"Sure, we can customize that" is the most expensive sentence in the deal

People don't measure "benefit now" and "cost later" with the same ruler: the further out it sits, the more the weight collapses. Layer an organizational mismatch on top — the cost doesn't even hit your team — and "sure, we can customize that" becomes a sentence with almost no friction behind it.

• Hard evidence · discounting is hyperbolic, and the near term is systematically inflated: Ainslie (1975, Psychological Bulletin) and Kirby & Herrnstein (1995, Psychological Science) show that people discount delayed gains and costs on a hyperbolic rather than exponential curve — weights near the present are inflated absurdly, which is why preferences reverse over time (the commitment that looked cheap today is regretted when it comes due). Peer-reviewed, replicated across species and populations. Hard.
• Hard evidence, with a caveat · immediate rewards are louder in the brain: McClure, Laibson, Loewenstein & Cohen (2004, Science) found stronger activation in the ventral striatum, medial prefrontal cortex (vmPFC) and posterior cingulate when an immediately available reward is on the table, while intertemporal comparison leans more on lateral prefrontal and posterior parietal regions. Flagged honestly: their "two systems" interpretation was challenged by Kable & Glimcher (2007, Nature Neuroscience), who showed a single valuation system with subjective discounting explains the same data. The phenomenon is rock solid; that particular mechanism story is contested. Cross-ref psychology / neuroscience.
• Weaker · complexity costs rise faster than revenue: Gottfredson & Aspinall (2005, HBR) argue that product and process complexity carries non-linearly rising cost, with an "innovation limit" past which it destroys value. Management analysis plus consulting cases, not peer-reviewed — treat as directional, not law. But everyone who has worked in delivery recognizes the curve.

Edges: this is not "never customize." Some custom work is a genuine moat, and some customers are big enough to be worth changing for. The rule is that customization has to be run as a piece of business — priced separately, scheduled separately, written into the contract separately — rather than tossed in as a sweetener to get a signature. Custom work given away free is you spending other people's time to buy your own quota.

Two independent mechanisms are billing you at once. First, mediocre or irrelevant information dilutes the force of diagnostic information. Second, the harder the conflict between options, the more people defer — and deferral doesn't go to a competitor, it goes to doing nothing.

• Hard evidence · the dilution effect: Nisbett, Zukier & Lemley (1981, Cognitive Psychology) — adding neutral, non-diagnostic information alongside genuinely diagnostic information pulls judgments significantly back toward the middle; the strong signal gets watered down. Same direction: Anderson (1965, JPSP) found that adding "moderately positive" traits to a description lowers the overall evaluation. The brain averages; it doesn't add. Peer-reviewed, well replicated. Hard.
• Hard evidence · conflict produces deferral and inaction: Tversky & Shafir (1992, Psychological Science) showed that when two options each have a clear advantage and the trade-off is hard, people are markedly more likely to defer the decision or choose neither — even when both beat the status quo. That is exactly what three parallel proposals produce. Classic, repeatedly replicated. Hard.
• A famous study to discount · the jam experiment: Iyengar & Lepper (2000, JPSP) ran the supermarket tasting study (24 jams drew more passers-by, but the 6-jam table converted far better) and it has been cited for two decades — but it must be flagged. The meta-analysis by Scheibehenne, Greifeneder & Todd (2010, JCR) found the average choice-overload effect is close to zero; Chernev, Böckenholt & Goodman (2015, JCP) narrowed it further: the effect is reliable only under specific conditions — complex options, unclear preferences, time pressure. A B2B proposal happens to hit all three, so the direction holds here; just don't recite "24 versus 6" as a universal law.

Edges: subtraction is not hiding information. The detail still has to exist — it belongs in the appendix and in the answer when they ask, not stacked across the decision page. One counter-case: when the buyer's core worry is literally "can you cover all our scenarios," breadth is the hard evidence. Then list it — on a capabilities annex, not mixed into the recommendation page.

A signal is only credible when it is more expensive for an impostor. Zero-cost adjectives are available to everyone, so they transmit nothing. And when the listener is motivated to process carefully — which spending money and carrying accountability guarantees — empty rhetoric isn't merely useless, it files you under "marketing copy."

• Hard evidence · a signal needs differential cost: Spence (1973, Quarterly Journal of Economics) — a signal separates high from low quality only if it costs the low-quality party more. That work won a Nobel in economics, and it is the formal version of "anything anyone can say for free carries no information." It also explains why self-limiting statements like "we don't do X" raise credibility: a bluffer won't pay that cost.
• Hard evidence · peripheral cues stop working under motivation: Petty & Cacioppo (1986), the Elaboration Likelihood Model — when an audience is both motivated and able to process carefully they take the "central route," where only concrete, checkable arguments shift attitudes, and those shifts last longer; polished slogans are peripheral cues that only work on people who aren't paying attention. B2B purchasing is almost always high-motivation. Peer-reviewed. Hard. Cross-ref psychology.
• Hard evidence · concrete material is processed faster and remembered better: the dual-coding work beginning with Paivio, Yuille & Madigan (1968, JEP Monograph) repeatedly shows concrete words are processed faster and recalled more reliably than abstract ones. The sentence your champion has to repeat in a meeting must be concrete — an abstract slogan evaporates in their mouth (how to arm a champion is covered in the value selling piece).
• Industry data, discount it · buyers already did the research: Adamson, Dixon & Toman (2012, HBR), citing CEB survey work, report that B2B buyers are on average about 57% through their purchase process before contacting a vendor. Consulting-firm survey, definitions shift by year, not peer-reviewed — order-of-magnitude reference only. But the direction is solid: the generic part of your introduction, they read before you opened your mouth.

Edges: concrete doesn't mean burying someone in technical specs. "99.99% SLA, active-active distributed architecture" is hard currency to a CTO and still noise to a finance lead — concreteness is relative to the listener, so get concrete about their ledger, their process, their day. And an exclusion only works if you genuinely made that trade-off; faking "we don't do that" to sound credible falls apart on the first follow-up question.

Sales technique isn't good or bad in the absolute; it's matched to the situation. The same moves raise win rates on complex deals and lower them on simple ones, because the return on diagnosis comes from finding what the buyer doesn't know — and when they know it all, the return is zero while the cost still bills.

• Large-sample industry research · technique effectiveness flips with deal size: Rackham's SPIN Selling (1988), built on field observation of roughly 35,000 sales calls, found that closed questions and classic closing techniques help in small transactions and actively reduce success in large complex ones. Discount it — not peer-reviewed, underlying method not fully published — but the sample is enormous and the direction has been confirmed repeatedly in practice. (How to use the question funnel itself is in the piece on asking better questions.)
• Hard evidence · working memory has a hard ceiling: Sweller (1988, Cognitive Science), cognitive load theory — working memory is capacity-limited, and load unrelated to the goal (extraneous load) crowds out the resources available for the actual decision. Forty minutes of discovery bolted onto a simple purchase is exactly that kind of extraneous load. Peer-reviewed, widely replicated. Hard.
• Hard evidence · less information sometimes judges better: Gigerenzer & Goldstein (1996, Psychological Review), the "less-is-more" effect — in particular environments, people with fewer cues judge more accurately, because irrelevant cues dilute the valid ones (same substrate as the previous section). Peer-reviewed, hard, but bounded: it is not a licence for "less information is always better."
• Mechanism · the discovery itself is meta-information: ask ten questions of someone who just wants to buy a screwdriver and what they learn isn't "these people are thorough," it's "these people are a hassle" — your process complexity gets read as the complexity of working with you.

Edges: don't treat this as permission to coast. The genuinely dangerous box is "very clear, and clearly wrong" — they open by asking for A, and A won't solve their problem. The test: can they state what they want and also why they want it? If they can answer "why," quote. If they can only answer "what," it's worth one probing question — this is where the Challenger material earns its keep, but as a question, not a meeting.