Why a big push breaks the system, and how to dose change instead

A new director arrives and turns everything over in a month: a new structure, new rules, a new tone of voice. A year later the company is standing still, and the best people have already left. What this director lacked was dosage. And that is why a strong blow almost always breaks the very thing it was meant to fix.

A living system does not rebuild itself from an impact. That holds for a person, a team, a company, and for the market you walk into with your advertising. A strong push produces shock rather than restructuring. In shock, the system reacts all at once, as one piece, crudely. It does not work out what exactly needs to change. It defends itself.

It is like repairing a watch with a sledgehammer: plenty of force, no precision, and fragments instead of fine tuning. For anyone selling into Russia from abroad the temptation to arrive with a big push is especially strong.

Why shock breaks instead of repairing

Consider what physically happens inside a system when it is hit hard.

When the influence is weak and targeted, only the area you aimed at responds. One department changes one process. One audience segment hears about you for the first time and calmly remembers. The change is local, easy to notice and easy to roll back if it goes the wrong way.

When the influence is strong, everything responds at once. The system cannot tell where the problem is, so it switches on a general alarm. People stop thinking about the work and start thinking about survival. Who stays here. Who gets cut. Is it worth investing effort, or time to look for another job. In that state no fine adjustment is possible, because all the resources go into defence.

The point is simple. The harder you press, the cruder the answer. Force and precision pull in opposite directions.

Why revolutions inside companies so often fail

Take the most common scenario. A new manager comes in with a brief: shake things up, speed things up, take the business to the next level. They try honestly: a new org chart, rewritten procedures, a fiery speech at the all-hands meeting.

Then the system goes into shock. The usual links between people break: who agreed what with whom over years no longer works. And the first to leave are the best, while the weakest stay.

This is the point almost everyone misses. A strong employee is sensitive. They read quickly that things have got worse, that their work has less meaning, that decisions are now made past them. Their threshold is low: a little is enough for them to decide and go. And the job market is waiting for them, the offers exist.

Those who stay often stay without believing in the change. They stay because they have nowhere to go, or because they simply endure. A high threshold and low sensitivity: they are harder to upset, and they also bring less.

A year later the reform is done, the slides are beautiful, and the team is weaker. Formally the change happened. In substance the system degraded.

I have seen this in projects more than once. An owner in despair calls in a “strong turnaround manager”, the manager cuts from the shoulder, and six months later the team has to be assembled again. The manager is rarely stupid. They were hitting a watch with a sledgehammer.

The same mechanism works on a customer base. A company decides to raise prices by a third, change the packaging completely and rewrite all its sales scripts in one week. The logic sounds sound: if we are changing, let us change everything at once. The result is that regular customers who bought out of habit for years suddenly meet an unfamiliar product at an unfamiliar price from unfamiliar people. They go off to compare you with competitors, something they never did before. The most loyal, the ones with the lowest threshold of trust, break away first, exactly like the sensitive employees.

How to change when sharp change is ruled out

A hard task is solved as a staircase rather than a single jump, and the pattern is always the same. A small step. A check of the result. A correction. The next small step.

Each step is chosen against two conditions at once. It has to be small enough that the system does not go into general alarm and start defending itself with its whole body. And it has to be noticeable enough that something actually moves in the area you are aiming at. Too small and nothing changes, you are marking time. Too large and you are back in shock. Real change management lives between those two limits.

The check after every step is there for a reason that has nothing to do with bureaucracy or reporting for its own sake. It catches the signal that you are moving in the right direction. The system needs confirmation: things got better, the direction is right. Without it, even correct steps feel like chaos, and people slide back into alarm.

Think of a hot water tap. Open it fully in one movement and you get scalded, so you turn it a little, touch the water, turn a bit more. That feel for the water is the feedback. Running a company or an advertising campaign works the same way: you turn a little at a time and keep feeling the result.

People also confuse the size of the step with the size of the ambition. Small steps do not mean small goals. The goal can be huge. What has to be small is each separate move toward it. A climber sets a big summit and still climbs it in short, precise holds.

That confirming mark is sometimes called a reward, and it is easy to misread. It means feedback that fixes the shift in place, which is different from a bonus or praise: a number went up, a customer came back, inquiries increased, a conflict in the department calmed down. The system sees the result and allows itself to keep changing. That is the fuel for the next step.

What this means for a company entering Russia

For a foreign business the big push usually looks like a launch. The Russian site goes live, the whole quarterly budget goes into Yandex Direct in the first weeks, a new local positioning is announced, and a sales team starts working leads in a language and a set of channels the headquarters has never tested. Everything is new on the same day, so nothing can be judged on its own.

Russia raises the cost of this approach in a few specific ways. Yandex Direct is a different auction from Google Ads, with its own automated strategies that need accumulated conversion data before they behave predictably. The buying habits differ: a lot of qualified contact moves into messengers, and a company that only counts forms reads its own funnel wrong. And Russian buyers who have never heard of you have no prior trust to absorb an unfamiliar offer, so the first impression carries more weight. If all these variables change at once, a bad month tells you nothing about which one failed.

The dosed version is less dramatic. Start with a narrow test in Yandex Ads on one segment and one offer. Look at the response. Correct the offer and the landing page. Widen the geography or the semantics. Each round gives data the next one stands on. The work in Direct is built this way anyway: a steady stream of small checks, where every week you see what worked and adjust, instead of one bet on everything.

Familiarity with a brand builds the same way. Nobody can be made to love a company with one video, even a brilliant one. Trust accumulates in touches: saw it, remembered it, heard it again, came across a review, visited the site, came back. Ten calm touches work better than one bomb after which a person burns out and forgets. When we build positioning and brand identity, we are designing that staircase of recognition rather than a single loud shot.

Sales follow the rule too. Trying to close a large deal with one heavy push frightens the client; they feel pressure and defend themselves. Small agreed steps, each a little larger than the last, lead to a yes far more reliably: first a short call, then a trial task, then a small contract, and only then a large one.

And a website. Tearing down the old Russian site and rolling out a new one in a day almost always ends with a dip: search positions fall, regular customers get lost in the new navigation, inquiries drop at exactly the moment you expected growth. Change it in parts. One section first, then watch behaviour and numbers in Metrica, then the next. That way you always know which edit helped and which hurt, and you can roll back one page instead of the whole project.

How to tell that you are pressing too hard yourself

Recall the last big change you started in your business, a repositioning or a new site overnight, and ask three questions.

Was it one strong blow or a series of checked steps? If it was a blow, you are already in the risk zone.

What happened to your best people and best customers afterwards? If the most sensitive and valuable began to drop away first, that is the effect described above: the ones with a low threshold and a choice leave.

And the main one: what would the same task look like broken into ten small steps with a check after each? Concretely: what is the first step, how will you know it worked, what is the second.

Often this exercise alone shows that the “bold decision” was impatience, and the system answered the only way it could: it contracted and started losing.

There is a mirror error too. Dosing does not mean endless shuffling and fear of doing anything. Some owners slice change so thin that a year passes without movement. A step must be small and real: after it something visibly changes. If the check shows nothing, the step was too timid and should be enlarged. The craft is in that sense of measure between too strong and too weak.

Force does not decide everything. A watchmaker repairs a watch with the most precise tool in the box, and the heaviest one stays on the shelf. Start with one small step. Check it. Take the next one.

Frequently asked questions

Sources

• Russian version of this article on belokrylovo.ru: Почему большой импульс разрушает систему