Client Ignores Your Advice: Push Back or Let It Go?
就之不用，去之反求。 — Approach him and you go unused; leave, and he comes seeking you. (《鬼谷子·內揵》, my translation)
Short answer: do it their way when the call is reversible, and put your reasoning in writing once. Fight only when the thing can't be undone. Being overruled is usually not a verdict on your judgment — holding power makes people discount advice, and it makes their decisions measurably worse.
That second part is from a 2011 paper with four studies behind it. The first part is from a Chinese persuasion manual whose commentary, written around the sixth century CE, describes the exact sequence where a client rejects your plan and then asks for it six months later.
Key Takeaways
- Power itself drives advice discounting — and it lowered accuracy in the one study that measured both.
- Rejection usually tracks fit and timing, not whether you were right.
- Sort every disagreement by reversibility before you decide whether to argue.
- A correct recommendation delivered as a verdict gets refused and costs you the relationship.
Why do clients ignore advice they're paying for?
Because they hold the power in the room, and power does something specific to advice-taking. See, Morrison, Rothman and Soll tested this across four studies. One was a field survey of professionals rated by their actual coworkers; others manipulated power directly. All of them showed the same negative relationship between power and advice taking, mediated by confidence. Their fourth study found that higher-power participants ended up with less accurate final judgments.
Their own summary is blunter than mine. Power, they write, "can thus exacerbate the tendency for people to overweight their own initial judgment, such that the most powerful decision makers can also be the least accurate."
So the client who overrides you isn't necessarily a difficult person. Occupying the buyer's seat raises confidence, and raised confidence is what predicts discounting your input — which means most of the internet's "how to handle a difficult client" framing is aimed at the wrong cause.
How much of your advice actually lands?
Rarely all of it, and there's a standard measure that makes the point. Researchers call it weight on advice — the proportional shift from someone's starting position toward the adviser's:
weight on advice = (final answer − initial answer) / (advisor's answer − initial answer)
A score of 1 means they fully adopted your recommendation; 0 means they didn't move at all. See and colleagues summarise the robust background finding plainly: people tend not to alter their judgments much on input from others.
Treating advice as binary — taken or ignored — hides what actually happened. Your client probably moved partway, and partway is the normal result rather than a failure, so the useful question is how far they moved and on which specific decisions.
Is your advice wrong, or just badly fitted?
Almost always the second, and the Guiguzi is unusually direct about it. The Neijian chapter states the failure condition: 「不見其類而爲之者，見逆；不得其情而說之者，見非」 — act without seeing what kind of thing you're dealing with and you meet resistance; speak without having grasped their actual situation and you are rejected.
Note what's being blamed. Not the content of the recommendation. The missing item is 情 — their real circumstances, pressures and motives. The text's remedy follows immediately: 「得其情乃制其術」, get their situation and only then can you command the method.
The text treats rejection as diagnostic information about your grasp of the client's position, not as a referendum on your expertise. That inverts the usual reaction, which is to restate the recommendation more forcefully — the one move that cannot fix a fit problem.
What does "their actual situation" mean on a real project?
Concrete things you probably don't have. Who promised what to whom before you were hired. Which budget line this comes out of, and when that line resets. Whether your contact is being measured on the metric your recommendation would improve, or on a different one. Whether someone senior already rejected your idea two years ago under a different name.
The Neijian chapter's rule is 「夫內有不合者，不可施行也」 — what does not fit internally cannot be carried out. Not "should not." Cannot.
A technically correct recommendation that collides with an internal commitment you can't see will lose every time, and it will look to you like the client being irrational. Before arguing the merits again, spend one question on what you don't know: what would have to be true inside their organisation for your version to be the easy choice?
Should you tell a client plainly that they're wrong?
Tell them the trade-off plainly. Don't deliver a verdict. The Guiguzi separates these two moves in a single line — 「隂慮可否，明言得失，以御其志」: privately weigh whether it's feasible, then plainly state gains and losses, and so steer their intention.
The private half is judgment. The public half is 得失, gains and losses — not "this is a mistake." And the Jue chapter names the cost of getting that wrong: 「若有利於善者，隱託於惡，則不受矣。致疎逺」. When something genuinely beneficial is couched in negative terms, it isn't accepted, and estrangement follows.
That's a sharper claim than the usual "be tactful." It says a correct recommendation wrapped in a negative frame produces two losses, not one: the advice is refused, and the relationship cools. Most of the standard advice online — document your objection, go on record — is precisely this failure mode performed deliberately.
Three ways to say the same true thing
| Framing | What the client hears | What it costs you |
|---|---|---|
| Verdict — "this won't work" | A challenge to their authority | Refusal plus distance (the 致疎逺 problem) |
| Trade-off — "this buys X and costs Y" | Information for a decision they own | Nothing; this is the 明言得失 move |
| Test — "let's find out cheaply by Friday" | A way to be wrong without losing face | A little time, and it ends the argument |
The third row is the one people skip. A cheap reversible test converts an argument about who is right into a question with an answer date, and it lets the client change position without conceding anything.
When should you just do it their way?
When the call is reversible and they've heard the trade-off once. The Jue chapter gives a three-part check before committing to a decision: 「度之往事，驗之來事，參之平素，可則決之」 — measure it against what has already happened, test it against what is coming, check it against the ordinary baseline; if it's workable, decide.
Applied to your situation, the useful axis is reversibility.
| Reversible | Irreversible | |
|---|---|---|
| Examples | Copy, layout, pricing page, campaign creative, most design choices | Data migrations, domain and URL changes, contracts signed, public launches, deleted records |
| Your move | State the trade-off once in writing, then build it their way well | Do not proceed on a verbal override; get the decision explicit and dated |
| Why | The cost of being wrong is a revision | The cost of being wrong is unrecoverable, and it will be attributed to you |
The reversibility test resolves most of these disputes without a confrontation, because most client overrides are reversible. And on a reversible call, shipping their version and being proven right beats winning the argument.
What if the decision can't be undone?
Then the fit question becomes a stay-or-go question, and the Guiguzi is willing to say so: 「事有不合者，聖人不爲謀也」 — where there is no fit, the sage does not plan for them. The Neijian chapter closes on withdrawal as the mature option, not the failed one: 「環轉因化，莫知所爲，退爲大儀」 — turn with the circumstances as a ring turns, and to withdraw is the great norm.
In practice that means declining the specific piece of work rather than the client. "I'll do the rebuild, but I can't be the one to run the migration without a rollback plan" is a boundary on one irreversible act. If the answer is that you must do it anyway, you've learned the engagement's real terms — which is the same signal that ends a business partnership worth exiting.
Why do they come back later asking for your original plan?
Because rejection at the time and correctness over time are different things, and the text names the pattern. The Neijian chapter's opening grid includes 「就之不用，去之反求」 — you approach and go unused; you leave and they come seeking. Its own later gloss: 「就而不用者，䇿不得也。去而反求者，事中來也」.
Then Tao Hongjing's commentary, transmitted with this edition, explains that last phrase: 「謂所言當時未合，事過始驗」 — what was said did not fit at the time; only after the event did it prove out.
That is a description of the "client hired someone else, it failed, now they want my original proposal" experience, written roughly fifteen centuries before anyone had a client. It also tells you what to do with the feeling: the gap between being right and being adopted is a timing property of advice, not evidence that you communicated badly.
The Jue chapter adds the reason the decision was never yours to win: 「凡決物，必託於疑者」 — all deciding rests on the one who holds the doubt. Your job is the situation and the trade-offs. The choosing belongs to whoever carries the consequence.
Does a paper trail actually protect you?
It protects your invoice and your reputation. It does not protect the outcome, and if you write it as an accusation it actively damages the relationship — that's the 致疎逺 problem again, in email form.
Write it once, neutrally — a record of the trade-off, not of your objection. Four lines do it: the option chosen, what it optimises for, the cost accepted, and what would change the plan. That version stays useful to the client later, instead of becoming a weapon you're holding.
One neutral written record of the trade-off is worth more than three escalating arguments, and it's the only form of "documenting it" that doesn't buy protection at the price of the relationship.
How is this different from scope creep?
Different problem, different fix, and they get conflated constantly.
| Situation | What's missing | What fixes it |
|---|---|---|
| Scope creep | A boundary — this is new work | Price it |
| Endless revisions | A finish line — nobody defined done | Define done up front |
| This one | Fit — your judgment lost to theirs inside agreed work | Sort by reversibility; convert verdicts into tests |
Pricing does nothing here, because the client already paid for this work. Defining "done" does nothing either, because you agree on what done means — you disagree on how to get there. And if the disagreement is really about who owns what you produce, that's a different question again.
What this doesn't license
Two limits worth stating.
The Guiguzi is a manual for influencing rulers, and it is frank about manipulation in a way most readers will not want to adopt. Its diagnosis of why advice fails is excellent. Its posture — study the powerful person's desires in order to attach yourself to them — is not a client relationship anyone should want. Import the analysis, not the stance.
And the mirror case is real. Sometimes the client's "bad idea" is correct because they hold information you don't: the 情 problem runs both directions. The Guiguzi notes that even the ancient kings used divination 「以自決也」 rather than trusting their own judgment alone. Tao Hongjing's gloss presses harder, asking whether anyone lesser should trust only themselves without consulting those who understand. Applied honestly, that line is aimed at you too.
Frequently Asked Questions
My client ignored my advice and the project failed. Whose fault is it?
Theirs for the decision, yours for the record. If you stated the trade-off once in writing and built their version competently, you did the job. The useful move now isn't a reckoning — it's noticing that the 「去之反求」 pattern usually follows, and deciding in advance what you'll charge to come back.
How do I tell a client they're wrong without damaging the relationship?
Don't. State the trade-off and offer a cheap test with a date on it. The Guiguzi's 「明言得失」 — plainly state gains and losses — is a different act from issuing a verdict, and the Jue chapter specifically warns that beneficial advice framed negatively gets refused and produces estrangement.
Is it unprofessional to build something I think is a mistake?
Not if it's reversible and you said so once. It is unprofessional to perform an irreversible act on a verbal override without an explicit, dated decision.
Why does power make people worse at taking advice?
Confidence appears to be the mechanism. In See and colleagues' four studies the link between power and reduced advice taking was mediated by confidence in one's own judgment, and in their fourth study the higher-power group was less accurate despite being more confident.
Who was Guiguzi?
A Warring States figure traditionally named as the teacher of the strategists Su Qin and Zhang Yi, and the attributed author of the earliest Chinese manual on persuasion and reading other people. Both the dating and the authorship are contested; the text survives with a commentary attributed to Tao Hongjing (456–536 CE).
Sources and a note on the texts
On the texts, the translations and the limits. The Chinese for 《鬼谷子·內揵第三》 and 《決篇第十一》 is quoted from the Chinese Wikisource edition of the Tao Hongjing commentary (嘉慶十年江都秦氏開雕), retrieved as raw wikitext and checked line by line; ctext.org is not used, as it blocks automated retrieval. That edition preserves variant graphs — 隂 for 陰, 逺 for 遠, 䇿 for 策, 疎 for 疏 — and I've kept them as printed rather than silently normalising. English renderings are mine and deliberately literal. The Guiguzi's date, authorship and textual integrity are all disputed, and some traditional scholars regarded it as a later compilation; nothing here depends on it being by one hand. On the modern side, I quote See et al. (2011) from the authors' own institutional copy. Its first study sampled 208 incoming business graduate students rated by coworkers rather than a random sample of clients, and the causal evidence comes from its experiments — the effect is well-evidenced for power in general, not specifically measured on buyers of freelance services. Yaniv & Kleinberger (2000) and Bonaccio & Dalal (2006) are cited here as the sources See et al. themselves credit for "egocentric discounting" and for the review literature; I verified their bibliographic records but could not obtain their full texts, so I make no claim about their specific numbers. None of this is legal advice.
- 《鬼谷子·內揵第三》, in 鬼谷子/卷01, Chinese Wikisource — https://zh.wikisource.org/wiki/鬼谷子/卷01 (retrieved 2026-09-30)
- 《鬼谷子·決篇第十一》, in 鬼谷子/卷02, Chinese Wikisource — https://zh.wikisource.org/wiki/鬼谷子/卷02 (retrieved 2026-09-30)
- Kelly E. See, Elizabeth Wolfe Morrison, Naomi B. Rothman and Jack B. Soll, "The detrimental effects of power on confidence, advice taking, and accuracy," Organizational Behavior and Human Decision Processes 116(2), 272–285 (2011) — https://doi.org/10.1016/j.obhdp.2011.07.006 · full text retrieved from NYU Stern: https://web-docs.stern.nyu.edu/pa/ksee_power_advice_taking.pdf (retrieved 2026-09-30)
- Ilan Yaniv and Eli Kleinberger, "Advice Taking in Decision Making: Egocentric Discounting and Reputation Formation," Organizational Behavior and Human Decision Processes 83(2), 260–281 (2000) — https://doi.org/10.1006/obhd.2000.2909 (bibliographic record verified 2026-09-30)
- Silvia Bonaccio and Reeshad S. Dalal, "Advice taking and decision-making: An integrative literature review, and implications for the organizational sciences," Organizational Behavior and Human Decision Processes 101(2), 127–151 (2006) — https://doi.org/10.1016/j.obhdp.2006.07.001 (bibliographic record verified 2026-09-30)
- Jack B. Soll and Richard P. Larrick, "Strategies for revising judgment: How (and how well) people use others' opinions," Journal of Experimental Psychology: Learning, Memory, and Cognition 35(3), 780–805 (2009) — https://doi.org/10.1037/a0015145 (bibliographic record verified 2026-09-30)
Related reading
- How to Negotiate Without Getting Anxious
- What Guiguzi Would Tell a Sales Rep Negotiating a Big Deal
- When a Prospect Goes Silent After a Great Call
- Telling a Client Something Is Out of Scope
- When Should You End a Business Partnership?
Put your own case to him
The reversibility test is general; your situation isn't. Whether to push back depends on what the client already promised upstairs, what you can't see, and how much of the relationship you're willing to spend. Take the actual disagreement to Guiguzi himself and argue it out, or read about who he was first.