When Customers Do Not Come, Divide the Visible Reasons From the Invisible Ones

Another month has ended with no enquiries. You are publishing. The site exists. The price is not out of line with the going rate. You hunted for what was missing and fixed everything you could think of. And still nobody makes contact.

In a stretch with no enquiries, the first explanation reached for is that you are not known yet. It has not reached anyone yet, nobody has found it, awareness is short. In the early period after going independent, that explanation genuinely accounts for the facts. Because it accounts for them, when several months pass with no contact, what gets doubted next moves across to what you supply. The quality is not good enough, in that form.

Before any of that, something can be separated. The same result of nobody coming emerges out of three different states. Not found. Found, but with no reason to come. Tried to come, but found no route. And of these three, only the first leaves a record without you doing anything. Unless you have prepared a place that keeps records, the remaining two appear as the same shape — nothing happening — however many of them there are.

So this is not a state in which the cause is unknown. It is a state in which the cause cannot be narrowed to one, because several causes can be derived from the same observation. If it is merely unknown, more information brings you closer. If it cannot be narrowed, no quantity of the same kind of information brings you closer. What is needed before acting is not more information but a repair to the asymmetry that leaves records on one side only.

For that reason, no list of methods for attracting customers appears here. Which medium to publish on, what to write, what to charge — every one of these is chosen after it is settled which of the three is occurring. Chosen before, the work simply accumulates with no way of confirming afterwards whether it landed. Why the person did not come is placed outside the range as well. What can be separated is the position — which stage the person who did not come was standing at — and what they were thinking as they left remains even after the position is known.

What gets placed is the content of the three states, the mechanism by which two of them never appear in observation, and how the ones that do not appear are handled. First, the range the advertised responses reach, and the fact that several causes can be derived from one result. Then the structure by which records accumulate on one side only, and the way that bias widens with the years. Finally, the difference between work not arriving and work not being won, the case of a person who decided to ask and left, and the handling of what cannot be narrowed to one — closing with how to check that the states have been separated.

Three Separate States Surface as One Result

Behind the single result of no enquiry occurring lie three separate causes.

The first state is not being found. Your existence has not entered the other person’s field of view at all. The occasion for being judged has not yet arisen.

The second state is arriving in front of them with no reason to move. They looked, they read, and then decided not to make contact. A judgement was made, and its result was no.

The third state is trying to come and finding no route. They wanted to commission the work but did not know how to ask. Or they could not see how long a reply would take after making contact, and stopped partway.

From your side, the three look identical. They appear in the same shape: nothing happening. The operations they require are nevertheless different. The first needs exposure, the second needs material for judgement, and the third needs the procedure put in order.

The second and third look as though they can be folded into a single phrase: not appealing enough. Fold them, and the distinction disappears. The second is a state in which someone understood the content and decided that this time it was not for them. Not everyone commissions work, so this is a normal outcome. The third is a state in which the decision to commission has already been reached and the procedure stopped it. The former was not chosen; the latter was chosen and never completed. Call both by the word appeal, and the prescription that comes out for the latter is to improve the content.

The three hold at the same time. One layer has not found you, another layer has judged and declined, and a small number are stuck in the procedure. A mixture of all three is not unusual.

The proportions are not fixed either. If you have just started, the first is large; several years in, the share of the second and third rises. Even so, what circulates as a response stays weighted towards the first. One likely reason is that many people searching for this information are at the early stage, so the distribution of demand leans that way.

That the proportions shift has one further consequence. A response that worked three years ago and does not work now has not necessarily gone out of date. The same response may still be working on the first exactly as before, while your own principal cause has moved across to the second and third. From the observation that it stopped working, both readings can be derived.

The phrase no customers are coming does work of its own as well. It folds the three together and then locates the cause in the value of what is being supplied. Treated as a question of value, the prescription becomes supply something better. When someone in the third state carries this out, the content improves while the state persists — already chosen, with the procedure still not in place.

The Seven Prescriptions Put the Cause on the Not-Known Side

The widely advertised responses come to seven, and most of them work on the first state.

Start by assuming awareness is short. Clarify who the work helps most. Convey what the other person is struggling with and what changes through you. Register on the places where people already come to look. Research the words the other person searches with, and write the answer to them. Publish actual cases in the form of a difficulty and a result. Since several causes are entangled, act on each of them.

All seven function as practice. The prescription of researching search terms and writing the answer is a direct operation against the first. Publishing cases works directly on the second, since it hands material for judgement to someone short of exactly that. The observation that several causes are entangled is structurally accurate as well.

With that granted, look at the range the seven address. Awareness, clarity of audience, conveying value, registration, search, cases. Almost all of them raise the probability of being found. And only the publishing of cases reaches the second.

No prescription among them addresses the third — the procedure for making contact itself.

The reason can be inferred. The third leaves no record. Seen from the side writing the responses, a phenomenon that cannot be observed does not become the object of a prescription.

If awareness increases, the second and third will increase in absolute numbers too — that view holds. They will. And what is lost increases in proportion with them. Improving the first does not change the share held by the second and third. Widening the entrance does not clear a blockage at the exit.

That does not put the seven wide of the mark. Where the first is the principal cause — where you have just begun and nobody knows of you yet — they are squarely within range. The drift starts after that: for someone several years into independence who has kept publishing, the likelihood that the first is the only principal cause falls.

The observation that several causes are entangled retains a shape that is hard to use as a prescription. Once the conclusion becomes act on everything, the ordering disappears. Repairing downstream while upstream is blocked does not move the figures, so acting simultaneously and acting in sequence are not the same thing. A prescription that lacks order is coming out of an accurate perception.

Carrying out this group of prescriptions increases the volume of work. More places to register, articles to write, cases to compile. And when the results do not move against the added work, the conclusion becomes still not enough. Prescriptions that work on the first are shaped so that they can be continued indefinitely, so a point of completion is unlikely to arise.

The Same Result Can Be Derived From Any of the Three Causes

The observation nobody comes contradicts none of the three causes. So stacking that observation up never narrows the cause to one.

This form of problem has a name. In a 1949 paper, the economist Tjalling Koopmans dealt with cases where the data in hand is the same while several different underlying relations can hold behind it (Koopmans, 1949, “Identification Problems in Economic Model Construction”, Econometrica, 17(2), 125–144). When an observation does not point at a single answer, what is short is not the quantity of observation but the number of observations that can tell things apart.

The fact that enquiries came to zero this month comes out as the same zero whether nobody saw you at all, or a hundred people saw you and every one of them declined, or five people stopped just short of sending. The figure zero does not distinguish the three states.

And continuing an observation that does not distinguish them for another year simply lines up twelve zeroes. The same kind of observation never reduces the branching, however many you collect.

With experience you come to sense it — that view exists, and it holds up to a point. But the material that sense is using is also the record from the side that reached you. About the side that did not reach you, not a single item of material is added however many years pass. What the years increase is the quantity of material on one side.

Nor does any of this make increasing observation pointless. Even with the same kind of observation, greater quantity raises the precision of the ratios. What it does not raise is the branching — which cause is at work. Precision improving and causes narrowing are not linked.

Mix the two, and in practice the judgement that comes out is wait and see a while longer. Another three months and the trend will show, in that form. The trend will show. What shows is the trend of an observation that cannot tell things apart. Twelve zeroes becoming twenty-four zeroes leaves the branching at three, since each of them can come out of any of the three states. Waiting resolves the question of precision only; the question of identification is not resolved by waiting.

What makes it worse comes next. When observation grows while the branching does not shrink, the explanation settles on whatever is most familiar. Either the first, which has records in hand, or the quality is not good enough, which requires no records at all. Neither was derived from the observation. They remain because the observation did not rule them out.

Records Are Kept on One Side Only

Why is only the first visible? It follows from an asymmetry in measurement.

Impressions can be counted. Visitors to the site can be counted. These are events that leave a record on your side. And what leaves a record rises into awareness as an object of improvement.

By contrast, the judgement of someone who looked, read, and decided not to make contact leaves no record. The very fact that a judgement occurred never reaches you.

Less recorded still is the person who meant to ask and stopped. These people came right up to the point of contact. Even so no record remains. The fact of stopping partway is written down nowhere.

So the distribution of what becomes an object of improvement and the distribution of what is actually happening are misaligned from the outset. Only the first is visible, and the second and third stay out of view however numerous they are.

It is not that no means of measurement exists at all. How far into the screens someone got before leaving can sometimes be captured as a record. Being able to capture it and not looking also happens. This asymmetry is less a constraint of the technology than a question of where things are placed.

On the other hand, being unmeasurable and being nonexistent are different things. Obvious as the distinction looks, it gets confused at the moment of judgement. The inference that nobody can be looking, because there is no response, erases the possibility of the second — that they looked and judged. From the fact of no response, being unseen and being seen and declined follow equally well.

There is no point worrying about people who did not look — that can be said. About individuals, there is genuinely nothing to be done. What can be known is the distribution. If you know at which stage people are dropping away, the place to repair can be identified. The individual reasons stay unknown while the position of the blockage becomes known. And since it is the position you repair, not knowing the individual reasons does not obstruct the work.

The voices that gather in your hands have the same property. The people who write impressions are the people who commissioned the work. From those who did not commission it, not one reason arrives. However many client voices you collect, the reasons for not coming never come out of them. How that bias distorts observation as a whole is treated squarely in how to read the drop-off rate, and what happens further at the stage of publishing the voices you collected in how client testimonials are used.

The Side With Thicker Information Becomes the Candidate Cause

This asymmetry widens with the years the business continues. That is why it does not resolve on its own if left alone.

Someone who has published for three years has three years of impression records in hand. Over the same period, the records of people who looked, judged and left number zero. Accumulation happens on one side only, so as time passes the information on the visible side alone grows thicker.

And the side with thicker information carries weight as material for judgement. Figures are lined up, movement over time can be seen, comparison against last month is possible. On the opposite side there is nothing to compare against at all. Lined up as material for argument, one side is concrete and the other is blank.

Consider what is the cause in this state, and the answer comes automatically from the side that has the concrete material. Candidates form not in order of likelihood but in order of how complete the material is.

Ignoring the side with no material is not what happens. That side never rises as a candidate at all. And what never rises is never rejected either, so no record is kept of whether it was considered. Ask the same question a year later, and whether you examined that possibility last year cannot be found anywhere in your own files.

It happens again on the side of whoever you consult. Explaining your situation to another person, what you can hand over is only the side with records. The trend in impressions can be shown; the count of people who judged and declined cannot be shown at all. Whoever receives it advises within the range of the material handed over. So the advice also tends to lean towards the first. This is not the ability of the person advising but a result produced by the shape of the material handed to them.

Material quantity skewing the judgement is not a matter of the judgement being careless. A comparison between something concrete and a blank is not a comparison in the first place. The blank side, having nothing to compare with, falls out before it reaches the ground of consideration. That it fell out is not recorded either, not even as a falling out.

This bias is not a problem of recording technology. Impressions can be measured because that event occurs on your side. Judgement occurs on the other person’s side, so there is no reason for a record to arise on yours in the first place. Introducing better measurement leaves this asymmetry standing as a matter of principle. What can be reduced is only the portion of judgement moved across to a location on your side.

As the years pass, a further element joins the bias. Improving what is measurable has cost effort every time. With three years of effort piled on one side, the explanation that it is still not enough looks more natural than doubting another cause. The quantity of effort invested becomes, as it stands, the persuasiveness of the explanation.

This bears directly on how time gets used. Improvement of the measurable advances while the unmeasurable stays untouched. And since the untouched side is not even recorded as untouched, the following year proceeds on the same allocation. That is what it means for something not to resolve if left alone.

Work Does Not Arrive and Work Cannot Be Won Are Different Problems

These two are used interchangeably in ordinary conversation, and the stage that is blocked differs.

Not arriving is a state in which the enquiry itself has not occurred. It corresponds to the first or the third.

Not being won is a state in which enquiries arrive and do not conclude. It corresponds to the second, and the material for judgement has already been handed over.

The operations required are opposites. Polishing the content of a proposal when work is not arriving leaves you with no occasion to propose at all. Increasing exposure when work is not being won simply increases the enquiries that do not conclude.

And the state of not being won is shaped so that the cause is easier to identify. Contact has occurred, so the reason for the refusal can sometimes be asked.

The reason you can ask for is not necessarily the real one. The person declining chooses a reason with no edges. The budget does not fit, the timing does not fit, it was not agreed internally. These serve as the standard forms of refusal, so they carry little information about content. What can be read from the exchange is less the reason itself than the position — how far the consideration had progressed.

Inside not being won there are two types as well: the case where the conditions did not fit, and the case where the material for judgement was short. The former is a normal outcome with nothing to repair. Not concluding with someone whose conditions do not fit is the result of selection working. Only the latter is something that can be repaired. And these two do not separate under questioning either, since a person short of material for judgement also says the budget does not fit.

The distinction is needed because handling them folded together produces opposite moves. When people whose conditions do not fit keep arriving, what is required is not more material for judgement but a change in who is reached. When people short of material for judgement keep arriving, changing who is reached produces the same thing again. The same result — not being won — corresponds to opposite operations.

Loosening the conditions is a route: work will be won. And the content of the work won drifts in a direction other than the one you intended. Loosening conditions to raise the count solves the problem of count and creates a problem of content. It is a judgement about which to take, not an improvement.

Of the prescriptions in circulation, those addressing not being won are the more numerous. How to write a proposal, how to present a price, how to respond to a refusal. These are techniques for after contact has occurred, so records remain and they can be written as a reproducible procedure. That prescriptions for not arriving are scarce reflects less a difference in importance than a difference in whether writable material was left behind.

Keep the two folded together under the name not arriving, and the response alternates. Work on increasing exposure, work on polishing the proposal, switched every time results fail to appear. And since both stop partway, neither effect ever gets tested.

People Who Decided to Ask Leave Before They Send

The third is a state in which the intention is settled and the handover procedure stops it.

The conditions are several. The means of contact sits somewhere hard to find. What to write and send is unclear. What happens after sending is not stated. The time until a reply is not indicated.

None of these has anything to do with content, and none takes long to fix. They go unfixed because the people who left at this stage never appear in observation.

Obstacles of this kind have a name on the side of behavioural economics. In 2018 the economist Richard Thaler drew a distinction between design that nudges a person’s choice along and the pointless obstacles placed on the side of carrying that choice out, which he called sludge (Thaler, 2018, “Nudge, not sludge”, Science, 361(6401), 431). What the distinction turns on is that the obstacle does not change what the other person wants. The judgement to ask stands, and only the execution stops.

So for a person at this stage, improving the content does not move the count. What has to move sits on the side of the procedure rather than the content.

Structurally it arises where a distance separates judgement from execution. When several decisions intervene between deciding to ask and actually sending, an opportunity to leave is created at each one. What to write, how much to write, whether now is a reasonable moment. These are small decisions, and to the extent they are not settled, the act of sending gets postponed.

That distance can be reduced by the design on the receiving side. It means showing in advance what needs to be settled before someone sends. The work of settling it is done once, before any commission arrives, and is not repeated afterwards. How far that can be brought forward inside a single email is treated in designing the automatic reply.

The distance cannot be made zero. Commissioning work involves decisions on the other person’s side as well. What can be reduced is only the portion arising from what is unsettled on yours. Cut into the other person’s judgement itself, and the operation becomes one of cutting away their room to judge.

The number of unsettled items and the number of places where a departure can occur correspond. The more items are unsettled on your side, the more points are handed to the other person to decide. The single phrase please get in touch leaves the content, the scope and the cost of the enquiry all three for the other person to determine. Count the items you have not settled, and you have, as it stands, the list of places where departures occur.

Someone who really wants to ask will make contact however hard it is to find — that view exists. People who need it strongly do get through. The ones who do not come are those at the comparison stage. Where people at the comparison stage are numerous, a design that only the strongly motivated reach fixes the ceiling on count at that point.

Repair the third and the count of enquiries rises. Among the additions are enquiries whose conditions do not fit. The obstacle had been working as an unintended filter. Reducing the distance weakens that filtering at the same time.

Not Narrowing to One Is Not the Same as Not Knowing

There is only one operation that reduces the branching: making what came out in the same shape come out in separate shapes.

Increasing the quantity of observation does not reduce the branching, but increasing the number of kinds of observation by one can cut away one of the three. If impressions are zero, the second and third cannot have occurred. If impressions exist and nobody reached the explanation, the third is gone, since not one person came as far as the point of sending. How many you can cut away is determined by how many kinds of output you hold.

Presence is what is working here, not precision. The observation that ratios cannot be read while the counts are small is correct. Even so, the difference between zero and one cuts branches at small counts. The gap between ten people and twelve cannot be read; the gap between zero people and ten can be. A crude discrimination is still usable for locating a position.

Increasing outputs does not necessarily mean building something new. Placing the same explanation in a location that keeps records or one that does not changes whether it can be verified afterwards. With identical content, where it is placed decides whether the number of kinds of output rises by one. Which stage to place it at so that it becomes countable is treated in how to read the drop-off rate, and the order of confirmation running from upstream downwards in when the conversion rate is poor, where to start doubting.

Knowing the position and knowing the reason are different. Where someone left can be read from the outputs; why they left cannot. The reason becomes a guess, and testing a guess means changing something and comparing. What can be reduced here is the number of branches, and the reason itself lies beyond that.

Even so, settling the position changes the quality of the guess. The candidate reasons differ entirely from stage to stage. The guesses that stand at the stage of being shown and not visiting, and those that stand at the stage of reading the explanation and not sending, share almost no items. Guess at reasons without settling the position, and candidates from both stages line up mixed together. Choosing from a mixed set of candidates leaves no way to judge afterwards whether the choice landed.

Nor can every judgement be moved into a location that keeps records. Introductions made face to face, and judgements completed inside the other person’s head, cannot be recorded by their nature. What can be moved is only the portion whose location you get to choose. And moving even that portion lets you estimate the scale of the unmovable portion by subtraction.

Increase the kinds of output and the poor figures become visible. Departures that had gone unobserved come out in front of you as counts. The fact that things were easier to bear while they were invisible can become the reason for stopping the counting. Separating the states carries this cost.

Once the States Can Be Separated, Three Changes Appear in the Records

What is confirmed is not the count of enquiries but three changes that appear on the side of the records.

First, look at whether this month’s result can be assigned to one of the three.

If it cannot, the kinds of output are short. Whether the count rose has no bearing on this test. Zero enquiries is fine, so long as you can say it was zero-from-zero-impressions; two branches have already been cut.

Second, look at whether the place repaired and the output that moved coincide.

If they coincide, the inference about the cause was correct. If a different output moved, what worked was not the reason you inferred. Confirming the coincidence bears directly on how often the next inference lands.

Third, count how many places remain where judgement occurs and no record is kept.

The lower that number falls, the higher the probability of being able to assign next month’s result. The remaining number is itself the width of what you can currently diagnose.

None of the three tells you the reasons of those who did not come. What can be known is which stage the person who did not come was standing at. The reason side requires different material.

Which of the things you fixed actually landed is still unknown. It is unknown not because the fixing was careless but because whether it landed never appeared on the side of the outputs. The result nobody comes has the same shape before the repairs and after them.

As operations of increase, the advertised responses and what has been placed here are no different. What differs is the object increased: one raises the count of people arriving at the entrance, the other raises the number of kinds of output kept in your hands. With the count of people, whether it rose is known on the spot. With the kinds, from the month after the increase, the same nobody comes starts to appear in separate shapes.

When the next month ends with nothing happening, what can be read from that month is not decided after the month ends. It has already been decided by how many kinds of output you held during it.

The specifics of the diagnosis that doubts from upstream downwards are treated in the diagnostic procedure for why nothing sells, and the design of holding several entrances while narrowing the confluence to one in customer journey design. Identifying which stage of the route is blocked is handled squarely in when attraction fails, what is lacking is not technique but a route, the structure of mistaking the figures of a place for the figures of a business in followers not growing is not a problem of the business, and the conditions under which publishing itself does not last in blogging not lasting comes from the difference in how fast things come back.