Reading competitors from public sources

A competitor scan collects what rivals show the public, plus what your own customers tell you openly. Everything else is off limits, however common it is.

Fair sources: their website, offer terms and published prices, public reviews, job ads (they reveal plans and pains), public posts, what comes up when you search the questions customers ask, and honest conversations with customers who compared you.

Never: posing as a customer to hear their pitch (pretexting), asking their staff for internal numbers, joining closed groups under a false name, or using a former employee's files or decks. These are deception or misuse of confidential material, and they put the company's reputation at risk for a few minutes of information.

At Klasna, Ostap offered to book the large online school's sales call posing as a parent. The answer is no. Their public offer pages and reviews tell the team almost everything it needs.

An AI assistant can speed up a competitor scan, under three rules.

1. Public pages in, customer data out. Give it the rivals' published pages and ask for a summary in the scan's fixed columns. Never paste your own customers' data: Klasna's family export doesn't belong in an outside tool.
2. Check every claim against the source. Assistants mix up sources, merge two companies and sometimes invent details. A fact goes into the scan only after you've seen it on the page, with its link and date.
3. Ask how it describes you. Parents ask the same tools which school to pick, and the answer comes from what's written about you on the web. If the assistant describes Klasna vaguely or wrongly, that's a finding for the plan too.

Used this way, the assistant saves reading time and the scan stays as reliable as one done by hand.

A useful scan is a routine, not a raid: the same questions asked of every alternative, repeated on a schedule.

The steps, in order:

1. List three to five alternatives customers actually name (not just the most visible rival).
2. Fix the same questions as columns: whom they speak to, their main promise, the proof they show, where customers complain, source and date, and what's unknown.
3. Collect answers from public sources, each with a link and the day you checked.
4. If an AI assistant did a first pass, check every claim against the original page.
5. Write down what you couldn't find honestly: the Unknown column.
6. Repeat regularly, because offers and prices change.

The fixed columns are what make rivals comparable; the dates show when a fact may be stale; the unknown column keeps the team from filling gaps with guesses or with methods it shouldn't use.

Public reviews are one of the richest fair sources in a scan, and one of the easiest to over-read. Treat a recurring complaint as a clue to test, not as a fact about the market.

Two habits keep it honest. First, count: write how many reviews you read and how many mention the theme, with the date ("12 of 40 reviews, checked 1 October 2026"). Reviews come from a skewed group: the very happy and the very annoyed write; most customers don't. Second, check with your own customers: the complaint is about a rival's customers; what matters is whether the same need matters to the families you want.

At Klasna, 12 of 40 reviews of the large online school complain that tutors changed mid-course. That suggests a steady tutor could matter. Before "your tutor stays with you" goes into any message, Klasna asks its own parents whether it does, and whether Klasna can actually promise it.

When every alternative makes the same promise, customers expect it and don't choose on it. Such a claim is an entry ticket: you must meet it, but you can't win with it.

Uliana's scan found that every rival headline says "experienced tutors", and so does Klasna's. Saying it in bigger letters ("the most experienced tutors") doesn't help: parents can't check it and every school can say the same. Dropping it would be a mistake too, since a school without good tutors fails the basic test.

This is one of the most useful outputs of a scan: it shows what is already claimed, so the team doesn't build its message on it. Klasna still needs qualified tutors; it needs a different reason to be chosen, one the alternatives don't already claim and that Klasna can prove. Finding that reason is the job of the positioning module later in the course.

"I don't have competitors, I'm too small" is a common belief and always wrong. Competition is defined by the customer's alternatives, not by the size of the business. Recall: a competitive alternative is whatever the customer would do without you, including nothing.

Dana teaches English at Klasna, and in September 2026 she started her own practice for adults: three private students from friends, a social-media account with a few posts, a messenger channel and some free evenings. Asked what each student would have done without her: the QA engineer had free group classes at work and an AI chatbot; the student preparing for an exam for a job abroad looked at a large online school's exam course; the bakery owner had been putting it off for two years.

So Dana's market sentence is: adults who'd otherwise use free classes at work, a big school's course, or nothing. It's the same lens Klasna uses, at the size of one person, and it costs her nothing but a few questions.

For a one-person business the lens is the same: for each customer, ask what they'd have done without you, then where that alternative falls short. Each gap suggests a different way to be the better choice, and each stays a hypothesis until the customer confirms it.

Dana's three students:

• The QA engineer had free group classes at work and an AI chatbot for emails. A general group syllabus isn't his job: Dana can offer practice on his own client calls.
• The student preparing for an English exam for a job abroad looked at a large online school's exam course. A standard course isn't her plan: Dana can build one around her exam date and weakest sections.
• The bakery owner had put it off for two years. Postponing wins when starting feels big: a small first step with a fixed weekly slot can beat it.

Dana's offer and message can differ for each type of student, but she should ask each one before relying on the guess.

Competing honestly has a second side: you don't take customers by breaking a commitment. Dana teaches at Klasna and runs her own practice for adults; her contract says she never solicits Klasna's families or learners.

Soliciting means offering your own services to them in any form: a post in a Klasna parents' chat about free evening slots, a word in a lesson, a quiet discount. Leaving out the price doesn't change that. When Klasna parents ask her privately, she points them back to Klasna and finds her adult students elsewhere.

The same principle runs through the whole lesson: public sources and honest questions, nothing taken by deception or from someone else's confidential relationship. That line is what lets Dana do both jobs with a clear conscience, and what lets Klasna stand behind its scan.

How it fits together: an honest scan uses public sources and open questions, follows fixed columns with sources and dates, checks AI output, and treats shared claims as entry tickets; a solo tutor uses the same lens and the same ethics. The next lesson is the module's case: rewriting Uliana's situation row for Leo.