Changes without a committee, and a decision log

A change request is any request to alter what the project agreed: its scope, date, budget or success criteria. One door means every such request reaches the project lead as a short written note, whoever it comes from: a developer, the marketing lead or the sponsor.

The door exists because changes rarely arrive as changes. They come as a remark in the kitchen or a quick message, and each one sounds small. Taken one by one into the work, they eat the buffer before anyone has seen their total cost. Writing the request down is the first filter: it makes the person say exactly what they want, and it gives the lead one place to see everything that is pending.

One door is not a freeze and not a board. Anyone may ask, and changes still happen; they just enter the work after someone has sized them and the right person has decided. At Klasna in January, Leo's idea of a parents' progress page and the marketing lead's free trial lesson both go through the same door.

Once a change request comes through the one door, it follows the same short path every time, so nobody has to invent a process under pressure:

1. The request in writing: what exactly should change and why.
2. The cost from the doers: the lead asks the people who would do the work how many days and how much money it takes, and what else it would push back.
3. An impact note: five lines at most, with the lead's recommendation.
4. A decision by the person allowed to make it: the lead below an agreed threshold, the sponsor above it.
5. A record: the decision goes into the decision log and the team hears about it.

The order matters. Deciding before the cost is known turns the decision into a guess; costing without a written request means sizing something nobody has defined; skipping the record means the same request returns next month as if it were new. For the free trial lesson at Klasna, the path takes two or three days, not a meeting series.

An impact note is the lead's short answer to "what would this change cost?" Five lines are enough: what changes, how many days, how much money, what else moves, and a recommendation.

The days and money come from the people who would do the work, not from the lead's memory or the requester's hopes. The requester wants the change and naturally sees it as cheap; the lead knows the plan but not every task; the sponsor decides, so if he also set the cost the note would check nothing. The doer knows the code, the calendar and what else is already on the desk.

The lead's job is to ask, question the reasons behind each number and translate them into the project's terms: days of buffer and money from the reserve. At Klasna, Danylo says a "trial" booking type needs four of his days, and in January those days don't exist; Hanna gives the cost per lesson. Neither number is the lead's guess, and both go into the note.

A change has two prices, and the impact note shows both: days (whose time, and what it pushes back) and money (what it takes from the remaining budget or reserve). Each is worked out per unit and then multiplied, so the reader can check the arithmetic.

A change that earns nothing is pure cost. At Klasna a trial physics lesson still pays the tutor 430 UAH and uses 31 UAH of video and AI, so each one costs 461 UAH. A hundred families on the waiting list would mean 100 × 461 = 46,100 UAH, nearly all of the 52,000 UAH left under the ceiling in January, plus four of Danylo's days for a trial booking type.

Put the total next to what is left, not next to the whole budget: 46,100 UAH looks modest against 600,000 and alarming against 52,000. The second comparison is the one that matters, because only that money can still be spent before opening. The note then ends with a recommendation, for example: not before opening, look again in March with real booking data.

Change control needs a named decider, not a committee. The brief already says who decides what: at Klasna, Leo decides changes to the date, the budget ceiling, the out list and the success criteria, in writing, within two working days.

Everything inside those limits belongs to the project lead. To stop small things from queuing at the sponsor's desk, agree a threshold once, in advance: for example, the lead may spend up to two days of buffer and up to 10,000 UAH from the reserve, decides alone and reports it in the Wednesday status. Above the threshold, or when something touches what only the sponsor owns, the change goes to Leo with its impact note.

So swapping two practice topics with no extra days, or giving Danylo one more day to fix the test's mobile layout, is the lead's call. A free trial lesson at 46,100 UAH breaks the money threshold, and a parents' progress page adds scope the brief never contained: both go to Leo. The sponsor gets fewer, bigger questions; the team gets fast answers on the rest.

A decision log is a running list of the project's decisions, a few lines each, kept where everyone can find it. Each entry answers six questions: when, what was decided, who decided, which options were dropped, why, and what it changes in the plan, budget or scope.

The log is for the people who weren't there, or who will have forgotten: the marketing lead who asks again in February, a new colleague in March, Hanna checking what was not spent. Without it the answer to "why did we launch without a trial lesson?" is another meeting; with it, it's a link.

Two things stay out. The chat thread that led to the decision is too long to reread and buries the reason. And who argued for which option turns a record into a scorecard of people, which makes the next discussion more cautious. At Klasna, Leo's no to a free trial before opening becomes one entry: the date, his decision, the dropped option, the 46,100 UAH reason and the plan to look again in March.

The one-door rule has no exception for the most senior person, and that is the point. The sponsor decides changes, but he should decide them seeing their cost. A request he makes in passing, if it goes straight into the work, skips the impact note: nobody sizes it, nobody sees what it pushes back, and the team quietly spends days the plan doesn't have.

The polite way through is simple: thank the sponsor, write the idea up as a request, and send him the impact note like anyone else's. The rule protects him too. At Klasna, Leo's parents' progress page would have cost Danylo a large part of January; seeing the note first, Leo can decide whether it is worth that.

A side door from the top also teaches the team that the door is optional. If the sponsor's remarks go straight into the work, everyone else learns to ask the sponsor instead of the lead, and change control is gone without anyone deciding to drop it.

A change rule needs a clock. At Klasna the sponsor answers change notes in writing within two working days. The clock matters as much as the decider, because the most common way a change gets decided is by nobody: the note sits unanswered, the work drifts, and by the time anyone looks, the calendar has made the choice.

So silence is neither a yes nor a no. When the time passes, the lead chases the decider, with a call if a message didn't work, and says what waiting costs: "each day without an answer moves the release a day". Treating silence as approval lets a change in that nobody owns; treating it as a refusal quietly drops something that may matter. Either way the decision log would have an entry nobody actually decided.

How it fits together: one door collects the requests, the doers' numbers feed a five-line note, a named decider answers on a clock, and the log remembers. The next lesson asks how to plan the final weeks when the date is fixed but the details keep changing.