# Opc Influence

> Business persuasion and negotiation for a one-person company. Work out what both sides actually want, what each walks away to, and the shape of an ask the other side can accept. Use when the user says 客户不接受 / 怎么让客户同意 / 报价谈不拢 / 涨价怎么开口 / 对方压价 / 合作分成谈不下来 / how do I get this client to agree / they want a discount / renegotiate this contract. Do not trigger for growth or funnel problems（归 opc-marketing）, for examining one's own judgment or life direction（归 opc-founder）, or for pure copywriting once the terms are already settled.

- Skill: `liustack/opc-influence` (Agent Skill, multi-file: 6 files)
- Install (CLI): `npx skillmds@latest add liustack/opc-influence`
- Raw SKILL.md: https://api.skillmd.com/api/skills/liustack/opc-influence/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Marketing & Growth
- Author: liustack (https://skillmd.com/u/liustack)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/liustack/opc-influence

---


# opc-influence: business persuasion and negotiation

One-person companies negotiate from a structurally weak position: no legal team, no sales floor, and the same person handles delivery afterwards. That shapes everything here. You are not looking for the move that extracts the most from this deal. You are looking for terms that hold up while you still have to deliver against them, with a counterparty you will likely meet again.

This skill covers a specific ask reaching a specific person: a client accepting a price, a partner agreeing to a split, a contract getting reopened. Growth and funnel problems belong to `opc-marketing`. Working out what you yourself want belongs to the thinking stack.

## Step 1: separate position from interest

The stated ask is never the whole thing. Write both columns before anything else.

| | Position (what they say) | Interest (what it is for) |
|---|---|---|
| Them | "we need 20% off" | budget already committed elsewhere, or a signal they doubt the value, or a mandate to extract something from every vendor |
| You | "I need the full rate" | this month's cash flow, or the precedent it sets for the next five clients |

Interests can be traded. Positions can only be split. Most deals that look stuck at a number are two interests that never got named.

If you cannot name their interest, that is the first thing to find out, and it is usually one question away: "what would have to be true for this to be an easy yes?"

The two reference files carry the material this runs on. `references/reading-people.md` covers what to look for in the other side and what moves them, used at Steps 2 and 4. `references/coordination.md` covers getting to agreement once both sides want a deal, used when Step 1 leaves you stuck on shape.

## Step 2: read what they are guarding

Resistance has three sources and they need different handling. Getting this wrong is the most common way an ask dies.

- **A judgment.** They weighed it and concluded no. Evidence can move this. Bring the thing they do not have.
- **A constraint.** Budget is fixed, procurement forbids it, someone above them already decided. Evidence changes nothing. Change the shape of the ask so it fits the constraint, or find who owns it.
- **A story about themselves.** Agreeing would mean admitting they were wrong, or looking soft to their own team. This one is invisible and it kills more deals than the other two combined. Give them a version they can say yes to without contradicting who they told people they are.

The tell: a judgment argues back with reasons, a constraint repeats the same sentence, a story changes the subject.

## Step 3: name both walkaways

Neither side is negotiating in a vacuum. Write down what each of you does if this falls through.

- **Yours.** Concretely: which other client, what runway, how many months. A walkaway you have not costed is not a walkaway, it is a bluff you will fold on.
- **Theirs.** Who else can do this work, at what switching cost, on what timeline. Their alternative is usually worse than they let on, and yours is usually better than it feels at 2am.

Whoever can more comfortably walk sets the terms. If yours is genuinely weak, do not pretend otherwise in your own head. Plan for a smaller ask, or spend the next month fixing the dependency rather than this conversation.

## Step 4: shape the ask

Now build the request. Three parts, in this order:

1. **Their why, not yours.** Open with what changes for them. Cash flow, risk, time, how it looks to their boss. Your costs are not an argument, they are your problem.
2. **A trade, not a concession.** Every give gets a get. Annual prepay in exchange for a discount. Faster delivery in exchange for a narrower scope. A one-sided concession teaches them the price was soft and invites the next round.
3. **An exit they can take.** Name the version you can live with if the main ask fails. Ending with a single all-or-nothing option forces a defensive no.

## Step 5: check what this costs beyond the deal

Before sending, three questions:

- **Can you deliver what you just promised?** A one-person company that wins terms it cannot service has bought a problem, not a client.
- **What does this set up for next time?** Precedents outlive deals. A discount granted once becomes the reference price forever.
- **Would this survive being seen?** If they later learned exactly how the case was built, would the deal still stand? If not, you have not persuaded anyone, you have set a trap that has not sprung yet.

## Hard rules

**Everything you say has to be true.** Adoption numbers you cannot produce, deadlines invented to hurry someone, authority you do not hold, a peer who did not actually do it. These are not aggressive versions of persuasion, they are a different act, and they cost you the relationship the moment they surface. A case that only works when invented was never worth making.

**Their agreement has to survive you leaving the room.** Pressure produces a yes that decays into silence, delay, and a bad review. You want the version they can defend to their own team when you are not there.

**Do not negotiate what has not been decided.** If the user has not settled what they actually want, that is thinking work, not negotiation. Say so and hand it off rather than helping them argue for something they have not chosen.

**Stop at the structure.** Produce the shape of the conversation, the trade, the fallback. If they want the finished email or the exact script, that is delivery work and it comes after the terms are settled.