From activity to profit

A marketing report can be busy and still say nothing about money. Sort every number into three kinds:

• Activity: what the team did. Posts, ad views, spend, followers.
• Leading metrics: what customers do early that points at the result. Enquiries, trial lessons.
• Lagging metrics: the result itself, which arrives later. Paying families, second packages, contribution.

Uliana's September report: 24 posts, 1.2 million ad views, 900 new followers, 600 enquiries. Only the enquiries are customers doing something; the rest is activity. Hanna's question, "what did it return?", can only be answered further down the chain.

So put the numbers in one chain, from what you spend to what Klasna keeps: ad spend and posts → enquiries → trial lessons attended → families who pay for a first package → families who buy a second → contribution. Every link should be something you can count.

Each step is a filter. You can be very busy at the top and still earn little at the bottom, and the chain shows exactly where people drop out.

Start from the result you need and work backwards through the chain to the leading targets. Then the early numbers tell you within weeks whether the goal is in reach.

For February 2027 the plan spends 300,000 UAH (150,000 paid social, 90,000 search, 60,000 school partnerships); last spring's tests suggest about 110 new paying families at those levels. Assume that 55% of trial lessons become paying families and 50% of enquiries book a trial:

• trials needed: 110 ÷ 0.55 = 200;
• enquiries needed: 200 ÷ 0.5 = 400.

Now "400 enquiries and 200 trials in February" are targets you can check by mid-month, long before the money shows up.

The rates are assumptions, and you write them down as such. Take them from last spring's records if you have them; if not, they're a guess, and you replace them with real figures in the first two weeks of February.

The common slip is multiplying instead of dividing: 110 × 0.55 × 0.5 gives about 30, which is the families 110 enquiries would bring, not the enquiries 110 families need.

The chain ends in contribution, so the lagging goal should be in money too, not only in families. That's the number Hanna can hold the plan to.

A family that starts in February is worth about 6,200 UAH of contribution: about 5,600 by the end of May, plus its chance of coming back in September. With the February goal of 110 new families:

110 × 6,200 = 682,000 UAH.

That makes the plan's promise concrete: 300,000 of February spend aimed at about 682,000 of contribution from the families it starts.

Two cautions. First, use the value of a February starter, not a family's full first-year value: a family that joins in February has only a few months before summer. Second, don't call all 682,000 marketing's return yet. Some of those families would have come anyway, through word of mouth or returning from last year, and only the families the campaign added are its result. Finding that number is the next lesson's job.

Writing the goal in money also lets you compare it with the spend later, honestly, once you know how much of it the marketing really caused.

The three kinds of metric from the chain (activity, leading, lagging) are easy to name and easy to mix up. One question sorts each number: who did it, and when does it arrive?

• If the team did it, it's activity: posts published, ads shown, money spent. It measures effort, not customers.
• If customers did it early, and it predicts the result, it's leading: trial lessons booked by parents of exam-age teens. It moves within days and gives you time to act.
• If it's the result itself, arriving weeks or months later, it's lagging: contribution from February's new families by the end of May.

A leading metric earns its place only if it really predicts the lagging one. Trials booked by parents in the target segment do; likes on a post usually don't. That's why "enquiries" is better as "enquiries from the target segments": an enquiry from someone outside the segments rarely turns into a paying family.

The slip: putting only lagging metrics on the plan. By the time contribution for spring arrives, the spring is over and there's nothing left to fix.

A metric nobody owns slips quietly. Give every metric in the plan exactly one owner: the person who acts when it slips.

Klasna's spring chain has three owners:

• Uliana owns enquiries from the target segments;
• the academic lead owns trial-to-paid, how many trial lessons turn into paying families;
• Hanna owns contribution.

Uliana also owns the plan as a whole. But each link still has one name next to it, so when trial-to-paid drops from 55% to 40% in the second week of February, nobody has to ask whose job it is: the academic lead looks at what's happening on the trials and acts, that week.

Why one name and not a group: "the team owns it" means each person can assume someone else is on it. And why the person closest to the link: Uliana can bring more enquiries, but she can't fix what happens in a trial lesson; Hanna sees the money only months later.

The slip: an owner without the power to act. Whoever owns a metric needs to be able to change what drives it.

When a number is rewarded, people find the cheapest way to move it, whether or not the result moves too. Any single number can be gamed.

Ostap, Klasna's ads freelancer, wants a bonus per enquiry. Pay per enquiry and you'll get enquiries: cheap, plentiful and from people who never pay. The ads drift to wherever enquiries are easiest, not to the parents who buy.

The fix is to pair a number with one it can't fake and to pay for results rather than activity:

• base the bonus on paying families from his campaigns;
• report enquiries from the target segments next to it, so the leading metric stays honest too.

Two fixes that don't work. Doubling the bonus per enquiry rewards the same gaming more. Paying per ad view rewards pure activity, the easiest number of all to inflate. And removing every metric so nothing can be gamed leaves you with no control at all.

The habit: before you attach money or praise to a metric, ask "what's the cheapest way to move this without helping Klasna?" If there is one, add the pair.

For a spring plan, a leading metric has two jobs: it must move early, weeks before payments, and it must predict the lagging result, so that a drop in it is worth acting on.

For Klasna's spring plan, two pass both tests:

• enquiries from parents of teens taking the exam in 2027: the target segment, the first step a buyer takes;
• trial lessons attended: the step closest to paying, and the one the academic lead can act on.

Three numbers fail. Likes on posts and followers gained move early but predict little: they are activity, and many of the people behind them will never be payers. Contribution for the school year predicts nothing, because it is the result, and it arrives too late to steer the spring.

A good test of a candidate: if this number fell by a third next week, would you change something, and would you expect fewer paying families a month later? If the answer is "no" or "not sure", it isn't a leading metric for this plan.

Gaming isn't only an ads problem. Every link in the chain can be pushed in a way that helps the number and hurts the result.

Take trial-to-paid, the share of trial lessons that become paying families. It's close to money, so it looks safe to reward. But it can be gamed by pressuring families on the trial: a hard sell, a "this price is for today only". The rate goes up, and so do refunds and families who never buy a second package. A family pushed into buying rarely buys again.

So pair it with numbers that catch the pressure: second packages and refunds from those families. If trial-to-paid rises while second packages fall, the rise isn't real.

The same pattern works up and down the chain: enquiries paired with paying families, trials paired with attendance, new families paired with how many stay.

How it fits together: the chain links spend to contribution, targets run backwards from the goal, each link has one owner, and every rewarded number has a guard. The next lesson asks how many of the families after a campaign it really added.