Being at the Mercy of Information Comes From Not Having Selected the Entrance

Every time you learn a new method, what you are currently doing starts to feel wrong. You see someone’s results and change direction, read another article and change back. The more you look into it, the less clear it becomes what to adopt. You are gathering information, and the number of things that get settled keeps falling.

This word folds together three separate states, and only one of them is a problem of volume. The one that goes unhandled to the end is the form in which, every time information arrives, the order of what matters gets rearranged. With the same material, a different order produces a different answer. Against this form, reducing the volume only raises the weight of each single item; the state itself does not change.

And the rearrangement of that order is not experienced as a change, because you view the world through the order that came after it. What you notice is having changed direction. That the order moved is not noticed. This is why the state lasts.

The question is a single one: who is moving that order. A selection always occurs at the entrance to information, and where you do not make it yourself, it is not that no selection happens but that something else makes it. One thing that performs it is a system optimising for volume of response. Even so, within what has been measured, the size of that delegation was smaller than it is imagined to be, and much of the narrowing came from the entrance you selected yourself. That is not bad news. It means what can be moved sits on your side.

Being able to move it yourself also means that nobody else will. While you are holding the system responsible, the list of who you follow stays as it was.

How the order itself takes shape is not pursued here. That it stands up afterwards as a record of judgements you placed is carried by Your own axis is not found but placed again. What is examined here is the route by which that order gets moved from outside. The two are two sides of one thing, and closing one leaves the other standing. Close the route and the order does not build itself; hold an order and it still gets rewritten while the route stays open.

The path runs as follows. The three states this one word names get separated, and the reach of the advertised responses gets checked. On that basis, the effect of the number of options on decisions is treated, along with the limitation found in later verification, and the mechanism by which the order gets rearranged is set out. Finally, how far the delegation of the entrance actually extends is examined against measurement, followed by the difference between cutting off and designing, how this shows on the business side, and how to confirm it.

What “At the Mercy of Information” Folds Together Is Volume, Truth and Order

Being at the mercy of information covers at least three separate forms.

The first is where the volume exceeds what you can process. What comes in cannot be read through, and it accumulates unhandled. The load of not keeping up is the main symptom.

The second is where the truth of the information cannot be assessed. Contradictory claims stand side by side and which is more likely cannot be told. The material that would settle it is not there.

The third is one where direction moves every time information arrives. Neither volume nor truth is the problem; each new thing changes what should take priority. Sometimes you notice the movement and sometimes you do not.

From outside the three are indistinguishable. Each produces the description of being at the mercy of information. The remedies, though, are separate. The first is a problem of volume, the second of verification, the third of order.

Reducing the volume does not lighten all three. It lightens the first. The third occurs independently of volume. Read three articles a month, and if direction moves each time you read one, the state is unchanged. If anything, the fewer the items, the larger the effect of each one.

The three overlap. When they do, the one you notice is the first. The sense of too much is experienced clearly, whereas the rearrangement of the order is not experienced at all, because you view the world through the order that came after. The sequence in which you notice them and the sequence in which they should be handled therefore do not match.

One clue separates them. List the information you encountered last week that is affecting your current judgements. In the first form, you cannot recall what you read. In the second, you remember the content and it sits unresolved as to whether to adopt it. In the third, items come up clearly, and you notice that they conflict with the direction you held the week before.

Calling it being easily swayed by information takes all three on together and puts the cause on the information environment. Treated as a matter of environment, the remedy becomes reducing exposure. And when reduction does not work, what comes next is reducing further. With remedies running only in the direction of less, the third form goes unhandled to the end. Unhandled, the failure of reduction gets explained by the environment as well. Too much information, bad platforms, the times we live in. Each is accurate as an observation, and none of them touches the order.

The Mainstream Responses Split Between Holding Criteria and Checking Truth

Count the responses on offer and there are five.

Hold a criterion for judgement. Decide your own axis and let in nothing that does not fit it. Check the source and cross-check against other information. Set the good and bad sides side by side and judge for yourself. Once you have the information, act and find out.

All five work. The first and second in particular run in the same direction as this article’s conclusion. The observation that necessary information settles once a criterion exists is accurate.

Take the five in turn and see what each touches. Checking sources and cross-checking, comparing both sides. These land on the second form, where truth cannot be assessed. Acting to find out is an operation for producing material yourself.

Holding a criterion and an axis reaches the third form. The remedy, though, has taken the form of a demand: hold one.

That failing to hold one when told to is the person’s problem does not follow. An axis is the order of what matters. An order does not stand up by resolving to have one in your head; it takes shape afterwards as a record of judgements you placed. That process is handled squarely in Your own axis is not found but placed again.

So “hold an axis and you will not be swayed” is correct, with the sequence placed backwards. You do not hold an axis in order to produce a state of not being swayed; being harder to sway is what results once the record of placed judgements accumulates.

None of this says the five are void. The verification procedure works on assessing truth. What is in question is not effectiveness but which form each one reaches.

Setting “let in nothing that does not fit the axis” as a rule, meanwhile, has a side effect. The axis itself stops being updated. What does not come in cannot be material for rewriting the axis. You stop being swayed at the price of closing the route by which change gets detected.

And “act and find out” is a strong remedy among the five, and an expensive one. You can find out one thing at a time, and results take time to return. This remedy therefore cannot be applied to all information. An operation for selecting what to apply it to has to come first, and that selection requires, in the end, an order. The remedy cannot produce what it presupposes.

Beyond this, the set of five produces wariness about contact with information as such. Wariness reduces contact and reduces material for judgement along with it. A judgement made with little material comes with more conviction and no more accuracy. And the shortfall in accuracy is not observed until results return.

More Draws People In; Fewer Produces Decisions

There is a study that measured the effect of the number of options on decisions: the experiment published in 2000 by the psychologists Sheena Iyengar and Mark Lepper (Iyengar & Lepper, 2000, Journal of Personality and Social Psychology, 79(6), 995–1006). A tasting table was set up in a grocery store, displaying six varieties of jam on some days and twenty-four on others.

First, stopping. Of the 242 people who passed the twenty-four-variety table, 60% (145) stopped. At the six-variety table, of the 260 who passed, 40% (104) did. More drew people in.

Next, purchase. Under the six-variety condition, about 30% (31 people) went on to buy jam. Under the twenty-four-variety condition, 3% did.

More brought people to a halt; fewer produced decisions.

This asymmetry bears directly on how a business is designed. The measure of attention gathered and the measure of decisions completed are measuring different things. And in most settings, what gets observed is the former. Impressions and volume of response both sit on the stopping side.

Buying jam and making a business judgement are different. What this experiment covers is a purchase made on the spot. What can be taken out as structure is only that increasing the number of options can act in opposite directions on the measure of interest and the measure of decision. That applies when you are on the receiving end of information and when you are the one putting it out.

That this effect depends heavily on conditions is something later verification established. It cannot be read as a general rule that fewer options is better.

The “decision” measured in this experiment is also one that completes the same day. A business judgement, where months pass between deciding and seeing a result, cannot be said to be the same kind of thing. Over long horizons, doubt has room to return after the decision, so a separate problem arises about where the completion of a decision gets counted. That problem is not in this experiment’s design.

And this experiment circulated in a simplified form. “Too many options and people cannot choose” is a memorable, applicable sentence. A memorable form travels with its conditions dropped. What dropped away were the conditions under which the effect holds.

However, “Too Many Options and People Cannot Decide” Is Not a General Law

The finding that fewer options produced more decisions carries an important limitation. Later verification showed that this effect does not hold in general.

The psychologists Benjamin Scheibehenne, Rainer Greifeneder and Peter Todd published a meta-analysis in 2010 (Scheibehenne, Greifeneder & Todd, 2010, Journal of Consumer Research, 37(3), 409–425). It pooled 50 experiments, 63 conditions and 5,036 participants, published and unpublished.

The result was a mean effect size of essentially zero.

Substantial variability between studies was reported at the same time, however. Effects appear in some settings and not in others, and the result reverses with conditions. The paper lists several conditions that could underlie the effect, and states that no condition guaranteeing it could be identified.

Two readings come out of this.

First, “too many options and people cannot decide” cannot be built into a design as a premise. Whether your own situation sits on the side where the effect appears cannot be known in advance.

Second, the jam experiment’s result has not been erased. It happened under specific conditions, and a mean of zero does not mean it fails to happen in individual settings.

So the finding is not unusable; its use changes. It cannot serve as a general rule, and it can serve as a hypothesis to check in your own setting. Whether reducing options increases decisions is measurable in your own business. The point is not to adopt as a general rule what you could measure.

Nor can the meta-analysis be read as proof that no effect exists. Large variability includes effects being real under some conditions. A mean of zero is not proof of nullity but a description of conditions not yet identified.

This case also serves as a model for handling other findings. An effect shown by one study often narrows into something conditional under later verification. Deciding a business design on the strength of a single experimental result is therefore itself a way of handing the order to outside information. The measurement of how much the entrance is delegated is read the same way — as a figure with conditions attached.

And limitations of this kind are hard to remember. “Fewer options is better” fits in one sentence; “it reverses with conditions, and the conditions have not been identified” does not. What circulates is the shorter one. Limitations are lost faster than the findings they belong to.

What Being at the Mercy Contains Is the Order Being Rearranged Continually

When you are at the mercy of information, what is actually moving is the order of what matters.

Judgement requires an order. Whether to prioritise speed or accuracy, whether to place income above freedom of time, whether to take the work available now or the work that leads somewhere. No amount of material settles these. The order settles them.

When the order moves each time information arrives, the conclusion changes every time. With the same material, a change in what sits on top changes the conclusion instead.

And that the order moved is not noticed. You view the world through the order that came after, so it feels as though that order had been yours all along. What registers is a change of direction; a change of order does not.

Suppose you read an article saying you should publish more often and raise your frequency the following week. A month later you read one saying quality is what matters, and you lower it. Both judgements feel reasonable at the time. What is moving is the order of frequency against quality.

Updating a judgement with new information is correct. What deserves separating is whether the material was updated or the order was. An update of material — new data on how well some method works — sharpens the conclusion as long as the order is fixed. An update of order moves the conclusion somewhere else entirely.

That the order should be fixed does not follow. An order ought to be updated by the experience of placing judgements. The problem is when the input for the update is other people’s claims rather than your own experience. Updates from experience are slow and updates from claims are fast, so left alone the latter dominates. The difference in speed becomes the difference in composition.

The experience side is not always the slow one, though. A judgement whose result returns within days — whether the quote you sent was accepted — updates quickly from experience too. Domination happens only for judgements whose returns are slow. And judgements that set the direction of a business tend to sit on the slow side.

Whether the order moved cannot be assessed in the moment. It can be assessed only when you produce a different answer to the same material. Detecting a change of order therefore requires a record of what you judged, and how, before. Without a record, the movement cannot be observed at all.

And where the order moves frequently, nothing accumulates. Every direction is abandoned before its outcome is known. Nothing being known, the experience needed to update the order does not build either. Being at the mercy of information stops the production of the very material that would end it. And with nothing built, the next move of the order comes, again, from an outside claim.

The Entrance Does Get Delegated. It Is Smaller Than Imagined

A selection always occurs at the entrance to information. Where you do not make it, it is not that no selection happens but that something else makes it.

One thing that performs it is a recommendation system. What decides what gets displayed is the measure that platform optimises. That measure is often time spent or volume of response, and whether something helps your judgement is not in it. The information coming in is skewed as a result. Strong claims, extreme cases, things shaped to move feeling. These gather response, so they get selected.

How large a share the delegation actually accounts for has been measured.

In 2015 Bakshy, Messing and Adamic decomposed, for users of one large platform, where the skew in what reaches the eye comes from (Bakshy, Messing & Adamic, 2015, Science, 348(6239), 1130–1132). The three parts were who you have chosen to connect to, the ranking system that orders what is shown, and which items you actually open.

The result was that most of the skew came from the first and the third — the user’s own choices — with the ranking system contributing less than those.

This result does not generalise as it stands. It is a count over one platform, one period, and only those users who declare their own position. The information involved is current affairs rather than material for business judgement. The handling does not change here: not a basis for design, but a hypothesis to check in your own setting.

On that basis, what the decomposition shows matters. “The system is doing this to me” places responsibility outside, and it sends what can be moved outside along with it. Within what could be measured, much of the narrowing sat in who had been chosen as a connection — an entrance selected, in the past, by the user.

Suppose you spend time watching what people in your field publish. What appears is reports of results and strong claims. Cases that did not work seldom appear, because they gather little response. What you see is a distribution shaped differently from the real one. Part of that distortion, though, was produced by the past choice of who to follow.

That recommendation matches your interests, and so is useful, is half right. What remains is that it matches interest rather than judgement. What draws interest and what serves as material for judgement overlap without being the same. And what is optimised is the former.

The skew is not produced by the system alone either. People publishing have their own reasons to write up what worked. Remove the system and the observed distribution stays skewed towards success. Ending the delegation fixes part of the skew, not all of it.

And a delegated state comes with the feeling of choosing. Whether to read what was displayed is your own choice. The sense of selection remains while the set of options was assembled by something else. Because the sense remains, the delegation stays invisible.

Not Cutting Off, but Designing the Entrance

From the three forms and the measured delegation, the form of the remedy is settled. Not reducing the volume, not cutting off, but selecting the entrance.

Cutting off has a problem. With nothing coming in, the order does not move. Nor, at the same time, does material for updating it. The order you currently hold gets fixed as a result. Fixed is stable, and it is also a state in which change cannot be detected.

Selecting the entrance means deciding for yourself which routes you receive information through. Decide what you read at the level of people and publications, and close the inflow routes for everything else. Rather than assessing items one at a time, settle the set.

Assessing one at a time does not work because by the time you assess, you have read it. You cannot assess without seeing the content, so by the time you do, the effect on the order is already complete.

Instead of watching what your field publishes, decide on three sources you want as material for judgement and check those regularly. The volume you read need not fall. What changes is who assembles the set.

Narrow the entrance and the area outside it becomes invisible. The area you cannot see as a result of choosing, though, differs in character from the area you cannot see for want of choosing. With the first, you know what you left out. Knowing, you can open it when you need to. With the second, you do not know you left anything out.

Selecting the entrance is not done once. A source you chose can quietly stop being useful. Periodic review is needed. The frequency of the review, though, can be far lower than the frequency at which information arrives.

Selecting the entrance is not aimed at comfort either. An entrance assembled only from claims you find agreeable is a result of choosing and carries no material for updating the order. The criterion is not “do I agree” but “does this source give its reasons when it reaches a conclusion different from mine”.

Only the set side is treated here — what gets let in. How the material then gets used — the line between material that may be borrowed and judgement that may not — is a separate problem, carried by What thinking for yourself actually contains. Tidy the entrance and the confusion of material with judgement still remains.

And narrowing the entrance produces, in the short run, an unease that there is now too little information. The sense that something may have changed while you were not watching. That sense is real and cannot be disproved. For what you are not watching, you cannot confirm that nothing changed either.

In Practice, the Wobble Shows in Method, Positioning and Pricing

The wobble in order shows up, in practice, in three situations.

First, switching methods. Each time you learn a new way of finding clients, you break off what you are doing and move. Every one of them switches before its outcome is known, so not a single piece of assessable material accumulates. And with nothing accumulated, the next judgement leans on outside information too.

Second, changing your positioning. You see someone’s success and move your positioning towards it. Where you moved to was built on that person’s order. It was not built on yours, so the reason to keep going is not inside. Where the reason to keep going gets placed is handled in How to build a concept.

Third, wavering on price. Seeing a high figure makes you want to raise yours; seeing something offered cheaply makes you want to lower it. What is setting the figure is whatever you saw at that moment.

The problem is not the changing but the input to the change sitting outside. That single point holds for all three. A change based on your own measurement is an adjustment, however many times it happens.

Increasing the number of attempts is correct. What deserves separating is an attempt from a switch. An attempt fixes the criterion for assessment first and runs for a bounded period. A switch moves on before the assessment. From outside both look like trying various things, and one leaves a result behind while the other does not.

Of the three, the first is easiest to handle. Writing down the period and the criterion in advance is enough to distinguish a switch from an attempt. The third moves most readily, because a figure is tied directly to psychological load.

For all three, a state where outside information has no effect at all is not the goal. Market conditions genuinely change, and stopping outside input leaves you unable to respond. What is wanted is not a cut-off but a state in which you can name, every time, what the input to a change was.

And switching the input to your own measurement slows judgement down. Measurement takes time. During it, people moving quickly on outside information look further ahead. Whether you can bear that appearance is the practical constraint. And during the period of bearing it, nothing appears as evidence that you are bearing it. Until a measurement returns, you are indistinguishable — from outside and from within — from someone simply moving slowly.

Distinguishing Whether You Are the One Moving the Order

There are three assessments.

First, look at the source of the reason for a change of direction, rather than the number of changes.

Take several recent changes of direction and write, for each, what you saw that made you change. Where your own measurement is the source, it was an adjustment. Where someone’s claim or example is the source, the order is being moved from outside. The count itself is not used in the assessment.

Second, look at your reaction to the same information three months later.

When a strong claim makes you want to change direction, hold that judgement for three months. If the conclusion is the same after three months, it followed your order. If interest has gone by then, what was moving was attention, not order.

Third, look at whether you can list your entrances.

Check whether you can write out the routes through which you currently receive information. If you can, the selection sits on your side. If “it just turns up in front of me” covers most of it, the set is being assembled by a recommendation system.

There are three markers: the source of the input to a change, reproducibility after an interval, and the visibility of the routes. The volume of contact itself is not used in this verdict. Contact is not the problem, so whether it fell is not something to assess.

The more material piles up, the less settled the choice becomes. That experience does not indicate that you were looking badly. Looking increased the material, while the order in which material gets arranged was moving on the same input. More material with a changing arrangement does not produce more answers. The sense that fewer things get settled captures the state precisely.

The advice to hold a criterion comes with a premise: a state in which the criterion already exists. What this article looked at sits one step before that. Until the route by which the criterion gets moved from outside is closed, a criterion gets rewritten as fast as it is placed. In sequence, the route comes first.

Reselecting the entrance does not reduce the information. What moves is not the volume but who assembled the set. And within what has been measured, most of that set came from connections a past version of you selected and from what you actually opened. What can be moved sits closer to your own side than it appears.

How the order of what matters takes shape is handled in Your own axis is not found but placed again, and the work of putting it into words and into the shape of a business squarely in Building a view of the world. The design of placing the order first and looking outward afterwards is handled in How to build a concept.

Adjacent states are treated elsewhere. The line between material that may be borrowed and judgement that may not is handled in What thinking for yourself actually contains, the mechanism by which the entrance to judgement disappears in Having stopped questioning things is not idleness, and the machinery by which choices align with those around you in What is happening inside someone who is easily swayed.