{"id":2464,"slug":"owner-launch-wait-and-time-to-meaning","title":"Owner launch-wait psychology and time-to-meaning","kind":"reference","scope":"business","status":"current","audiences":["kevin","smb-owner","candid-team","client-prospect"],"topics":["owner-launch-wait-psychology"],"reference_body":"# Owner launch-wait psychology and time-to-meaning\n\n**Owner launch-wait psychology** is the body of cognitive and emotional experience that a small-business owner undergoes during the weeks and months between the launch of a new website and the moment that site begins to produce measurable commercial results. The interval is sometimes labelled the **time-to-meaning** of a new web property — the lag between technical go-live and the first reading at which performance data can be interpreted as evidence of success or failure. The phenomenon is significant because the normal duration of that lag, as established in the search-ramp and trust-accrual literatures, is substantially longer than the duration that most owners psychologically tolerate; the result is a predictable pattern in which a structurally healthy launch is misread as a failed one, and corrective action is taken on the basis of an emotional disconfirmation rather than on the basis of accumulated data.\n\nThe topic sits at the intersection of judgment-and-decision-making research (anchoring, expectation-disconfirmation, action bias, sunk cost, loss aversion, hyperbolic discounting, availability, illusion of control) and the operational reality of search-engine ramp curves and trust accrual. It is the owner-side companion to the technical brief [[seo-j-curve-and-new-site-ramp]] and the trust-side brief [[new-site-trust-accrual-and-first-impressions]], and it draws on the broader cognitive substrate documented in [[psychology-of-marketing-aversion]] and [[behavioural-economics-for-smb-marketing]].\n\n## Overview\n\nA new website launches on a Tuesday. By the following Tuesday, organic traffic remains negligible, the inbox is quiet, the phone has not rung, and the owner — having paid for design, copywriting, photography, development, and the disruption of running a business through a rebuild — begins to feel that the project has failed. The feeling is acute, it is not unreasonable on its face, and it is, in the great majority of cases, an artefact of a small set of well-documented cognitive mechanisms acting in combination on a normal ramp curve.\n\nThe defining feature of the launch-wait period is that the owner's **subjective experience of progress** runs ahead of the **observable evidence of progress** by a factor that depends on the slope of the ramp and the depth of the trust-accrual interval. During this gap, the owner is asked to behave as if the site is working while having no proximal evidence that it is. The literature identifies eight cognitive mechanisms that, taken together, predict that owners in this period will experience a normal ramp as failure and will be tempted to act on that feeling.\n\nThe most common manifestation is **mid-window tinkering** — changing the headline, swapping the hero image, rewriting the home page, adding a new section, demanding a colour change. Each change feels like a response to data. It is almost never actually a response to data, because the traffic levels reached in the first weeks of a new SMB site are below the threshold at which any small change could be statistically detected even if it had occurred. The behaviour is **anxiety management dressed up as decision-making**, and the discipline that resists it is the pre-committed measurement window.\n\n## The launch-and-wait experience\n\nThe launch-and-wait experience is the lived period between go-live and the first defensible reading of commercial impact. Its characteristic phases are:\n\n**The taut interval (week 1).** The site is new, the owner has notified the network, traffic spikes briefly from announcement effects, the inbox shows the first reactions of friends and existing customers. The owner reads these as early signal. They are not — they are launch-day attention from people already in the orbit.\n\n**The drop (weeks 2–3).** Announcement traffic decays. The site's actual organic baseline becomes visible for the first time, and it is small. The contrast against the launch-week spike registers as decline. The owner begins to ask \"is something wrong?\"\n\n**The trough (weeks 3–8).** Traffic is at or near the long-run baseline of a brand-new domain with no accrued trust signals — which, on the search-engine side, is essentially zero qualified organic traffic. The site exists but no one is finding it through search. This phase is **mechanically inevitable** under the conditions documented in the J-curve brief and the trust-accrual brief; it is not evidence of anything wrong with the site. It is, however, the phase in which the owner's emotional model of the project most strongly diverges from the empirical model.\n\n**The slow build (weeks 8–24).** Search trust begins to accrue. Branded queries arrive. The first non-branded organic visits appear. Engagement metrics on those early visits start to be readable. The signal becomes interpretable, in the technical sense of having enough sample to support an inference.\n\n**The first defensible reading (month 6 onward).** The site has been live long enough that a comparison against pre-launch baselines, against industry base rates, and against the original measurement plan can be made without small-sample noise dominating the conclusion. This is the earliest moment at which \"is this site working?\" is a question the data can answer.\n\nThe mismatch between the four-to-six-month time-to-meaning and the two-to-three-week patience window of the typical owner is the single most-discussed dynamic in the launch-wait literature. The mismatch is structural, not pathological — it reflects a real misalignment between the speed of human emotion and the speed of search-engine and trust signals.\n\n## The trough of meaninglessness\n\nThe phase between weeks 2 and 8 is sometimes referred to in informal language as the **trough of meaninglessness**: a period in which the site is live, the data is accruing, and yet there is no readable signal to act on. The trough is the period in which the gap between the anchored expectation set at sale (\"you'll see results fast\") and the actual ramp curve registers most painfully.\n\nThe trough has three distinguishing properties:\n\n**It is mechanically produced.** The trough is not a function of how the site was built, marketed, launched, or designed. It is a function of how search engines and trust networks evaluate new domains. A perfectly built site and a poorly built site experience the same trough, because the trough is in the evaluator, not the product. (See the sister briefs [[seo-j-curve-and-new-site-ramp]] and [[new-site-trust-accrual-and-first-impressions]] for the underlying mechanics.)\n\n**Its data is not yet interpretable.** Inside the trough, the volume of traffic is too low to read any change as signal. This is not a matter of needing better analytics — it is a matter of statistical power. Standard conversion-rate-optimisation work at well-resourced organisations finds that only roughly 10–20% of properly powered A/B tests yield a positive result; below that, sample sizes are too small for the question to even be asked. A new SMB site in week 3 is not within several orders of magnitude of the sample size required to read a small change. The data is genuinely uninterpretable, not merely inconvenient.\n\n**It feels exactly like failure.** This is the phenomenological centre of the problem. The owner's interior experience during the trough is indistinguishable from the experience of actually having a failed launch. The same flat dashboards, the same quiet inbox, the same silent phone. The brain has no innate mechanism for distinguishing \"nothing visible because the evaluator hasn't read the signal yet\" from \"nothing visible because there is nothing to read.\" Both look like nothing.\n\nThe combined effect — mechanically produced, statistically opaque, emotionally indistinguishable from failure — explains why the trough is the period in which the most damage to a healthy launch is typically done. The owner, in the absence of countervailing discipline, begins to rebuild a site that was already working as designed.\n\n## Why human pattern-recognition fails on slow signals\n\nThe cognitive substrate that makes the launch-wait so difficult is not unique to small-business owners; it is general to human judgment under slow-feedback conditions. Eight mechanisms have been documented in the judgment-and-decision-making literature that, taken together, predict the launch-wait failure mode. They were inventoried in the original mechanism mapping for this topic.\n\n**Source:** Compass artifact research document, June 2026, synthesising the underlying primary literature. **Confidence:** High on enumeration; per-mechanism confidence varies. **Caveat:** Two of the eight (illusion of control and the magnitude of loss aversion) are replication-flagged and should be cited as directional rather than as quantitatively settled.\n\n### Anchoring\n\nThe first quantitative claim a person hears about how quickly a website will produce results becomes the reference point against which all subsequent observations are judged, and subsequent judgments insufficiently adjust away from that anchor even when its origin is known to be unreliable. The original demonstration, Tversky and Kahneman's 1974 \"wheel of fortune\" study, found that participants asked to estimate the percentage of African countries in the UN were measurably influenced by an arbitrary number on a spinning wheel that they had watched land in plain sight — a vivid demonstration that the anchor operates even when its irrelevance is transparent.\n\nFor new-site owners, the anchor is almost always a vendor's pre-sale promise — some variant of \"you'll see results fast\" — which sets the reference point against which the slow ramp will be evaluated. The more inflated the original promise, the wider the gap with the realistic ramp, and the harsher the felt failure.\n\n**Source:** Tversky, A., & Kahneman, D. (1974). \"Judgment Under Uncertainty: Heuristics and Biases.\" *Science* 185(4157):1124–1131. **Confidence:** Verified. One of the most robust effects in judgment research; replicated thousands of times in many contexts. **Caveat:** Magnitude varies by context; anchoring is a direction, not a fixed quantity.\n\n### Expectation-disconfirmation\n\nSatisfaction is governed not by absolute performance but by the gap between prior expectation and perceived performance. When a high, vendor-inflated expectation meets a slow normal ramp, the resulting negative disconfirmation registers subjectively as dissatisfaction and is interpreted as failure — even when the underlying performance is on its expected trajectory. A 2024 *Journal of the Academy of Marketing Science* meta-analysis of 150 records and N=58,597 confirmed a positive expectation-satisfaction relationship of approximately r=.29 and found no evidence supporting contrast effects.\n\nThe mechanism explains why over-promising at sale is so corrosive: the larger the original anchor, the larger the disconfirmation when normal ramp behaviour is observed, and the more painful the felt failure independent of any underlying problem.\n\n**Source:** Oliver, R. L. (1977). \"Effect of expectation and disconfirmation on postexposure product evaluations.\" *Journal of Applied Psychology* 62(4):480–486; Oliver (1980). *Journal of Marketing Research* 17(4):460–469; meta-analysis in *Journal of the Academy of Marketing Science* (2024). **Confidence:** Industry-consensus/Verified. Dominant paradigm for 40+ years; well-supported by recent meta-analysis. **Caveat:** Direction of effect depends on assimilation vs. contrast processes; the magnitude of the felt gap depends on how inflated the original anchor was.\n\n### Action bias\n\nAction bias is the preference for doing something over waiting, even when waiting is the utility-maximising choice. It is driven by an asymmetric regret response — inaction following a bad outcome feels worse than action following the same outcome — and it is the engine that turns the trough's \"things feel bad\" into \"I must change the home page now.\"\n\nThe canonical real-world demonstration is Bar-Eli et al.'s analysis of 286 professional soccer penalty kicks: goalkeepers jumped to one side or the other in 93.7% of kicks, despite the empirical finding that goalkeepers saved 33.3% of penalties when staying central, compared with 12.6% jumping right and 14.2% jumping left. Centre was the utility-maximising strategy by a wide margin, and the elite professionals studied chose it less than 7% of the time. The result is the canonical real-world finding that expert decision-makers in high-stakes settings prefer doing something to holding position even when holding position is empirically better.\n\n**Source:** Bar-Eli, M., Azar, O. H., Ritov, I., Keidar-Levin, Y., & Schein, G. (2007). \"Action bias among elite soccer goalkeepers: The case of penalty kicks.\" *Journal of Economic Psychology* 28(5):606–621. **Confidence:** Verified (named-author peer-reviewed paper; n=286 high-stakes real-world observations). **Caveat:** \"Centre is best\" is a specific finding of this dataset, not a universal action-bias coefficient.\n\nThe underlying preference structure was modelled by Patt and Zeckhauser. **Source:** Patt, A., & Zeckhauser, R. (2000). \"Action Bias and Environmental Decisions.\" *Journal of Risk and Uncertainty* 21(1):45–72. **Confidence:** Verified within its domains. **Caveat:** Context-dependent. The opposite bias — omission bias, the preference for inaction — dominates in other settings such as vaccine decisions. Action bias is a tendency, not a universal law.\n\n### Illusion of control\n\nIllusion of control is the tendency to overestimate one's ability to influence outcomes that are substantially determined by factors outside one's control. For a new-site owner, changing the headline, the hero image, or the button colour feels like influencing the outcome even when the actual outcome — search-engine indexing, trust accrual, ranking — is governed by factors the owner cannot move quickly. The mechanism is what makes tinkering feel productive rather than merely soothing.\n\n**Source:** Langer, E. J. (1975). \"The Illusion of Control.\" *Journal of Personality and Social Psychology* 32(2):311–328. **Confidence:** Replication-flagged. The general direction (people overestimate their influence on chance-determined outcomes) is Industry-consensus. The specific Langer Study 2 effect has repeatedly failed to replicate — Kühberger, Perner, Schulte and Leingruber (1995) ran four experiments and none reproduced the original finding. Treat Langer's effect size as Directional, not Verified. **Caveat:** Cite illusion of control as a real cognitive tendency whose direction is robust but whose magnitude in any given context is unestablished.\n\n### Sunk cost\n\nThe sunk-cost effect is the tendency to continue an endeavour because of prior irrecoverable investment of money, time, or effort, even when forward-looking analysis says to stop. The canonical demonstration is the Arkes and Blumer theatre-season-ticket field study, in which subscribers who paid full price for season tickets attended significantly more plays than those who received random discount cards — even though the marginal cost of each play attended was identical for both groups. The decision to attend was governed by the prior irrecoverable payment, not by the forward-looking value of the play.\n\nFor new-site owners, the sunk-cost effect operates in the direction of over-investment: the more the owner has already spent on the new site, the new brand, the new content, the harder it becomes to walk away, and the more vulnerable the owner is to pouring additional money into a site whose problems are structural.\n\n**Source:** Arkes, H. R., & Blumer, C. (1985). \"The Psychology of Sunk Cost.\" *Organizational Behavior and Human Decision Processes* 35(1):124–140. **Confidence:** Verified for monetary sunk costs in lab and field. Single-source/Directional for purely behavioural (time/effort) investments; Soman (2001) and some failed replications find the effect weaker or inconsistent there. **Caveat:** The corrective discipline is the same as for the opposing error of premature abandonment — judge on evidence accumulated over a pre-set window.\n\n### Loss aversion\n\nLoss aversion is the finding that losses loom larger than equivalent gains in subjective experience. For a new-site owner, a slow start is framed as a loss relative to the anchored vendor expectation, amplifying the urge to abandon or over-correct. The mechanism is robust as a direction; its widely-cited magnitude is not.\n\n**Source:** Kahneman, D., & Tversky, A. (1979). \"Prospect Theory: An Analysis of Decision under Risk.\" *Econometrica* 47(2):263–291. **Confidence:** Industry-consensus that loss aversion exists as a direction. Directional on magnitude. The widely-cited \"losses loom roughly 2× larger than gains\" coefficient is no longer accepted as a universal value: Gal and Rucker (2018) challenged the universal 2:1 claim and argued the field had over-extended the original finding; meta-analytic re-examinations by Walasek, Mullett and Stewart find the coefficient highly moderated and far lower (approximately 1.3), or near 1 under symmetric, unordered conditions; Kahneman himself conceded the effect is \"not a law of human nature that you have to find it in every context.\" **Caveat:** Cite the direction confidently; do not cite the 2:1 coefficient as settled. How loss aversion scales to the slow-launch framing specifically is unmodelled — a known gap in the literature.\n\n### Hyperbolic discounting\n\nHyperbolic discounting, or present bias, is the finding that people steeply discount future rewards relative to immediate ones, producing dynamically inconsistent impatience: an owner wants results \"now\" and is tempted to abandon before a slow payoff arrives. Search-engine returns and brand-trust returns are inherently slow on the time horizons documented in the sister briefs; present bias is the mechanism that creates the urge to quit before the slow payoff has had time to materialise.\n\n**Source:** Ainslie, G. (1970s) original psychophysical work; Laibson, D. (1997). \"Golden Eggs and Hyperbolic Discounting.\" *Quarterly Journal of Economics* 112(2):443–477. **Confidence:** Industry-consensus that present bias exists (strong evidence including demand for commitment devices). Directional on exact functional form. **Caveat:** The specific quasi-hyperbolic functional form (the β-δ \"beta-delta\" model) is contested; some studies favour a small fixed-cost present bias instead. The direction is robust; the precise math is not.\n\n### Availability\n\nThe availability heuristic is the tendency to judge probability by how easily examples come to mind. An owner who has previously launched a website that failed will weight that personal experience far above the base-rate distribution of new-site outcomes — and walk away from a normal ramp because the recall is vivid. The corrective is base-rate framing: presenting the actual distribution of outcomes for sites at the relevant week of a launch, which works against the vivid single recall by providing a different anchor.\n\n**Source:** Tversky, A., & Kahneman, D. (1973). \"Availability: A heuristic for judging frequency and probability.\" *Cognitive Psychology* 5(2):207–232. **Confidence:** Verified. Well-established direction; magnitude situational. **Caveat:** The corrective is base-rate framing, which gives the owner a different anchor against which to evaluate the present situation.\n\n### Combination\n\nThe eight mechanisms do not operate in isolation; they compound. An anchored expectation creates the gap; disconfirmation makes the gap feel bad; loss aversion amplifies the bad feeling; action bias makes inaction feel worse than action; illusion of control says the action will help; sunk cost says quitting wastes everything invested; hyperbolic discounting says the payoff is too far away; availability brings prior bad experiences forward as evidence that the present situation will end the same way. The owner is, in effect, simultaneously primed to over-invest in changes (sunk cost, action bias, illusion of control) and to abandon entirely (hyperbolic discounting, loss aversion, availability) — sometimes within the same week. Owners who recognise the eight-mechanism stack in themselves typically de-escalate their own panic without further intervention.\n\n## Sunk-cost framing and rational persistence\n\nA common informal claim during the launch-wait is that an owner who keeps changing the site is \"responding rationally to data.\" The literature does not support this framing. The verdict, on review, is **usually false**.\n\nEarly in the wait, too little traffic has accumulated to read any signal. Small and new sites can typically only detect large effects in their analytics, and even those require sample sizes most SMB launches never reach within their first months. The urge to change things is action bias and illusion of control reducing anxiety, not rational data-driven inference. The mechanism is verified; the case-by-case application is directional — a genuine large structural problem, such as a broken form submit, can be diagnosed without large samples, but the typical week-3 dissatisfaction does not arise from such a problem.\n\n**Source:** Synthesis of the action-bias literature, the illusion-of-control literature, and standard A/B-test power calculations. **Confidence:** Verified on the mechanism. Directional case-by-case. **Caveat:** The corrective is not \"never change anything\" but \"change in response to structural defects (broken forms, missing pages, obvious usability defects) — not in response to vibes about ranking trajectory.\"\n\nThe framing the literature does support is **rational persistence**: a decision to continue is not the same as a refusal to look at evidence. Rational persistence is a commitment to read the evidence at the pre-committed window and to act on what it shows, neither sooner nor later. It is distinguishable from sunk-cost escalation in that it is forward-looking — the decision to continue is justified by the expected future payoff over the planned window, not by the prior expenditure — and it is distinguishable from blind faith in that it has a stop condition (the window's end).\n\n## The empirical milestones that mark genuine progress\n\nIn place of week-by-week vibes, the launch-wait literature points to a small set of empirical milestones that mark genuine progress and that can be read with reasonable confidence as the ramp accrues.\n\n**Indexation and structural integrity.** Within the first weeks, the site should be fully indexed by major search engines, core pages should be reachable and crawlable, structured data should validate, and forms should submit. These are binary, low-traffic-tolerant signals — they can be read at any traffic level — and they are the **only** signals that should drive action inside the trough. A structural defect identified here justifies a fix; a slow ramp does not.\n\n**First branded queries.** The first non-trivial volume of branded queries (people searching for the business name) indicates that the brand is reaching its existing network and beginning to leak into search behaviour. The milestone typically appears within the first weeks but its strength scales with the launch communications around the new site.\n\n**First non-branded organic visits.** Non-branded organic visits — people finding the site via category, problem, or location queries rather than by name — are the first signal that the search engines have begun to assign the site to relevant query intents. The milestone typically appears in the weeks-to-months window covered by the sister brief [[new-site-trust-accrual-and-first-impressions]].\n\n**First conversion event.** A form submission, a phone call, an email reply, a booking, or whatever the site's primary commercial event is. Within reasonable variance, the first such event arrives well inside the first six months for a properly built SMB site, and its arrival is the first event that can be discussed in commercial rather than technical terms.\n\n**Sample sufficient to detect change.** This is the milestone at which the trough ends. It is the first moment at which any subsequent change to the site could be tested with enough sample to produce a defensible inference. For most SMB sites it is months, not weeks, from launch.\n\nThe list is significant because it gives the owner a vocabulary other than \"results\" for describing what is happening. \"Indexation complete\" is a result. \"First branded query\" is a result. \"First non-branded organic visit\" is a result. They do not pay the bills directly, but they are evidence that the underlying system is functioning, and they are interpretable inside the trough where the commercial outcomes are not.\n\n## Patience vs pivot — when each is justified\n\nThe discipline that distinguishes warranted patience from warranted pivot has been formalised in a set of rules drawn from the launch-wait literature.\n\n**Pre-commit to a measurement window before launch.** The owner decides, in advance, how long the wait will be, what sample size will constitute a readable signal, and what specific structural defects would override the window. The pre-commitment is made at the only moment the owner is calm enough to think clearly — before the disconfirmation and the loss feeling hit. At low traffic, the owner literally cannot detect anything but very large changes, and standard CRO base rates from elite shops put the positive-test rate at only 10–20% even at scale. A defensible window for a typical new SMB site is **eight to twelve weeks minimum** before reading anything other than structural-defect signals.\n\n**Resist mid-window tinkering.** Constant changes destroy the ability to attribute cause and reset the clock — every change makes the next measurement window start over. The urge to change things is anxiety management, not inference. A genuinely failing site shows structural problems, not merely a slow ramp. The benchmark to change course is **structural defect identified**, not \"I feel bad about how slow this is.\" If the answer to the question \"what specifically is broken?\" is \"nothing, it's just slow,\" the rule is to change nothing.\n\n**Separate \"feels like failure\" from \"is failing.\"** When the new site feels as if it is failing in weeks two through eight, the feeling is to be treated as an expectation-disconfirmation artefact first. The check is against base rates — what does a normal week-3 actually look like for a comparable SMB site? — and not against the vendor's original promise, which is the inflated anchor. The operational form of the rule is the question pair: \"compared to what?\" and \"what would week-3 look like if everything were going exactly to plan?\" The answer to the second question is often: exactly what the owner is seeing.\n\n**Watch both sunk-cost escalation and hyperbolic-discounting abandonment.** The two errors point in opposite directions but require the same corrective. Sunk-cost escalation pours additional money into a structurally broken site because of prior investment; hyperbolic-discounting abandonment quits before a slow payoff arrives because the future feels too far away. The discipline against both is to judge on evidence accumulated over the pre-set window, not on how the wait feels in any given week. Both errors are anxiety management dressed up as decision-making; the pre-committed window is the only discipline that works against both simultaneously.\n\nThe composite rule is straightforward: outside the window, act only on structural defects; at the end of the window, re-evaluate against base rates and against the original plan; inside the window, do not re-evaluate other than for structural defects. The window is the contract the owner makes with themselves at the calm moment before the launch; honouring it is the operational form of rational persistence.\n\n## The role of weekly read-outs\n\nThe literature on launch-wait psychology converges on the practical conclusion that a structured weekly read-out, delivered to the owner during the trough, is one of the more effective interventions against the eight-mechanism stack. Its function is not to deliver new commercial data — there is none to deliver — but to absorb the owner's anxiety into a defined cadence and to provide vocabulary other than \"results\" for the period in which results are not yet readable.\n\nA typical read-out reports: indexation status, structural-integrity checks, any branded-query volume, any non-branded query impressions, any structural defects identified and fixed in the prior week, and the position of the launch within the pre-committed measurement window. The read-out's purpose is to translate the eight-mechanism experience — anchored expectations, felt disconfirmation, action-bias urge, illusion-of-control temptation — into a structured artefact that confirms what is and is not happening and what is and is not yet readable.\n\nThe mechanism by which the read-out works is the substitution of a small reliable signal (the weekly artefact) for the absent commercial signal. Inside the trough, the owner's nervous system is reading for a signal that is structurally unavailable. The read-out provides a different signal — one that is reliably available, that is bounded in scope, and that confirms the pre-committed plan is being executed. It is, in the language of the action-bias literature, a way of doing something — preparing and discussing the read-out — that is the utility-maximising form of \"doing something\" during the wait.\n\n## Compass research findings\n\nThe June 2026 Compass synthesis on which this page draws made several findings that bear restatement.\n\nFirst, the eight-mechanism inventory is the most useful single deliverable in the underlying brief for client conversations. Owners who recognise the stack in themselves typically de-escalate their own panic without further intervention. The act of naming the mechanisms appears to interrupt their compounding, presumably because the felt experience of \"this is panic\" is qualitatively different from the felt experience of \"this is reality.\"\n\nSecond, two of the eight mechanisms are replication-flagged: the magnitude of loss aversion is no longer accepted at the canonical 2:1 ratio (with meta-analytic re-examination placing it closer to 1.3, and near 1 under symmetric conditions), and the specific Langer 1975 illusion-of-control effect failed to replicate in Kühberger et al. (1995). The direction of each effect remains industry-consensus; the magnitude does not. Practitioners working with these mechanisms in client conversation are advised to cite the direction confidently and not to cite specific coefficients.\n\nThird, the contested claim that \"an owner who keeps changing the site is responding rationally to data\" was adjudicated and found usually false. The verdict relies on a synthesis of action-bias literature, illusion-of-control literature, and standard statistical-power calculations, and is verified on the mechanism while remaining directional on case-by-case application.\n\nFourth, the rules-against-tinkering apparatus (pre-commit measurement window, resist mid-window tinkering, separate feels-like-failure from is-failing, watch both sunk-cost and hyperbolic-discounting) is internally consistent and converges on a single discipline: the pre-committed window, honoured against both opposing errors, with structural defects as the only override. The convergence is the brief's central operational finding.\n\nFifth, a known unmodelled gap in the literature is how loss-aversion magnitude scales to the slow-launch framing specifically. The mechanism is robust as a direction; how it sizes under the launch-wait conditions is unestablished. The gap is acknowledged rather than papered over.\n\nThe Compass synthesis sits alongside the sister briefs on the J-curve and on trust accrual ([[seo-j-curve-and-new-site-ramp]], [[new-site-trust-accrual-and-first-impressions]]) and complements the broader cognitive substrate documented in [[psychology-of-marketing-aversion]] and [[behavioural-economics-for-smb-marketing]]. Together, the four pages frame the owner-side experience of a new website launch as a structured phenomenon with documented mechanisms, documented corrective rules, and documented limits — distinct from the technical question of whether the site is working and equally consequential to the project's outcome.\n","rationale_body":null,"metadata":null,"links":{"outgoing":[],"incoming":[]},"created_at":"2026-06-25T18:44:11.854Z","updated_at":"2026-06-25T18:44:11.854Z","resolved_via_alias":"loss-aversion-kahneman-tversky-1979","resolved_via_alias_anchor":"why-human-pattern-recognition-fails-on-slow-signals"}