Reducing the buyer's risk

People who buy for a company are mostly avoiding a loss, not chasing a gain: a rollout nobody uses, a wasted budget, looking foolish in front of the CEO. Their signature is on the decision, and a failure is remembered longer than a success. That is the buyer's risk, and it stalls more deals than price does.

Veloria's finance director read Klasna's proposal and asked one question: what if the dispatchers stop attending after two weeks? She didn't ask what Veloria would get. At 800 UAH a lesson, a lesson nobody shows up for is money paid for nothing.

A longer list of benefits leaves her question unanswered. What lowers risk is proof she can use: something she can check against her own records, someone she can call, something she can test on a small scale, or a promise she can hold Klasna to. Louder claims add nothing; specific, checkable evidence and a way out if it doesn't work are what let a cautious buyer say yes.

Each person on the buying side guards against a different loss, so each needs a different kind of proof. One deck for everyone answers nobody.

At Veloria:

• Pavlo (HR and learning) proposed the program; if it flops, it's on him. He needs a reference he can call: an HR manager at a current Teams client who'll say honestly how it went. A phone number, not a quote on a slide.
• The finance director fears paying for lessons nobody attends. She needs numbers she can check against her own records (lessons scheduled, attended, billed) and an exit if it doesn't work.
• IT security and legal fear employees' data leaking. They need a written data answer: what's collected, where it's stored, who sees it, when it's deleted, ready before they ask.
• Shift leads need lessons scheduled around the rota; dispatchers need to know a poor tutor fit gets fixed fast.

Honesty is part of the proof. Teams has a few small clients and none in logistics; saying so plainly costs less than being caught pretending.

Some "proof" raises the buyer's risk instead of lowering it, because a careful buyer checks.

• Vague claims. "Trusted by leading companies" names nobody. A finance director reads it as "they couldn't name one". It adds a question rather than answering one.
• Invented testimonials. Someone at Klasna drafted a quote "from the HR director of a large logistics company". Klasna has no such client. One phone call exposes it, and then every other claim in the deck is in doubt too.
• Guarantees you can't keep. Promise only what you control. Klasna has many English tutors, so it can promise to replace a tutor within a week if the fit is wrong. It can't promise that every dispatcher will be fluent in three months: that depends on the learners, their time and their starting level.

The test for any item in a proof pack: can the buyer check it, call it, test it or hold you to it, and will it survive that check? If not, it doesn't belong there.

A trial lowers risk only if it decides something. "Let's try it for a month and see how people feel" sounds reasonable, but it has no measure and no end: a month later, everyone has an opinion and nobody has a decision.

A pilot (a proof of concept) has four parts fixed before it starts:

• a group: who takes part;
• a length: how long it runs;
• a measure agreed in advance: what result counts as success;
• a decision date: when someone says yes or no, and who that is.

For Veloria, Klasna's working assumption is one dispatch team of twelve, two lessons a week, for four weeks: 96 lessons. The measure and the decision come from the finance director's own question about attendance, so the pilot answers exactly what she is worried about.

A pilot with all four parts turns "we'll see" into a test with a verdict, which is what lets a careful buyer commit to the small step now and the big one later.

A free trial looks like the easiest way to lower the buyer's risk. It moves the cost to the seller and, worse, takes the stakes away from the buyer.

The cost is easy to count. Veloria's assumed pilot is 96 lessons (12 dispatchers × 2 lessons a week × 4 weeks). Each lesson costs Klasna 461 UAH in tutor pay and video. A free pilot means 96 × 461 = 44,256 UAH out of Klasna's pocket, with no guarantee of a contract.

The stakes matter even more. With nothing paid, nobody at Veloria has a reason to watch attendance or push people to show up, so the pilot drifts and proves little. A paid pilot at the usual 800 UAH a lesson (96 × 800 = 76,800 UAH) makes the result someone's business. To keep the risk low for Veloria, the payment is credited to the contract if it signs: the buyer risks only the pilot, and if it works, that money was never extra.

A pilot should predict the rollout, so the group has to look like the people the full program would cover. For Veloria, that means one real dispatch team on its normal shifts, keen or not.

The temptation is to pick the people who will make the numbers look good: volunteers who already speak some English, or the most enthusiastic employees. Hand-picked enthusiasts attend, so attendance looks great, and it tells the finance director nothing about whether ordinary dispatchers on night shifts will show up. When the full rollout starts, the real numbers arrive, and with them the doubt about everything the pilot "proved".

Two other choices miss the point too. A mixed group of anyone who signs up, from any department, doesn't match the program Veloria is deciding about. Managers trying the lessons themselves may be pleasant for them, but says little about the dispatchers. Choose the group whose result the buyer would trust for the decision she actually has to make.

A pilot decides something only if its rules are fixed before it starts and written down. The agreement with Veloria needs:

• The bar, set with Veloria in advance. Attendance is line one, since that is the finance director's worry: the share of scheduled lessons people actually attend. The number should be Veloria's own target, for example three in four scheduled lessons attended, not an industry average.
• The decision date and the person who decides.
• The exit: what happens if the bar is missed. Veloria stops and owes nothing more.

Each piece closes a gap. Without a bar, any result can be read either way. Without a date, the pilot drifts into an unpaid extension. Without an exit, the buyer's risk isn't really limited. Two common slips: an open end ("until the team feels ready") and an unsourced benchmark, which nobody can check and which turns the pilot into an argument about the benchmark. With a clear exit, the finance director can say yes to the pilot, because a no costs her nothing more.

Attendance shows that people came; it doesn't show they learned. The pilot's second line is progress, and the honest way to measure it is the same short level check at the start and at the end: in week one and in week four. Same test, same conditions, so the difference means something.

Weaker measures are tempting because they're easier. A satisfaction survey only at the end tells you whether people liked the lessons, not whether their English moved. The number of lessons booked repeats the attendance line. A tutor's impression is sincere but uncheckable, and the finance director can't verify it against anything.

The third line is the decision: who decides, on which date, and what a no means. Pavlo says the finance director reads the last line first, so put it last and make it plain.

How it fits together: a company buyer avoids losses, so you lower each person's risk with proof they can check, and with a pilot that has a real group, a written bar, a progress measure and a decision. Next: the module case, mapping Veloria's whole committee before the first meeting.