Before Building a List, Look at Whether You Hold the Route of Delivery Yourself

About six months in, the same state keeps showing up. There are subscribers on the email list, and what to send them still is not settled.

The state presents itself as a shortage of material to send. At least one of the causes sits at a point before the gathering started. What has to be settled first in list building is not the method of gathering but what will flow along the route that lets you reach the people you gathered on your own timing.

That the order has been reversed is invisible from outside. Sign-up counts rise, and the risen counts stay in the record. Even with nothing settled about what flows, the work of gathering holds together completely. Holding together, it never produces an occasion to notice what was left unsettled.

Three things get answered here for certain. What the verb gather makes a contact address look like. Why lowering the friction of sign-up changes the character of the people who arrive, and not only their number. And where a borrowed route and a held route divide, without looking at any document of ownership.

Some things go unanswered. How many people to collect, how often to send, what to hand over in exchange. With the field being handled and where the people already are, the same numbers mean different things, so no figures appear here. No subscriber target appears either; what happens the moment a target is set is one of the objects this article handles.

The damage arrives on a delay. A route decays while it goes unused. The decay runs in two stages, and only the second reaches the indicator. In the first stage the row count does not fall, so the thinning goes undetected. By the time it is detected, nothing remains to be done for that person.

There is a second loss, and it comes from the gathering side. Lowering the friction of sign-up raises the number of sign-ups and changes the character of who arrives. What changed does not appear in the figures. What appears is a fall in response some months later. When it falls, the first thing suspected is the substance, so the gathering method never becomes an object of examination.

Stock and route are two readings of the same list. Read as stock, only the operations that increase it get planned. Read as route, the operations that maintain it carry the same weight. In front of the same figure, different work gets summoned: as stock, you go looking for the next way to gather; as route, you look at when you last delivered anything. And which one you are holding cannot be read off the row count or the unsubscribe rate. It can be read off what comes back.

List Building in the Order “Gather First, Decide What to Send Later”

The account in wide circulation treats gathering as an independent stage of work. First you gather, then you send something. At the gathering stage, the substance of what you will send is not asked about. Gathering can stand as an independent stage because collecting is placed ahead of what gets sent. What happens when those two swap places is taken up by relationship first, sale after.

Set the advice down accurately. List marketing means an approach built around a set of contact addresses. Because you can reach the same person repeatedly, you stop having to gather afresh every time.

The methods offered run roughly as follows. A sign-up field on your own site, sign-up in exchange for a document, answering a survey, registering for an event, direct sign-up from an advertisement. What gets handed over in exchange includes documents, discounts, videos and tools.

The advice that recurs is to lower the barrier to signing up. Fewer input fields, a stated reason for signing up, a concrete description of what you are handing over. All of them reduce the friction of the act of signing up.

And the phrase a list is an asset gets attached. Acquiring new people costs money; reaching people already on the list does not. That comparison is the grounds.

Every line of this checks out on the ground. Being able to reach people repeatedly is genuinely valuable, and lowering friction does raise sign-ups.

Nor is separating the stages wrong in itself. For a business that has settled what it sends and has moved into volume, making gathering independent is efficient. The slippage arrives when the gathering stage is made independent while what you send is unsettled.

The gathering work then closes on itself. More sign-ups means results; fewer means a problem with the gathering method. The side of what you send never becomes an object of examination.

One objection says you can decide what to send while you gather. There are things you learn from having readers. What you learn, though, is the response, not the problem you are handling. Responses come back to what you put out. About the range you have not put out, nothing comes back.

So deciding as you send is, in practice, a procedure for exploring the neighbourhood of whatever you put out first. What you put out first settles the range of the search that follows, and the correct answer outside that range does not get found by this procedure.

An independent gathering stage can also be handed to someone else. Handing it over is a legitimate option, and the handed-over stage returns no information about what judgement the arriving people made when they signed up. Only the count comes back.

Which Part of a Contact Address the Verb Gather Makes Look Like Stock

Gather is a verb used on objects. And objects carry a property that contact addresses do not.

Gathered objects sit on your side. Objects that sit there can be used or not used at your discretion. While unused, they remain. That is the property of stock.

Apply the verb to contact addresses and the same property gets assumed: what you gathered accumulates on your side and is available whenever you want. The word asset rides on that assumption too. Contact addresses do not carry the property.

What sits there is a state in which someone said you may send to them. Being a state, it can be released from their side. The release is unilateral and does not require your agreement.

Stock does not shrink; a state does. And the speed of the shrinking depends on what you do. Send nothing and the state is forgotten and effectively gone. Send too much and it gets released explicitly. A metaphor of stock cannot express a property that shrinks in both directions.

Does a larger base not mean more remaining, making gathering first the right move? If base and remainder were proportional, that reading would hold. They are not always proportional, because the method of gathering changes the character of who arrives.

Concretely, it shows up like this. Gather with the friction pushed all the way down and the act of signing up is not an expression of intent. A sign-up that is not an expression of intent does not function as the ground for later contact. The base grows, and the base is a collection of people who do not substantively hold the state you may send to me.

Keep using the stock metaphor and the operation of growing and the operation of maintaining get treated as separate things. Gathering campaigns and sending operations get planned as different work.

The split shows up in the allocation of resources. Money attaches to the gathering side and not to the maintaining side, because maintaining looks like just using what is already there. In fact, maintaining is what settles the total size of the routes, since what remains is what you gathered minus what decayed.

The metaphor has one more by-product. Stock does not change in value while it sits. With that assumption in place, the decision gather now, send later appears whenever there is nothing to send. What you gathered early decays for the whole stretch before you start sending. The further apart the gathering and the first send, the lower the proportion that survives.

There is a reason that decision gets chosen. Gathering can be started with nothing to send. Deciding what to send makes no progress until it is decided. The startable one gets started, and it produces a sense of progress. By as much as that sense of progress, the unsettled items get pushed back.

What You Hold Is Not a Name but a Route You Can Deliver Along

A list is the set of routes along which you can deliver information to someone at your own discretion. Redefine it as a set of routes rather than a roster of names.

Under that definition, what you hold is not the contact address. An address is an identifier of a route, not the route. Hold the identifier and nothing arrives if the route is not alive.

A route has two parties, the sender and the receiver. It holds only while both acknowledge that it holds, and it disappears when either steps off.

So owning a route differs in kind from owning an object. Owning an object is complete with one owner. Owning a route stands on the other person’s continuing agreement.

The difference changes how the thing gets handled. With objects, growing and maintaining are separate operations: buy and it grows, leave it and it is maintained. With routes, growing and maintaining are the same operation. Delivering is itself the act of maintaining the route.

You cannot manage this without treating it as a roster — the objection is fair. As a management tool, a roster is the only form available. The trouble is the shape of the tool moving into the shape of the judgement. Treated as a roster, whether the row count rose or fell becomes the indicator. Row counts do not represent the state of a route.

Row count and route state diverge most during a stretch when you send nothing. The count does not fall, and the routes thin. With the row-count indicator still, this thinning leaves no record. It surfaces later as the weakness of response the first time you send after six months.

The definition works one level further. Read as routes, what you gather changes. What you gather is agreement, not addresses. Gathering agreement requires something to agree to, and there the need for what you send to be settled first appears.

Agreement being the condition of a route does not mean that agreement alone gets things read. Agreement is permission to deliver, not a guarantee of being read. Permission and attention have to be procured separately.

Without that distinction, the response slips by one step. When response falls, the operation chosen is raising the permission side — the subscriber count. The attention side is what fell, and raising permission does not raise attention. The added permission accumulates as routes carrying no attention.

The total number of routes carries a practical ceiling too. Maintaining takes hands, so grow past the number you can maintain and the excess goes to decay. Held as stock, the ceiling is invisible. Invisible, planning only ever runs in the direction of growth.

What Lowering the Sign-up Friction Does — When Cost Separates Types of People and When It Does Not

The point about friction has backing from the side of economics.

In 1973, the economist Michael Spence formalised as signalling the mechanism by which some other act stands in as a cue for someone who cannot check the substance directly (Spence, 1973, The Quarterly Journal of Economics, 87(3), 355–374). The mechanism works under one condition: the cost of the act must differ by type. Type here means the division between people who care about the problem enough to read and judge, and people who do not.

If the cost is the same for everyone, the act conveys nothing, since anyone can perform it identically. When the cost differs by type, those who could show it and those who could not divide. Dividing is what produces information.

It does not follow that sign-up barriers should be high. Spence handled the labour market, a bounded setting, and it does not transfer directly. What transfers is the condition that a difference in cost is what separates types.

Apply it to sign-up. Adding one input field imposes a cost on everybody equally. The effort falls the same on someone with intent and someone without. So this friction does not separate types, and lowering or raising it does not change the character of who arrives.

Having a piece of writing that has to be read and judged before signing up imposes cost differently. For someone interested in the problem, reading is part of the point; for someone uninterested, it is pure burden. With the cost differing by type, this friction separates.

The test is not the height of the friction but what the friction confirms. Both get called friction with the same word, and they are different in kind. The advice to lower the barrier works on the first, and lowering the second along with it removes the separating work itself.

None of this says higher friction is better. Set it high and people with intent fail to complete sign-up. Those people are not observable from your side, so the loss from over-raising friction never appears in the figures.

And separating friction carries a cost on your side too, because it requires putting a piece of writing worth reading and judging in place first. That cost falls not on the gathering stage but on the stage of deciding what to send. With what you send unsettled, there is nothing to put there. The order of settling what to send first connects to the design of friction here.

A Borrowed Route and a Held Route Divide on Two Points

The distinction usually gets explained as a matter of ownership — whether the thing is yours. The criterion that works in practice sits elsewhere.

The first criterion is who settles the conditions of arrival. Of the people signed up, how many receive what you put out. If someone other than you settles that proportion, the route is borrowed.

The second criterion is whether you can observe what failed to arrive. Some routes record the count of failures and the reasons for them; some do not. Where they are not recorded, what remains with you is the single fact that there was no response.

The second works harder in practice, because it settles where the inference about cause will go.

With failures recorded, the reason for no response divides in two: it did not arrive, or it arrived and was not read. Divided, the next move divides too.

Unrecorded, the two collapse into one. Infer a cause from the collapsed state and the inference tends to run toward the substance. The material was poor, the timing was poor, the writing was poor. The same conclusion comes out even when the route is where things stopped.

Does not using places where many people already are mean giving up on reach? Not using them is not the point. The point is separating using something as an entrance from holding it as a route. Places with many people make excellent entrances, because the other person is already there. That separation is handled in customer journey design.

Even a held route does not guarantee arrival. The receiving side’s settings can prevent it. Held means you can tell that it failed to arrive, not it always arrives.

Neither criterion can be settled by looking at a document of ownership. Even in a system you contracted for yourself, if the system settles the conditions of arrival, that part is borrowed. Conversely, on someone else’s ground, if the count of failures comes back to you, the observation on that part holds.

So a route is more accurately held as a per-item state than as a binary of borrowed and held. The conditions of arrival are borrowed and the record of failures is held. That combination genuinely occurs.

Taken as a binary, the only decision available is migration: all borrowed, or all held. Taken per item, the decision becomes moving only the item where the loss is currently largest. Moving in items rather than all at once is what actually progresses.

What You Handed Over in Exchange Stays as the Standard for Judging

Handing something over in exchange for sign-up is a widely used method. And what you handed over does not end when it is received. It stays as the standard for later judgement.

The reason the method works comes from the condition that a difference in cost separates types. At the moment of signing up, what will arrive afterwards is unknown. Unknown, the judgement runs on the material at hand, and if what is handed over in exchange has been described, it is the most concrete material available.

That is where the method’s rationality lies. There is one consequence.

Someone who judged on the exchanged item goes on judging by the quality of that item. If the first thing received was good, the same level is expected next. The baseline of expectation gets fixed onto the exchanged item.

This works hardest when the quality of what you hand over is high. Hand over something excellent and that level becomes the baseline, and ordinary sends fall below it. At the moment they fall below, the route begins to thin.

Should you then hand over nothing in exchange? That is not it. That is not it. What matters is that the exchanged item and what arrives afterwards be of the same kind. Of the same kind, the exchanged item works as a sample, and someone who judged on a sample goes on judging the same way about the body.

Of a different kind, the exchanged item is not a sample but a payment. Someone who received a payment holds no reason to maintain the route once the payment has been received in full.

Whether the kind is the same is settled not by format but by whether the material available for judgement is the same. A document and a piece of writing differ in format; handling the same problem from the same angle, they are of the same kind. Match the format while handling a different problem and the kind differs.

Design the exchanged item as a sample and a ceiling appears on operations to improve it, since you cannot hand over something better than the body.

At first glance that looks like a constraint. In fact it works as a test. When the exchanged item has become better than the body, it is not a sample but a separate product. Giving away a separate product for nothing is a sign that the position of the body has not been settled.

The structure carries one more consequence. In a state where what arrives is unknown, everyone enters at minimal involvement at first. Low involvement just after sign-up is not a failure of the gathering method. It follows structurally. The problem is a design that does not presume low involvement. Put out material assuming high involvement from the start and the route begins with the assumption and the reality already apart.

The Three Things to Settle Before You Gather — the Problem, the Angle, and What Accumulates

What has to be settled first is what will flow along the route. Concretely, three things.

First, whose problem, and which one. With that unsettled, the way you select what to deliver changes every time. Changing every time, the receiving side cannot predict what comes next. A route that cannot be predicted drops in priority for attention.

Second, from which angle you handle the problem. Handle the same problem without a settled angle and the material drifts toward general information. General information is available elsewhere, and a route carrying what is available elsewhere gets substituted by another route.

Third, what accumulates for someone who keeps receiving. On a route carrying only information that closes each time, no number of receipts changes the person’s state. Unchanged, there is no reason to keep receiving.

All three settle before you gather. What you gathered while they were unsettled remains, once they are settled, as people who do not fit.

Settling early makes later change impossible, runs the worry. Change is possible. People leave when you change, and that is all. People leaving is not in itself a bad event. The problem is that the reason for leaving is unobservable.

Change the material without stating the change and departures get recorded as the material was poor. State it and they get recorded as the problem being handled changed and it stopped fitting. The same departure, and a different next move.

You do not need all three fully settled before starting. Full settlement comes after you have started delivering. What is needed is knowing which items are unsettled. Knowing that, you can read the responses that arrive later as information about those items.

The third is the one that tends to be skipped. Whose problem and from which angle get examined reasonably well. What accumulates for someone who keeps receiving often does not.

The reason is plain: the work of producing each instalment contains no step for thinking about accumulation. Closing each instalment is lighter for the writer, because it holds for someone who has not read the earlier ones.

And material closed instalment by instalment does hold. Holding, it never surfaces as a problem. What surfaces is that no difference appears between someone reading for a long time and someone who just started. With no difference, there is no reason to read for long. This structure is common to anything intangible, and is handled in detail in where the value of an intangible product resides.

Make the Subscriber Count the Target and the Number Eats the Route

Set the subscriber count as the target and see what happens. Numbers are measurable. What is measurable becomes a target easily, and what becomes a target gets optimised.

The operation that maximises the subscriber count is the operation that lowers the friction of sign-up. And if what gets lowered is friction that separates types, the separating work goes with it.

Push the friction all the way down and signing up stops being an act of deciding whether to read. It becomes an act of receiving what is offered. Because it completes at the moment of receipt, what arrives afterwards was never part of the sign-up judgement.

Here the quality and the number of routes point in opposite directions. The operation that grows the number thins the character of the routes, and the thinned character stays off the figures, showing up only later as a drop in response.

When response drops, the first suspect is the substance. Rebuild the substance and, with the gathering method unchanged, the character of who arrives stays the same. Rebuilding repeats, and gathering never gets examined.

As the number grows, reading one person at a time also stops being possible. The problem is the case where you never read one before it stopped being possible. Having read, you know what sits behind an aggregate figure. Not having read, figures are figures from the start. The structure by which the stage you measure becomes the object of optimisation is handled separately in cost per acquisition.

None of this says do not look at the number. The number shows whether the routes are widening. It says do not make it the target. Set as a target, everything other than the number falls out of the scope of adjustment.

With nothing else measurable, the number is all you have — this counter holds together. Other things are measurable. Counts of replies, counts of replies that come back in words you did not use, the level of response after a deliberate gap. All countable. They are harder to count, not impossible to measure.

Select indicators by ease of counting and the same indicators get selected every time. Subscriber count, open proportion, click proportion. These aggregate automatically, so selecting them costs nothing. Look only at what costs nothing and the range you are looking at gets settled by the convenience of your own measurement.

Set the number as the target and the shortest path to the target gets chosen. The shortest path is usually the removal of friction. And the removal of friction has no floor. What is left after it has been pushed all the way down is invisible from the side of the number.

Routes Decay in Two Stages, and Only the Second Reaches the Indicator

The decay of an unused route runs in two stages.

The first stage is memory. The fact of having signed up gets forgotten. Arriving in that state, the thing gets treated as something unrecognised. At this point it is arriving and not being read.

The second stage is judgement. Something unrecognised arriving repeatedly gets unsubscribed from. The unsubscribing is an explicit act, and at this point the route is gone.

The two stages matter. In the first stage, nothing has happened on the indicator. The row count has not fallen. Only in the second does the figure move. By the time the figure moves, the first stage is already complete.

So the response to rising unsubscribes is not an operation that reduces unsubscribes. You have to go back to the first stage. The way back is not raising frequency but making it legible, on every arrival, who this is from and what it is for.

Does raising frequency raise unsubscribes? Sometimes it does. What raises them, though, is raising frequency alone while it stays unclear who the material is for. Material with a clear direction rarely becomes a reason to unsubscribe at high frequency. What is being judged is not frequency but whether the substance is predictable before opening.

Predictable does not mean identical. Write the same thing every time and prediction works while the reason to read disappears. What should be predictable is the problem and the angle, not the conclusion.

Decay running in two stages means the window for response divides in two as well. Inside the first stage, making the character of what arrives legible brings people back. Someone who has entered the second does not come back, because unsubscribing cannot be reversed from your side.

So the unsubscribe rate is always a record of being too late. By the time it is recorded, nothing can be done for that person. It nonetheless remains the rate people watch, because it is one of the few decay indicators that aggregate automatically.

Catching the first stage means looking at something else. Arriving and not being read shows up as a slow fall in the proportion opened. Not the movement instalment to instalment, but the slope across some months. A slope does not emerge on its own, so unless you deliberately line the figures up, it is invisible. An invisible indicator gets treated the same way as an indicator that does not exist.

Read Whether the Route Is Alive from the Substance of What Comes Back

Whether a route is alive appears in neither the row count nor the unsubscribe rate. It appears on the side of what comes back. Three places carry it.

First, whether replies arrive to what you put out. The count is irrelevant. Even one means the route is recognised as two-way. A route with none is, from the receiving side, a one-way notice board. A notice board is a format nobody misses once it stops being read.

Second, the vocabulary the other person uses about what arrived. Words of yours coming straight back mean the substance passed through as information. Words translated into their own situation mean something is accumulating along the route.

Third, the response after a stretch of not sending. Send after a deliberate gap and, if the response matches what it was before the gap, the route is held on the side of memory. A clear fall means what was held was frequency, not a relationship.

Row counts appear in none of the three. What they look at is what is happening at both ends of the route. Given that a list is a set of routes of delivery, the route is what needs checking.

That is also why row counts are kept out of the main indicators. A row count turns two different states into the same figure. Someone who signed up with intent and keeps reading, and someone who received what they came for and simply stayed, count as the same row. Two routes raise the row count, and one of them thins the character of the set.

The ease of replies differs entirely by field, and on that point the difference is real. It does. The thing being looked at, though, is not the count but whether any arrive. Even in a field where replies are rare, zero and one differ in kind. If not one has ever arrived across a long stretch, the way you are writing is not built to expect a reply.

The three places are settled on a handful of actual cases. They are not an exercise in producing proportions, so they hold at small scale. As scale grows, reading one at a time gets harder. Read a few while it is small and the resolution changes when you look at aggregates later. What to hand over first after gathering overlaps with the structure handled in the design of a trial offer.

Read this looking for how many people to aim at and you will not find it. Frequency and the substance of what you hand over in exchange get the same treatment. What a headcount means shifts with where the people already are and with what flows along the route. The same thousand people are an asset in one case and a liability in the other.

There are subscribers, and there is nothing obvious to send. What is short in that state is not material to send but the three things that should be settled before the gathering began.

The material sits in records you already have. Read the replies to what you sent before by their substance rather than their count. Are your own words coming straight back, or words translated into that person’s situation? If only the first, what is travelling along the route is information and nothing else.