Insights / startup-foundations

Why Positive Feedback Is Weak Evidence of Demand

By Daniel Tonkopiy | 2026-11-03

Positive feedback is weak evidence of demand because praise costs the person giving it nothing, and politeness and the interview itself push answers toward agreement. Founders who have heard a lot of praise before making any offer should sort each remark by what it actually shows: how often the problem happens, what the current workaround costs and who carries that cost, and whether anyone gave up money or time. The main constraint: only records and completed steps count.

The symptom and what it does not prove

This guide is for founders at the idea-to-validation stage who have talked to potential customers, heard a lot of praise and have not yet asked anyone to pay. The notes are full of 'great idea' and 'would use that', and the question is whether that adds up to demand. If an offer has already gone out and nobody bought, read When Customer Interviews Say Yes but Nobody Buys instead. That guide diagnoses a failed offer. This one diagnoses the praise that comes before any offer.

Positive feedback proves less than it feels like it does. The person giving it pays nothing, and the setting pushes toward agreement. Pew Research Center's guide to writing survey questions describes social desirability bias and says research has shown it can be greater when an interviewer is present. The same guide describes acquiescence bias, a tendency of some respondents to agree with agree-disagree statements, and says that behavior is even more pronounced when an interviewer is present. It also lists the likelihood that people will vote in an election among the things respondents may overstate. A founder asking about an idea face to face is, in effect, an interviewer asking for agreement about a future action.

Stated intentions are a weak forecast even when they are sincere. Vicki Morwitz's review of research on purchase intentions and behavior in Foundations and Trends in Marketing, published in 2014, says in its abstract: 'Purchase intentions are correlated and predict future sales, but do so imperfectly.' That review concerns consumer purchase intentions, not founder interviews, but the lesson carries over: a promise to buy is a signal to test, not a sale.

The U.S. Small Business Administration's market research guide asks two separate questions: 'Demand: Is there a desire for your product or service?' and 'Pricing: What do potential customers pay for these alternatives?' Praise speaks to desire at most. Whether the problem is worth paying to fix is the second question, and praise does not answer it.

The useful distinction is between a frequent inconvenience and a paid problem. Both produce complaints, and a frequent inconvenience often produces more of them, because the people who feel it every day are usually the people a founder gets to talk to. A paid problem shows up as money or paid hours in the account's own records, and the person who carries that cost will give something up to reduce it.

Sort the praise before reading it

Before deciding anything, put every positive remark from the notes on one sheet and sort it by type. The sort below is a problem severity scorecard adapted for diagnosis: instead of filling in blank fields for a new account, it re-reads feedback already collected and shows which severity question each remark can answer. For the full blank sheet, use the Problem Validation Scorecard: Pain, Frequency, and Budget.

Tag each remark said, seen, done or assumed, as in the Customer Interview Notes Template for Founders. Then apply one rule: the first three types count toward nothing. A compliment, a complaint and a stated intention can point to where to look, but none of them can move a block of the scorecard up from Missing. General frequency talk and told events can make a block Partial. Only records and completed steps can make it Strong.

Also note who said each remark: the person who feels the problem, the person who pays for it, or both. Praise usually comes from the first group. A sort in which every remark comes from people without a budget has found a frequent inconvenience for those people, not yet a paid problem for anyone.

Check problem frequency

Frequency is where praise misleads first, because 'all the time' sounds like a count. Turn each frequency remark into a request for a record:

• Ask for the last occurrence. The date, what happened and who dealt with it. A remark that cannot be dated stays tagged said.
• Ask for a named period. How many times in the last full month or quarter, from a log, calendar, ticket queue or job notes.
• Ask about the pattern. Whether it clusters in a season, on certain days or with certain customers, so that a busy week is not read as a yearly rate.
• Write down the scaling. If you turn a month into a year, tag the yearly figure assumed and note how you got it.

A high count is not the finish line. A problem can happen every day and cost the account almost nothing, which is exactly what a frequent inconvenience is. Frequency multiplies cost. It does not replace it.

Investigate cost of the current workaround

Every account that has the problem already handles it somehow, even if that means putting up with it. Cost the workaround per occurrence, then multiply by the count:

• Money out. Credits, refunds, fees, overtime, extra trips or contractors, from invoices or credit notes.
• Paid hours. Hours that would really be cut or moved to paid work, at the account's own rate.
• Revenue lost. Jobs, orders or customers lost because of the problem.
• Who carries it. The account, its own customer, a supplier, or an employee who absorbs it.

The last line settles more diagnoses than the others. The warmest praise often comes from people who feel the problem but do not pay for it. If the account bills its own customer for the extra time, passes a fee on, or relies on staff who are on shift anyway, the problem is real and frequent, but the account has little reason to pay to remove it. Compare the yearly cost with the price you have in mind, set the required margin before the answers come in, and check the largest cost item against a record. The SBA question about what potential customers pay for alternatives is the comparison the buyer will make, whether or not the founder makes it first.

Fix the gap in willingness to commit money or time

When the sort shows praise but no commitment, the fix is to ask for something that costs the person giving it. Praise is free, so only a step with a cost can test it. Make the request to the person who carries the cost, after the cost is written down, so the request has a figure to stand on.

Daniel's go-to-market mentoring page describes part of this step: 'Name the user, the budget owner and the event that makes the problem urgent.' That is his own description of his method. In praise terms, the user is usually the one who praised, and the budget owner is the one to ask.

Keep the request honest and small enough to be a real test. A free trial, a waitlist signup or 'send it over when it is ready' costs the buyer nothing and stays on the praise side of the sort.

A hypothetical example: praise from moving crews

The following is a hypothetical example with invented details. It illustrates the method and does not describe a real company, customer or measured result.

A founder plans a service that arranges a parking space at the door for small moving companies on move day, at an intended $99 a month, or $1,188 a year, with a required margin of twice the price, or $2,376. On a ride-along with the first company, the founder watches two crew members in green T-shirts push a customer's wooden table on its wheeled legs half a block down a tree-lined sidewalk, because the truck could not park at the door. Over two weeks the founder collects praise from two companies and sorts it.

Sorted by enthusiasm, Company A looks like strong demand and Company B looks indifferent. Then the founder asks both owners for records and for one costly step.

The company that praised most has the higher count and almost no cost, because its customers pay for the long carries by the hour. The lukewarm owner carries the cost directly and paid for the pilot. Read alone, the praise would have sent the founder after the wrong account. The founder keeps the assumptions next to the result: that the last 4 weeks are typical, although moving work is seasonal; that 45 minutes per carry is fair; that the crew hours at Company B would otherwise go to paid work or save overtime; that the dispute credit is not a sign of a larger hidden cost at Company A; and that a $200 pilot says something about paying $1,188 a year. Two accounts are no verdict on the segment. The next step is more records from flat-rate movers, and a separate look at whether the customers who pay for long carries by the hour are a buyer in their own right.

Evidence that the targeted fix is working

Decide in advance what would show that the fix worked, and write it as completed steps, not opinions. The targeted fix is the request for a costly step, so the evidence lies in what buyers do next.

Give the fix a fixed window, such as the next round of conversations, and judge it by completed steps per account asked, not by the number of positive remarks. If accounts with a real cost still refuse every step, the problem may be paid while the offer, the timing or the buyer is wrong, so check those before changing the product. If the cost keeps landing on someone else, consider whether that someone is the real buyer.

Common mistakes and limits

• Counting compliments as votes. More praise adds nothing if the seen and done rows stay empty.
• Treating the loudest complaint as the largest cost. The people who feel a problem most are often not the people who pay for it.
• Asking whether it is a good idea. The question invites agreement, and the interview setting already pushes that way.
• Accepting a promise as commitment. A promise to sign up, a waitlist entry or a free trial costs the buyer nothing.
• Reading frequency alone. A daily annoyance with no cost is still an inconvenience.
• Building before asking. More product before any costly request turns praise into a plan without a test.
• Discarding the lukewarm. A flat, careful answer from the person who carries the cost can be worth more than warm praise from someone who does not.

The method has limits. It suits business buyers with an identifiable budget owner and records to look at. Consumer products may need purchase and usage data instead of interviews. With a handful of accounts, a reading is a judgment about what to test next, not a demand estimate. In this guide, the Pew, SBA and Morwitz statements are independently sourced facts, Daniel's method wording is his own description, the sort, its rule and the readings are editorial opinion, and the moving company case is hypothetical.

Related guides and next step

• The wider build, narrow or stop decision: Validate a Startup Idea Before Building
• Comparing several candidate problems: Is Your Startup Solving a Problem People Will Pay to Fix?
• Running the test across a set of accounts: How to Test a Customer Problem Before Building a Product
• The blank sheet for one account: Problem Validation Scorecard: Pain, Frequency, and Budget
• When an offer went out and nobody bought: When Customer Interviews Say Yes but Nobody Buys

Daniel's go-to-market page says the first session is free and takes 30 minutes. Book a startup mentoring session to review your problem validation decision with your actual evidence and constraints, and bring the sorted praise with the records behind it.

Frequently asked questions

Why is positive feedback weak evidence of demand?

Praise costs the person giving it nothing, and survey research describes people giving agreeable or socially desirable answers, especially with an interviewer present. A compliment, a complaint or a promise to sign up says the problem is familiar. It does not show how often the problem happens, what it costs the account or whether anyone will give up money or time to fix it.

What evidence should a founder collect for problem validation?

A dated count of occurrences from a log or other record, the cost of the current workaround with its largest item checked against an invoice or payroll line, the person whose budget carries that cost, and a completed step from that person, such as a paid pilot or a held meeting on a set date. Tag each item said, seen, done or assumed, and keep refusals in the record.

What are common mistakes in problem validation?

Counting compliments as votes, taking the loudest complaint as the largest cost, asking people who feel the problem instead of people who pay for it, accepting a promise to sign up as commitment, treating a general frequency claim as a count, and changing the product before asking for anything that costs the buyer something.

How can a founder tell a frequent inconvenience from a paid problem?

Check who carries the cost. A frequent inconvenience happens often, but its cost is small, passed on to someone else or absorbed by staff who are paid anyway, so nobody gives anything up to fix it. A paid problem shows up as money or paid hours in the account's own records, and the person who carries that cost completes a step with a price or a date.

Sources

1. Writing Survey Questions, Pew Research Center Checked 2026-11-03
2. Consumers' Purchase Intentions and their Behavior (Vicki Morwitz, Foundations and Trends in Marketing), abstract record on EconPapers Checked 2026-11-03
3. Market research and competitive analysis (Plan your business guide), U.S. Small Business Administration Checked 2026-11-03
4. Go-to-Market Mentor and GTM Strategy Advisor Checked 2026-11-03

Problem validation | Positive feedback | Startup foundations | Diagnosis | Founder playbook