What the measurement cannot see is whether that client would have found you anyway, which is the question that determines whether the commission was worth paying.

Some platform clients are genuinely incremental and some are people who would have booked directly if the platform had not been in the way.

Distinguishing them is difficult and the one available signal is whether they searched for you by name before booking, which occasionally shows up in the platform's own referral data.

Which is worth asking a platform about and worth asking the client about at the takeout, since people frequently remember whether they were looking for you specifically.

The economics of that question are substantial across a season.

What direct conversion involves is set out in the direct bookings piece.

How long is the window?

Longer than any measurement tool assumes.

The gap between first hearing of a guide and booking one is frequently months and sometimes years, particularly for a destination trip that requires arranging leave and travel.

Which breaks every attribution mechanism that relies on a browser remembering something, since those horizons are measured in weeks at best.

It also means a channel switched off today continues producing bookings for a considerable period, and the absence only becomes visible much later.

Which is the specific trap in reading a channel report after a change: the effect of stopping something appears well after the decision, by which point it is attributed to something else.

The correction is to make changes one at a time and to allow a full season before concluding anything, which nobody has the patience for.

Asking clients how long they had known about you is the cheapest available substitute and it is directly informative about this.

The lead time evidence is examined in the lead times piece.

What should the record look like?

Three fields, and one of them is allowed to say unknown.

Reported source, recorded as what the client said rather than as what you concluded.

Mechanism, recorded only where one exists, being a platform, a distinct number or a dedicated page.

And prior awareness, recorded as same week, months, or named referrer, from the conversation on the water.

Where the second field is empty and the third is unknown, the booking is unattributed, and recording it as unattributed is the honest entry.

A season with a third of bookings unattributed is a normal season, and reports that show none are reports that guessed.

Three fields, filled at booking and at the takeout, produce something no published figure could.

Where the fields live is set out in the spreadsheet CRM piece.

What happens to the answer at the end of a season?

It gets read once, and read wrong unless the counts are kept.

The November review is where the record either produces a decision or produces a story, and the difference is whether the unattributed bucket was kept honest.

A report showing a clean split invites reallocation, and reallocation based on a guessed split is worse than no reallocation at all.

Which is why the unattributed bucket has to survive into the report rather than being distributed proportionally, which is the standard tidy-up and the standard error.

What the honest report supports is a narrower decision: whether a channel with a genuine mechanism paid for itself, which is answerable, and nothing else.

Everything without a mechanism goes into a qualitative paragraph rather than a number, and the paragraph is more useful than the number would have been.

Reading it against the previous two years is where any pattern appears, and one year in isolation supports nothing.

Where do channel figures go wrong?

Six ways, and last-touch is the first.

Crediting the final step, which the client reports accurately and which was rarely the cause.

Assigning unattributed bookings to whatever the client guessed, rather than recording them as unknown.

Reading platform volume as platform value, without asking whether those clients were incremental.

Ignoring the length of the window, so the effect of a change is attributed to whatever happened next.

Recording referral as a category rather than as a name, which discards the most actionable information available.

And comparing your mixture to somebody else's, when the mixture is a consequence of the business model.

The report that avoids all six is described in the marketing report piece.

What is the working method?

Two questions, three fields, and a visible unknown bucket.

Ask the ordinary question at booking and record the answer as what was said rather than as a conclusion.

Ask how long they had known about you, on the water, when nothing is at stake, and record same week, months, or a named person.

Build mechanisms where you can, being a distinct number, a dedicated page or a platform, and treat everything else as unattributed.

Record referrers by name, look at the pattern once a season, and thank the three or four people who recur.

Change one thing at a time and allow a full season before concluding anything, because the window is longer than any tool assumes.

Compare only against your own prior seasons, since the mixture is a property of your operation rather than of the trade.

The Communications Act provision behind the sponsorship rules is 47 U.S.C. 317, with the section mirrored on govinfo.

Feeding it into something readable is the job of the marketing report piece.