Proof
Security, legal, reference calls, and a pilot whose success criterion cannot be measured by anyone in the room. The file waits for a sentence like “this worked,” and nobody owns the sentence. More feature detail rarely moves it.
Claude · Friction
The purchase stalls in a specific place. Name that place.
Teams often describe a slow deal as “they need more education” or “procurement is hard.” Those sentences hide the mechanism. Claude work treats friction as a repeated obstacle inside a decision: a proof nobody can sign, a person who was never in the first meeting, a switching cost the champion waved away, a calendar that was never going to open, or a doubt that the problem is a product problem at all. Patterns here come from interviews and scenes. They are not percentages of a market.
Five frictions worth separating
Consensus
The champion’s map of the buying group was smaller than the real one. A finance partner, a works-council equivalent, a regional operator, or a security architect appears in week six. See how roles are named in decision-maker interviews.
Switching
Data that will not move, a team that must be retrained during a peak season, a contract that overlaps, an incumbent who is “good enough” in the one workflow that matters. The alternative is often a spreadsheet with a person’s name on it.
Timing and category doubt
Budget cycles, a prior failure that is still politically hot, or a buyer who does not believe the pain is something a vendor should fix. Timing friction waits. Category friction argues. Do not collapse them into “no urgency.”
A composite, labeled as a composite
The following sketch is a teaching composite. It borrows the shape of evaluations researchers hear. It is not an account of a named company, and nothing in it is a rate you may quote as “how buyers behave.”
A data-access tool is sponsored by an analytics lead who is tired of ticket queues. In week two the shortlist has three vendors and a plan to “just use the warehouse team.” A pilot is scoped as “time saved for analysts.” Security sends a questionnaire the champion has never read. Eleven weeks pass. The pilot produces enthusiastic quotes and no clock anyone will defend. Legal asks who owns residual data if the trial ends. Finance asks for a payback slide. The analytics lead builds one. The operator who would have to change the Monday meeting says, in a side channel, that the last tool “burned” the night shift. The purchase does not die in a single no. It thins out. The incumbent process remains.
A Claude reading separates the stalls instead of blaming “the enterprise.” Proof friction: a success line nobody could sign. Consensus friction: the night-shift operator was not in the kickoff. Switching friction: a prior burn still sets the political cost. The payback slide was a costume worn late. It did not cause the stall, and it would be a mistake to leave the study believing the lesson is “show ROI earlier.” That belief is how teams train buyers to ask for a number neither side can audit. If you need the words the operator actually used, keep them in the decision language bank rather than translating “burned” into “implementation risk.”
How a pattern earns its name
You may call something a pattern when you have heard the mechanism in more than one decision, you can describe a case that does not fit, and you can point to an artifact or a sentence rather than a mood. “Several champions said the security form arrived before anyone agreed what the pilot had to prove” is a pattern candidate. “Enterprises care about security” is a poster. Write the exception on purpose: the team that sailed through security because a sister company had already signed the same architecture. Exceptions keep you from smuggling a universal into a footnote.
Saturation, in this guide, means new conversations are repeating a friction you can already narrate, not that you have covered a segment. Say that aloud in the memo. The design choices that make this honest are in research design. The place the stall becomes visible — questionnaire, standup, redline — belongs on the touchpoint map.
Questions people bring to this chapter
Should we score friction on a scale?
A scale implies units you do not have. Rank the stalls only inside one decision: what froze the file longest, in that chronology. Across decisions, compare mechanisms, not points.
Is “price” a friction?
Price is sometimes the costume of another stall. Ask what would have to be true for the number to be acceptable, and what they did on the deals where the number was fine and they still waited. If the answer is a proof they cannot sign, you are not in a pricing study.
Can we tell sales to “remove” the friction?
Some stalls are removed by a clearer artifact: a rollback plan, a named owner for the pilot clock, a security answer that matches the architecture. Some are facts about the buyer’s year. Positioning can acknowledge a stall. It cannot promise the stall will pay for itself.
Turn the stall into a sentence
Positioning synthesis starts from a friction you are willing to be about, in language a buyer already uses.