Insights · 15 Sep 2026 · 8 min read · Iordanis Passas

The trust paradox: why acquaintances hesitate, enterprises buy, and what RUCF says about selling the truth

In three points

Trust in a product dips in the middle of the familiarity range rather than rising along it: close peers and complete strangers both said yes, and the people in between produced the most doubt.
A 15-person agency stalled on hypothetical edge cases while a 250-person firm signed in one meeting. The difference was whether anyone in the room had observed the cost, and size had nothing to do with it.
Both inversions are the evidence hierarchy at work. An opinion about you carries 0.30. A bill someone has already paid carries 1.00. The fix, in sales and in development, is to move the conversation from the first to the second.

Over the past few months I have been pitching Trackilero, a task, leave and capacity tool that tracks time to the exact minute and deliberately leaves out the chat, the automation credits and the custom-field builder that come with a $10-a-seat suite. Nothing about the product surprised me. The pattern in who resisted it did.

The resistance did not follow anything I would have predicted from sales lore. It followed two curves, and both of them are inverted from the story people tell about how trust works.

Fig. 01

The familiarity curve

Close peers have watched you ship for years, which is observed evidence about the builder. Strangers watch the software run, which is observed evidence about the product. Acquaintances have neither, so they fall back on the weakest class there is: an assumption about whether someone like you could plausibly hold the answer.

Why acquaintances hesitate while strangers trust

The assumption in most selling is that trust is linear. The better someone knows you, the more they trust what you make. What actually happened was a valley. People who had worked alongside me for years said yes quickly. People who had never heard of me said yes almost as quickly. The people in between, the former collaborators, the contacts from a previous job, the names in the same network, produced almost all of the hesitation.

The acquaintance is in a difficult position, and it is worth being fair to them. They remember you in a different role: a peer, someone who once sat in the same meetings, someone in the same city. Social proximity makes it hard to accept that you now hold the answer to a problem they have, because if the answer were that available, they feel they should have had it too. They also lack what the close peer has, which is direct knowledge of how rigorously you build. So they cannot evaluate you as an authority and they cannot evaluate you as a known quantity.

What they can do is form an opinion, and in RUCF an opinion about a product from someone who has not used it is assumed evidence at 0.30. That is the whole mechanism. The acquaintance is not more sceptical by temperament. They are evaluating with the weakest instrument available, and a weak instrument produces a cautious reading. They look for the reasons the software might fail rather than for whether it removes a cost, because looking for failure is what you do when you are judging a person and not a thing.

The stranger has none of this. They open Trackilero and see whether three-level subtask rollups add up without double-counting, whether the timer refuses a 13-hour entry, whether capacity planning already knows who is on leave. There is no history to colour any of it. The software either solves the operational headache in front of them or it does not, and either way they find out in an afternoon.

Why the small team stalled and the large one signed

The second inversion was stranger, because it ran against the one piece of startup folklore everybody agrees on. Small teams are supposed to be fast and unbureaucratic. Enterprises are supposed to have procurement committees and eighteen-month cycles. A 15-person agency spent three meetings in speculative doubt and did not buy. A 250-person company saw one demonstration and committed.

Fig. 02

The scale inversion

The agency was slower because nobody in it had ever seen a number for what their current tooling cost, so the only evidence available was opinion, and opinion is what the meeting produced. The enterprise had paid the bill and remembered the amount.

The 15-person team: the evaluator effect in the room

In a company of fifteen, operations are rarely instrumented. Nobody has measured how many billable minutes fall through rounded quarter-hours, or how often a project margin eroded because the subtask hours in a spreadsheet were summed twice. The leadership has a feeling about their tooling, and the feeling is that it is basically fine.

Without observed evidence, a purchasing decision becomes a group exercise in speculation, and speculation has a documented failure mode. Hertzum and Jacobsen reviewed eleven studies of people inspecting the same system and found agreement between evaluators ranging from 5% to 65%, with each evaluator confident their own list was the real one. That is what a pitch meeting with no data looks like from the inside. Each person invents a hypothetical their neighbour did not think of: fifteen custom statuses, nested automations, an AI summary of the week. Each feels like diligence. Collectively they are a list of things the team has never needed, weighted as though they were requirements.

In that room, feature count reads as competence. Because nobody knows the price of their own friction, a tool that removes things at $2 a seat, with five fixed statuses and no chat, feels like a risk rather than a discipline. There is no number in the room large enough to make subtraction look responsible.

The 250-person firm: the weight of an observed cost

A company of 250 people lives inside the consequences of scale, and it has receipts. A project manager promised a client a Thursday delivery for an engineer who had booked Thursday and Friday off. An invoice went to a client with subtask hours billed twice, and the client noticed first. They are paying $19 a seat for 250 seats, which is $4,750 a month, for a suite where most of the staff touch the timer and the task list and nothing else, while the automation credits and the AI meter fluctuate in ways nobody can forecast.

None of this was hypothetical. All of it was observed evidence at 1.00, already in their finance system and in at least one uncomfortable conversation. So the operator sitting across the table did not need to decide whether to trust me. They had never met me. They looked at a tool that handles tasks, exact minutes, leave and capacity without arithmetic errors, compared it to the arithmetic errors they had already paid for, and the decision was made by subtraction. They never had to trust the person, because the sums did the work.

Moving the conversation from 0.30 to 1.00

Once both inversions are seen as one mechanism, the remedy is shared as well. Both the acquaintance and the small team are evaluating with weak evidence because weak evidence is all the conversation offered them. A pitch built on the founder's confidence or a feature list invites opinions, and opinions weigh 0.30 no matter who holds them. The job is to put something in the room that weighs more.

RUCF calls the distance between what people believe a product costs them and what it observably costs them the awareness gap, and it is the right frame for a pitch. Do not debate features. Audit the silent friction. When a prospect asks for custom-field sprawl or an AI assistant, the wrong move is to defend the design philosophically. The right move is to reframe it as cost: most teams believe their time tracking is accurate, and in the teams we have measured, untracked gaps and rounded timers leak somewhere between eight and twelve per cent of billable revenue every month. Trackilero exists to stop that leak and nothing else. That sentence is about their money, not my product, and it is checkable.

The same reframe dissolves the acquaintance problem, because it takes the founder out of the sentence. "Here is what I built" asks them to assess me. "Here is where the data shows billable teams bleed time" asks them to assess a finding. A diagnostician presenting an audit does not trigger the social comparison that a peer presenting a project does. The awkwardness came from who was standing next to the product, never from the product itself.

Building for silent friction, not vocal demands

The development version of this mistake is more expensive than the sales version, because it ships. Founders get derailed by casual testers and early prospects who ask for more buttons, more integrations and more settings, and each request arrives with a confident voice attached. The say-do matrix is the tool for sorting them, because it insists on comparing the request to the behaviour before acting on either.

State	What they said	What they were doing	The trap	The rule
Silent friction	"Our current tool is fine. We just need more custom fields."	Losing hours to double-counted subtasks and overlapping holiday bookings	Build the custom fields	Subtract the step. Automate the capacity arithmetic, cap the timer at twelve hours. Fix what they bleed, not what they ask for.
Phantom friction	"Five fixed statuses feels limiting."	Moving work through the pipeline faster because every status means one thing to everyone	Build a custom status builder that breaks rollup reporting	Change the framing, not the code. Explain that fixed statuses are what keep the reports honest.

Both rows describe a request that sounds reasonable and a behaviour that contradicts it. In the first, the pain is real and the request points away from it. In the second, the pain is imagined and the request would create a real one. Neither is answered by building what was asked. Trust in a product accumulates because it works quietly for months, not because it capitulated to every suggestion made during the demo.

Kill the consensus sign-off

The last place the trust paradox shows up is the approval, whether that is an internal review of a feature or a buyer's committee deciding on a tool. Subjective review destroys momentum for the same reason the 15-person meeting did: it lets a room vote on a feeling. The manifesto puts it bluntly. A design with no named signal has no way to fail, and without numeric acceptance criteria written before the work starts, review becomes a matter of taste.

Fig. 03

A gate instead of a vote

A committee can always find a reason to defer. A gate written down before the week starts cannot be argued with afterwards, which is the point: it converts the buying decision from a judgement about the founder into a reading off the product, and a reading is something a stranger and an acquaintance will agree on.

When seeking buy-in, refuse the assumed consensus. Do not let a group approve or reject a tool on whether they like the interface. Offer a falsifiable gate instead: three team members run Trackilero alongside the current spreadsheet or suite for one week. It passes if subtask rollups match the invoice drafts without anyone recalculating by hand, and if the time to log a task drops below ten seconds. It fails otherwise. Write both conditions down before Monday, and the meeting on the following Monday is ten minutes long.

Trust is an artifact of measurement

Strangers buy and enterprises commit for the same reason, and it is not credulity. They are measuring the distance between a cost and an outcome, with nothing personal in the way. Acquaintances second-guess you because they are evaluating you. Small, unmeasured teams hesitate because they are evaluating each other's opinions. In both cases the evidence in the room is the weakest class RUCF recognises, and the reading it produces is exactly as cautious as a 0.30 weight says it should be.

The lesson for selling and the lesson for building turn out to be one lesson. Stop trying to win the argument about opinions. Change the ground to observed behaviour, quantify the silent friction that everyone in the room has felt and nobody has counted, and let the arithmetic close.

Framework this draws on

The evidence classes in Layer 2 (observed 1.00, reported 0.70, inferred 0.50, assumed 0.30), the say-do matrix for sorting requests against behaviour, and the manifesto's requirement that acceptance criteria be numeric and written before the work begins.

Find your gap