Repeat Client Rate Benchmarks

Securities disclosure requires management to explain why the reported numbers will not necessarily repeat. The discussion must focus specifically on material events and uncertainties known to management that are reasonably likely to cause reported financial information not to be necessarily indicative of future operating results.

A benchmark is the opposite manoeuvre. It takes a past number from somebody else's business and presents it as though it describes yours, with no account of the events and uncertainties that made it what it was. This page does not contain a repeat client rate benchmark for guided fishing, because nothing of the kind turned up in any source consulted, and manufacturing one would be worse than useless. What it contains is why the comparison fails, and what to do instead. Related pages live under the guide industry data hub.

What does the disclosure rule require?

An explanation of why the past will not simply continue.

Section 229.303(a) of Title 17 states that the objective of management's discussion and analysis is to provide material information relevant to an assessment of financial condition and results of operations, including an evaluation of the amounts and certainty of cash flows.

It requires the discussion to focus specifically on material events and uncertainties known to management that are reasonably likely to cause reported financial information not to be necessarily indicative of future operating results or future financial condition.

And it requires descriptions and amounts of matters that have had a material impact on reported operations, as well as matters reasonably likely to have a material impact on future operations.

Which treats a reported figure as something requiring interpretation rather than as something that speaks for itself.

Section 229.303 is carried on the eCFR.

The measurement itself is worked through by the repeat rate piece.

Orland, CA

What does the offsetting-changes rule add?

A warning that a stable number can conceal two moving parts.

Paragraph (b) requires that where the financial statements reflect material changes between periods in one or more line items, including where material changes within a line item offset one another, the underlying reasons be described in quantitative and qualitative terms.

The clause about offsetting changes within a line item is the transferable one, because it describes exactly what happens to a repeat rate.

An operation whose rate held steady at thirty per cent across two years may have lost a third of its long-standing clients and replaced them with new returners, which is a completely different business behind an identical figure.

Which is invisible in the ratio and obvious in the underlying counts, and it is the single strongest argument for keeping the counts rather than the percentage.

The same paragraph invites discussion of segments or other subdivisions where the registrant judges it necessary to understanding, which is the equivalent of splitting a rate by trip type.

Why the split matters is set out in the repeat rate piece.

JS-Outdoors, MN

Why is a guide benchmark impossible in principle?

Because the denominator is a business model choice.

A destination operation selling a once-in-a-lifetime trip has a structurally low repeat rate and is not failing, because the product is designed to be bought once.

A local trout guide working the same forty people has a structurally high one and is not necessarily succeeding, because a high rate on a shrinking base is decline.

A tourist-market charter is limited by whether people return to the destination at all, which has nothing to do with the guide.

And a group specialist has a rate that depends entirely on whether you count the organiser or the anglers, which is a definitional choice rather than a fact.

Which means the number is not comparable even in principle, before anybody worries about sample sizes or response rates.

Any published figure averaging across those four describes none of them.

How that averaging problem works generally is set out in the methodology piece.

Why do guides want the number so badly?

Because it feels like a report card, and report cards are comforting.

The appeal of a benchmark is not analytical, it is emotional: an operator working alone has no colleagues, no appraisal and no external signal about whether the work is any good.

A published figure appears to supply one, which is why an unsourced number circulates so effectively and why people are reluctant to give it up.

Recognising that is worth doing, because the underlying need is real even though the instrument is not.

What genuinely answers it is a series of your own figures moving in a direction, plus the specific feedback that comes from clients who return and say why.

Neither is as satisfying as a percentage with a comparison attached, and both are actually informative.

The substitute most operators reach for instead is comparison to whatever a neighbouring guide claims, which is anecdote with a competitor's incentive attached.

What does a high rate actually cost?

Growth, if it is the only thing being optimised.

An operation whose repeat rate is very high is frequently one that has stopped acquiring, since a base of loyal clients fills the calendar and removes the pressure.

Which works until attrition arrives, and attrition always arrives: people move, age, lose interest, or die, and a base with no inflow shrinks by a predictable percentage every year.

The rate stays high throughout, because it measures the share of a shrinking base that returned, which is exactly the wrong reassurance at exactly the wrong moment.

Which is another instance of the offsetting-changes problem, and it is the reason the absolute count of distinct clients matters more than the ratio.

An operation should watch the headcount as the primary number and the rate as a secondary one, which is the reverse of how it is usually presented.

A growing headcount with a falling rate is a healthier position than the reverse in almost every case.

Where the new headcount comes from is set out in the first clients piece.

Does the definition of a season matter?

More than people expect, and it is a further reason figures cannot travel.

A season is a calendar year for some operations, a spring-to-autumn window for others, and something spanning a new year entirely for anybody working a southern winter fishery.

Which means two operations computing a rate over a season are computing over different intervals, and the interval affects the answer directly.

A client fishing every fourteen months appears as a returner under one definition and as lapsed under another, without changing their behaviour at all.

Which is not a small technical point: for operations whose clients fish roughly annually, the boundary decides a substantial share of the classification.

Writing down what your season is, once, and never changing it, is the same discipline as writing down what a client is.

Both belong in the same sentence beside the number.

What comparison is actually available?

Your own prior years, and it is a better comparison than a benchmark would be.

A guide comparing this year to last year holds constant the water, the product, the market and the definition, which is more control than any cross-operation comparison could achieve.

Which means the absence of an industry figure costs far less than it appears to, since the industry figure would have answered a question nobody actually has.

The question people think they are asking is whether their rate is good, and the question that matters is whether it is moving.

Three years of your own figures answers the second completely and the first not at all, and the second is the actionable one.

Where a genuine external comparison is wanted, the only defensible version is a conversation with a comparable operation on comparable water using the same definition.

Which happens more often than outsiders expect and is worth more than any published figure.

What that definition should be is set out in the repeat rate piece.

What should be recorded to make it work?

Three counts and one field, kept identically every year.

How many separate people fished with you this year, how many did last year, and how many of that second group appear in the first.

Plus one field recording how many prior seasons each returning client has fished, which is the input that reveals the offsetting-changes problem above.

Without that fourth field the rate is a single number and cannot be decomposed; with it, the composition of the returners is visible.

Which is what turns a metric into a diagnostic, since losing long-standing clients while gaining second-year ones is an entirely different situation from the reverse.

All four come out of the trip records without any additional work at the time.

The only requirement is that the trip records identify people consistently, which is a spreadsheet discipline rather than an analytical one.

That discipline is set out in the spreadsheet CRM piece.

What about figures from platforms?

They measure platform repeat, which is a different thing entirely.

A booking platform can see whether a client booked through it twice, and cannot see whether that client booked directly the second time, which is what a satisfied client frequently does.

Which means a platform's repeat figure systematically understates retention for exactly the operations that convert clients to direct booking, being the better-run ones.

It also overstates it for operations whose clients never learn their name, since those clients have nowhere else to book.

Two opposite biases in the same figure, which is why the number cannot be corrected and can only be labelled.

None of which makes platform data worthless; it makes it data about platform behaviour, which is a legitimate and narrower thing.

What direct conversion does to all of it is set out in the channel share piece.

Does a low rate mean anything on its own?

Only in combination with the counts, and usually less than feared.

A rate that fell while the client base grew is the arithmetic of growth rather than a quality problem, and it is the commonest misreading in the whole subject.

A rate that fell while the base stayed flat is a retention question worth investigating.

A rate that fell while the base shrank is about demand, and retention measurement will not diagnose that at all.

Which is three completely different diagnoses from one direction of movement, and none of them is available from the percentage alone.

Which is again the disclosure rule's point: the reported figure requires the underlying reasons stated in quantitative and qualitative terms before it means anything.

Two counts and a sentence is the whole apparatus.

What to do about each diagnosis is set out in the win-back piece.

What would a credible figure look like?

Narrow, defined, dated, and with the counts published.

A defensible statement would name the fishery, the trip type, the definition of a client, the number of operations reporting, the number contacted, and the season.

Which is a substantial amount of description for one number, and it is exactly the description that separates data from an impression.

Nobody currently publishes anything approaching it for this trade, which is why no figure is offered here.

Where a group of guides on one water wanted to produce one, the exercise is genuinely achievable and would be the only defensible figure in existence for that fishery.

The barrier is not statistical sophistication but agreement on the definition, which is where such efforts usually collapse.

Agreeing the denominator first, in writing, is what makes the rest work.

The standards such an effort should meet are described in the methodology piece.

What about lifetime value instead?

A better question with the same data problem, and worth computing anyway.

The figure that actually matters commercially is what a client is worth across the whole relationship, which combines the return rate with frequency and with rate.

Which is computable from your own records the moment they hold trip counts and amounts per person, and which nobody publishes for this trade either.

Its advantage over a repeat rate is that it responds to all three levers rather than one, so an operation improving frequency sees it move where the repeat rate would not.

Its disadvantage is that it is backward-looking by construction and gets more so as the relationships lengthen, which means it lags every change you make.

Reading both together, with the counts beside them, is the working position, and neither in isolation is worth much.

Neither should be compared to anybody else's, for all the reasons above.

The arithmetic for it sits alongside the material in the numbers piece.

Does any of this apply to a new operation?

Not for two years, and pretending otherwise wastes attention.

An operation in its first season has no prior-season clients and therefore no denominator, which makes the metric undefined rather than zero.

In its second season the figure exists and is computed on a base so small that it swings enormously on individual decisions, which makes it noise.

By the third season it starts to mean something, and by the fifth the series becomes genuinely informative.

Which is worth knowing because new operators frequently agonise over a retention figure that is not yet measuring anything.

What is worth recording in those first years is the raw data, so that the series exists when it becomes readable.

Recording it from season one costs nothing and cannot be done retrospectively.

What else the first years should establish is set out in the first clients piece.

Where do benchmark comparisons go wrong?

Six ways, and the incomparable denominator is the first.

Comparing operations whose business models produce structurally different rates by design.

Reading a stable rate as a stable business, when offsetting movements can hide a complete change in composition.

Treating a fall after a growth year as a quality problem rather than as arithmetic.

Quoting platform repeat figures as retention, when they carry two opposite biases at once.

Using a figure whose definition of a client is unstated, which is the single choice that shifts the number furthest.

And computing anything whatever during an operation's first two seasons.

The full measurement method is set out in the repeat rate piece.

What is the working position?

No benchmark, your own series, and four fields recorded from season one.

Accept that no defensible industry figure exists and stop looking for one, because the search costs more than the answer would be worth.

Record the two headcounts, the overlap between them, and how many prior seasons each returner has behind them.

Compare only against your own prior years, holding the definition constant, and write the definition down so it stays constant.

Read direction rather than level, and read the counts alongside the ratio so that offsetting movements are visible.

Ignore platform repeat figures as a measure of your retention, since they measure platform behaviour.

Wait until the third season before drawing any conclusion, and record the data from the first anyway.

The registration-statement authority sits at 15 U.S.C. 77g, and the item is mirrored on govinfo.

The arithmetic itself is set out in the repeat rate piece.