The NOVI 8-Decision Growth Model

An academic article situating the NOVI 8D methodology within the contemporary behavioural-science literature, with attention to phase architecture, mechanism, measurement, and methodological limits.

Abstract

This paper presents the theoretical foundations, mechanism-level rationale, and methodological caveats of the NOVI 8-Decision Growth Model - a phased decision-mapping methodology applied to conversion architecture in commercial, charitable, and governmental contexts. The framework treats every buyer's progression toward action and beyond as the successful resolution of eight sequential decisions organised within three operational phases: a Capture phase comprising belonging recognition (D1), problem recognition (D2), and causal reframe (D3); a Convince phase comprising proof (D4), source credibility (D5), and proportionate ask (D6); and a Close phase comprising timing (D7) and identity-and-social risk (D8). Each decision is paired with a specific functional execution layer that translates the psychological diagnosis into deployable intervention - from core-offer selection through audience and channel architecture, creative testing, landing-page optimisation, trust-marker placement, friction reduction, post-purchase nurture, and lifetime-value extension. The paper argues that the framework's contribution is not the proposition that any single decision matters - each is well-established in the underlying psychological literature - but the operationalisation of the eight as a phased sequence whose failure mode can be diagnosed empirically, remediated structurally, and extended beyond the conversion event to encompass retention and advocacy. The framework is grounded in the dual-process tradition (Kahneman, 2011; Petty and Cacioppo, 1986), the influence and persuasion literature (Cialdini, 2007, 2016), the choice-architecture programme (Thaler and Sunstein, 2008), the habit-formation work of Eyal (2014) and Wood and Rünger (2016), and the implementation-science synthesis (Michie, van Stralen and West, 2011). Critical attention is given to the calibration of expected effect sizes against the publication-bias-corrected evidence base (Maier et al., 2022; DellaVigna and Linos, 2022) and to the conditions under which phased decision mapping is empirically warranted versus where it operates as a heuristic structuring device. The paper concludes that the framework's practical value lies in its diagnostic discipline - the requirement to identify which decision is failing, in which phase, before any intervention is designed - rather than in any claim that the eight decisions exhaust the determinants of buyer action.

Keywords: decision mapping, conversion architecture, applied behavioural science, dual-process theory, sequential persuasion, customer lifetime value, methodological reflexivity

1. Introduction

The dominant paradigm in commercial conversion optimisation, as developed across the past two decades of digital marketing practice, treats the buyer-to-action sequence as a continuous funnel whose efficiency can be improved through micro-level adjustments to creative, copy, channel, and offer. The funnel metaphor has been operationally useful - it generates measurable conversion rates and lends itself to A/B testing - but its theoretical foundation is comparatively thin. Funnels are described; they are not, in any meaningful sense, explained. Equally significantly, the conventional funnel terminates at the conversion event itself, treating what happens after as a separate concern handled by a separate function (retention marketing, customer success, lifecycle email). This separation obscures the empirical reality that the decisions a buyer makes after conversion materially affect the lifetime value the firm extracts from the relationship.

The NOVI 8-Decision Growth Model (hereafter, the 8D model) makes a different theoretical commitment. It treats the buyer's progression not as a continuous funnel but as a discrete sequence of eight decisions, organised into three operational phases - Capture, Convince, and Close - that extend across both the pre-conversion sequence and the post-conversion lifecycle. Each decision is paired with a specific execution layer that translates the psychological diagnosis into deployable practice. The framework's central operational claim is that conversion failure (and, equally, lifetime-value failure) can be diagnosed at the decision level, that the specific decision blocking a given engagement can be empirically identified, and that interventions designed against the identified blocker will outperform interventions designed against the apparent execution-level symptom.

This paper presents the theoretical foundations of the framework, its mechanism-level rationale, and its methodological limits. Section 2 sets out the phase architecture and the eight decisions within it, with attention to the functional execution layer associated with each. Section 3 examines the framework's central structural claims - phased sequentiality, decomposition, and diagnostic identifiability. Section 4 addresses the framework against the contemporary methodological-critical literature, with explicit attention to publication-bias-corrected effect-size expectations. Section 5 considers the framework's limits and the conditions under which it is and is not warranted. Section 6 concludes.

A note on register. The framework was developed in commercial practice and is operationally orientated. The present paper translates it into a form that engages the academic literature honestly - including, where the underlying empirical base is contested, acknowledging the contestation. The claim is not that the framework rests on settled science. It is that the framework operationalises a body of contested but substantively serious science in a way that produces operational discipline at the point of intervention design.

2. The Three Phases and Eight Decisions

The 8D model organises the eight decisions into three operational phases. The phases are not merely temporal markers but functional categories that map the psychological state of the buyer to a specific set of execution tasks. Table 1 summarises the architecture.

2.1 The Capture phase

The Capture phase governs the buyer's transition from non-awareness to qualified attention. It addresses the three decisions a prospect makes between first encountering a stimulus and acknowledging that the stimulus speaks to a real situation they recognise as their own.

D1 - The First 3 Seconds (Core Offer Selection). The first decision is whether the visitor recognises themselves in the page within the first several seconds of contact. The framework's operational test for D1 is whether the headline mirrors the buyer's situation before it states the solution, and the measurement proxies are ad click-through rate (with target ranges of 3-6% on Google Search and 1-3% on Meta) and time on hero section. The functional execution layer is the audit of market demand, margin viability, and value-proposition alignment - the upstream work that determines whether the offer being presented can plausibly be recognised as relevant by the target audience in the first place.

The psychological foundation for D1 is the categorisation literature in social cognition, particularly the work of Macrae and Bodenhausen (2000) on the speed and automaticity of self-relevance judgments. Self-relevance is computed rapidly and pre-attentively; visitors who do not recognise themselves in the surface signals of a page exit before any further information can register. The dual-process tradition (Kahneman, 2011) frames this as a System 1 judgment: fast, automatic, and resistant to subsequent System 2 correction. A page that fails D1 cannot be rescued by stronger arguments later because the visitor does not remain to receive them.

The post-2020 publication-bias literature has not seriously contested the existence of fast self-relevance judgments. The contested terrain is the magnitude of effect attributable to specific D1 interventions in real-world conditions. The DellaVigna-Linos (2022) in-field benchmark - approximately 1.4 percentage points on take-up outcomes - provides a more appropriate anchor for expected D1 intervention effects than the larger figures in the academic-journal literature.

D2 - Recognition (Audience and Channel Architecture). The second decision is whether the visitor recognises their specific situation in the problem statement on the page. Generic problem descriptions invite generic dismissal; specific problem descriptions create the recognition that opens the visitor to subsequent claims. The framework's measurement proxies are time on page and scroll past 30%, and the functional execution layer is the construction of audience and channel architecture: identifying which segments hold the buyers most likely to recognise themselves, mapping the semantic territory in which their recognition occurs, and configuring paid channels to reach those segments at the moments when recognition is most actionable.

The relevant literature is the construal-level theory of Trope and Liberman (2010) and the broader work on concrete versus abstract framing. Concrete, low-construal descriptions activate self-relevant retrieval; abstract descriptions do not. A visitor who recognises a specific failure mode they have personally experienced is reading a different document, in psychological terms, from a visitor reading the same words at a higher level of abstraction.

D2's empirical foundation is comparatively robust. The construal-level distinction has replicated reasonably well across the post-2015 methodological reform (Soderberg et al., 2015), and the specific finding that concrete problem descriptions outperform abstract ones in conversion settings has been confirmed in registered-report contexts.

D3 - The Reframe (Creative and Message Testing). The third decision is whether the visitor accepts the reframe - the explanation of the cause of their problem that differs from the explanation they brought to the page. The framework treats D3 as the most common blocking decision in NOVI engagements and the highest-leverage fix. The functional execution layer is structured creative and message testing: scripting and asset generation across multiple formats (user-generated content, static visuals, motion graphics), framework variation, and rapid-fire validation against the target audience.

D3 is theoretically the most demanding decision in the framework, because it requires the visitor to revise an existing belief. The relevant literatures are causal attribution theory (Kelley, 1973; Heider, 1958), the elaboration likelihood model (Petty and Cacioppo, 1986), and the more recent work on belief updating under conflicting evidence (Sharot, 2017). Belief revision is psychologically costly; visitors will adopt a new causal explanation only if it is presented as both novel and credible, and only if the source has cleared the credibility hurdles addressed elsewhere in the framework.

The post-2020 methodological literature has identified D3 as the area where applied behavioural science is most vulnerable to overclaiming. The priming literature on which much of Cialdini (2016) relies has shown systematic replication problems (Doyen et al., 2012; Open Science Collaboration, 2015), and the practitioner working in the D3 domain should be aware that the dramatic effect sizes reported in the popular-press literature on reframing and persuasion are unlikely to be reproducible at scale. The Maier et al. (2022) publication-bias correction is particularly relevant here.

This does not mean D3 is unimportant. It means the practitioner should treat D3 interventions as producing meaningful but moderate effects in the field, not the dramatic shifts implied by the published literature, and should design implementation expectations accordingly. The framework's emphasis on rapid-fire creative validation is, in this light, methodologically sound: structured iterative testing produces more reliable insight than the search for a single transformative creative.

2.2 The Convince phase

The Convince phase governs the buyer's transition from qualified attention to active intent. It addresses the three decisions that determine whether a buyer who has acknowledged a problem and accepted a reframe will commit to the proposed solution.

D4 - Proof (Landing Page Optimisation). The fourth decision is whether the visitor recognises proof that the proposed approach has worked for people like them. The framework's emphasis is on specificity: numbers rather than adjectives, contextually matched cases rather than generic testimonials. Measurement proxies are scroll to proof section (45% or beyond) and interaction with case studies. The functional execution layer is landing-page optimisation: UI/UX structuring, bespoke page construction, rapid load times, and the systematic removal of drop-off vectors that would prevent the proof from being encountered.

D4 is grounded in the social proof literature (Cialdini, 2007), the work on testimonial persuasion (Hoeken and Hustinx, 2009), and the broader research on contextual analogy in decision-making. The mechanism is straightforward: a buyer evaluating an unfamiliar offer estimates its likely effect for themselves by reference to its observed effect for others, weighted heavily by the perceived similarity between themselves and those others. Specific, context-matched proof produces stronger inference than generic proof of comparable magnitude.

D4 is one of the framework's more empirically robust components. The social-proof finding has been replicated in registered-report contexts including the Goldstein, Cialdini and Griskevicius (2008) hotel towel-reuse research, and the in-field nudge-unit literature has consistently identified normative-comparison interventions as producing reliable effects (Hallsworth et al., 2017).

D5 - Credibility (Social Proof and Authority Friction). The fifth decision is whether the visitor trusts the source making the claims. The framework identifies four elements that establish source credibility: the named human, the audited methodology, the affiliations, and the years. The functional execution layer is the strategic placement of credibility markers along the conversion path: case studies positioned at the moment of greatest scrutiny, video testimonials where written testimonials would feel insufficient, industry badges and certifications, and objection-handling FAQs that address the specific friction points the buyer is likely to raise.

D5 draws on the classical source-credibility literature (Hovland, Janis and Kelley, 1953; McGuire, 1985), updated by more recent work on the role of trust in digital persuasion (Metzger and Flanagin, 2013). The mechanism is well established and uncontested: information from sources perceived as credible is weighted more heavily than identical information from sources perceived as less credible. The contemporary applied question is which signals of credibility actually function in digital contexts, and here the literature is less settled.

The 2024 study by Salvi et al. on the conversational persuasiveness of large language models adds an interesting wrinkle to D5. Their finding that personalised AI-generated arguments outperform human-generated arguments in changing minds suggests that traditional human-credibility signals may be partially substitutable by algorithmic personalisation. The framework's emphasis on the named human as the principal D5 mechanism may require revision in domains where the buyer's primary interaction is mediated by AI systems. This is a research gap rather than a settled finding.

D6 - The CTA (Lead Capture and Direct Conversion). The sixth decision is whether the next step the page proposes feels sized correctly for where the visitor is in their sequence. The framework's central operational claim about D6 is that asks should be reframed as diagnostics rather than commitments - the visitor should be asked to receive something rather than to begin a process. The measurement proxy is landing-page-to-CTA click rate, with a target benchmark of 8 to 15 per cent. The functional execution layer is the systematic reduction of conversion friction: form-field minimisation, single-click checkout integration, conditional logic gating that hides irrelevant steps, and high-value lead magnets that make the proportionate ask still feel materially worthwhile.

D6 is grounded in the foot-in-the-door literature (Freedman and Fraser, 1966; Burger, 1999), prospect theory's analysis of loss aversion (Kahneman and Tversky, 1979), and the choice-architecture work on default-versus-active choice (Thaler and Sunstein, 2008; Johnson and Goldstein, 2003). The mechanism is the asymmetric weighting of immediate cost against deferred benefit - a high commitment now, even for a high-value outcome later, exceeds many buyers' willingness threshold, while a low-commitment diagnostic with the same eventual destination does not.

D6 is among the more empirically robust components of the framework. Default effects - the most closely related single mechanism - have replicated reliably across multiple natural-experiment contexts (Johnson and Goldstein, 2003) and have been confirmed in registered-report nudging studies. The general pattern that smaller asks produce higher conversion rates while preserving downstream sequence integrity is well documented (Cialdini, 2007).

2.3 The Close phase

The Close phase is where the 8D model departs most significantly from the conventional conversion-funnel paradigm. Where the funnel metaphor treats the conversion event as the terminal point of the sequence, the 8D model treats it as the midpoint of a longer relationship in which two further decisions - about timing and identity - continue to determine the value the firm extracts from the buyer over the lifetime of the relationship.

D7 - Timing (Post-Purchase and Lead Nurture Loops). The seventh decision is whether the buyer accepts that now is the right time to act - both initially, in committing to the first transaction, and subsequently, in expanding the relationship through repeat purchase, upgrade, or deeper engagement. The framework's fix when D7 is blocking before conversion is to state the cost of waiting in concrete terms; the framework's fix when D7 is blocking after conversion is to capitalise on the high-intent state immediately following the initial commitment, before the post-decision rationalisation closes the window of openness to further offers.

The functional execution layer is the design of nurture loops: automated email sequences calibrated to the specific moments at which the buyer is most receptive, instant SMS workflows for time-sensitive triggers, secondary educational series that maintain engagement during the consideration window for the next decision, and down-funnel retargeting that reinforces the relevance of the relationship when attention has migrated elsewhere.

D7 draws on the procrastination literature (Steel, 2007), hyperbolic-discounting models from behavioural economics (Laibson, 1997), and the post-decision-dissonance work of Festinger (1957) extended by Brehm (1956). The pre-conversion mechanism is the predictable mis-weighting of immediate inconvenience against deferred benefit; the post-conversion mechanism is the period of openness immediately following commitment, during which the buyer is actively constructing the rationalisations that will support the choice they have made and is therefore unusually receptive to consistent signals from the source. The window is finite: post-purchase rationalisation closes within hours to days, after which expansion offers are received with the same scepticism as cold offers.

The post-2020 literature has been kinder to D7 than to D3. Commitment-device interventions, including the Save More Tomorrow programme (Thaler and Benartzi, 2004), have replicated robustly and have produced large effects in long-term real-world deployment. The framework's emphasis on the immediately-post-conversion window is supported by the converging evidence on post-decision rationalisation as a behavioural phenomenon.

D8 - Identity and Social Risk (Lifetime Value and Referral Triggers). The eighth decision is whether the buyer judges the proposed choice - and the continuation, repetition, or recommendation of that choice - as defensible to their team, board, or peer group. The framework identifies D8 as the silent killer in high-ticket B2B decisions and prescribes framing the choice as the defensible one, with explicit reference to the peer group that has made it.

In the extended model, D8 governs not only the initial conversion but the lifetime value the firm extracts from the relationship through repeat purchase, expansion, and advocacy. The functional execution layer is the construction of systematic loops that prompt repeat purchase and organic brand advocacy: segmented email flows that recognise loyalty status, seasonal upsell triggers that re-engage at moments of natural relevance, subscription prompts that convert episodic buyers into continuous ones, and structured referral incentive programmes that translate the buyer's identity investment in the brand into observable behaviour the buyer's peer group can register.

D8 draws on the identity-based persuasion literature (Cialdini, 2016; Chance, 2022), the impression-management tradition in social psychology (Schlenker, 1980), the more recent work on social risk in organisational decision-making, and the consumer-behaviour literature on brand identification and advocacy (Aaker, 1991; Fournier, 1998). The mechanism is that buyers operating within institutional or social contexts evaluate choices not only on their objective merits but also on the social or professional consequences of the choice being scrutinised by others. A choice that appears risky in isolation may become defensible if the buyer can reference a peer group that has made the same choice; conversely, a choice that appears objectively optimal in isolation may be rejected if the buyer anticipates difficulty justifying it.

The extension of D8 from pre-conversion identity verification to post-conversion lifetime value rests on the observation that the same identity-based mechanism that governs whether a buyer commits also governs whether they continue, expand, and advocate. The buyer who has resolved D8 successfully at conversion has not merely permitted the transaction; they have begun an investment in the brand as identity infrastructure. The work of converting that investment into repeated and amplified behaviour is the lifecycle-marketing function the framework's execution layer is designed to perform.

D8 is less empirically settled than some of the framework's earlier decisions. Identity-based mechanisms are central to the most recent Cialdini synthesis (the addition of Unity as the seventh principle of influence in Pre-Suasion, 2016), but several of the supporting findings draw on the priming literature that has shown replication problems. The advocacy and referral literature, however, has its own robust empirical foundation in the work on word-of-mouth marketing (Berger, 2014) and net-promoter dynamics (Reichheld, 2003), which provides D8 with empirical support that does not depend on the contested priming findings.

3. The Framework's Structural Claims

The eight decisions, considered individually, are uncontroversial. Each draws on a well-established psychological literature, and most are familiar to any practitioner trained in the applied behavioural sciences. The framework's novel contribution lies in three structural claims about how the eight relate to each other.

3.1 Phased sequentiality

The first structural claim is that the eight decisions are sequential - that they must be addressed in order, that earlier decisions gate later ones, and that the failure of an earlier decision is sufficient to prevent any later decision from being reached regardless of how well the later decision is designed. The phase architecture (Capture, Convince, Close) adds a second layer to this claim: that the three phases represent distinct psychological states governed by distinct execution disciplines, and that an intervention executed at the wrong phase will fail even when it addresses the correct decision.

This is not a trivial claim. A funnel metaphor implies a continuous gradient of probability; the 8D framework implies a series of discrete gates, organised into phases that further structure the work of identifying which gate is closed. The practical implication is that intervention effort allocated to a later phase while an earlier phase is failing is wasted effort. A campaign that fails in Capture cannot be rescued by improvements to Convince; the prospects who would have benefited from Convince are not present to receive it.

The phased-sequentiality claim is consistent with the elaboration likelihood model (Petty and Cacioppo, 1986), which proposes that persuasion proceeds through stages and that earlier stages must be successfully traversed before later ones become consequential. It is consistent with the funnel-stage literature in marketing (McGuire, 1968). It is also consistent, more loosely, with the dual-process tradition, in which fast System 1 judgments precede slower System 2 evaluation. The extension into the Close phase finds support in the customer-lifecycle marketing literature (Reinartz and Kumar, 2003) and the post-decision-dissonance tradition that establishes the immediately-post-purchase period as psychologically distinct from both pre-purchase consideration and later steady-state consumption.

The strongest empirical defence of the phased-sequentiality claim is the diagnostic regularity it produces in practice. NOVI engagements consistently identify a primary blocking decision in a specific phase early in the engagement that, when addressed, produces measurable improvement disproportionate to the magnitude of the change made. This pattern is more consistent with a discrete-gate model than with a continuous-gradient model. It is not, however, a controlled empirical finding; it is a practitioner regularity that warrants more formal investigation.

3.2 Decomposition

The second structural claim is that the eight decisions exhaust the set of decisions relevant to buyer action and ongoing buyer behaviour - that a buyer who has successfully resolved all eight across the three phases will proceed to action and continue in the relationship, and that any failure to proceed or to extend implies the unsuccessful resolution of at least one of the eight.

This is a stronger claim and warrants explicit scepticism. The behavioural-science literature contains many additional candidate decisions that the 8D framework does not name: cognitive load (Sweller, 1988), affective state at the moment of decision (Lerner et al., 2015), the visitor's current goal hierarchy (Carver and Scheier, 1998), the broader macroeconomic and consumer-confidence environment (Akerlof and Shiller, 2009), habit-formation dynamics that may govern post-conversion behaviour more strongly than identity (Wood and Rünger, 2016), and others. The 8D framework either subsumes these under one of its eight (cognitive load might be subsumed within D2's specificity requirement; affect might be subsumed within D1's belonging recognition; habit might be subsumed within D7's timing or D8's identity loop) or treats them as background conditions rather than active decisions.

The practitioner-facing question is whether the simplification is operationally costly. The answer appears to be that for most commercial-conversion contexts it is not. The eight decisions cover the variance that is actionable at the page, sequence, or lifecycle level; the residual variance is largely environmental and not subject to practitioner control. A more complete model would be more accurate; the 8D model is more usable, and the trade-off is justifiable for applied work.

The trade-off is less clearly justifiable for academic research, where the cleaner question is whether the framework predicts more variance than competing decompositions. This is an open empirical question that the framework's commercial development has not yet been pressed to answer formally. The extension into the Close phase represents the area where this question is most open: the lifecycle-marketing literature contains several alternative decompositions of post-conversion buyer behaviour (recency-frequency-monetary models, lifecycle stage models, habit-formation models), and the comparative-fit question between these and the 8D model's D7-D8 framing has not been directly tested.

3.3 Diagnostic identifiability

The third structural claim is that the primary blocking decision in any given engagement can be empirically identified from observable behavioural data - that scroll depth, time on section, click-through rates, return visit rates, post-purchase engagement metrics, and similar proxies are sufficient to localise the failure point to a specific decision within a specific phase.

The claim is best understood as a working hypothesis rather than a settled finding. The measurement proxies the framework specifies - ad CTR for D1, time on page and scroll past 30% for D2, scroll to 60% for D3, scroll to proof section for D4, scroll to CTA area for D5, CTA click rate for D6, return visit rate for D7, About-page visits for D8 - are pragmatic operationalisations rather than validated instruments. They map intuitively onto the underlying decisions but the mapping has not been formally evaluated against alternative operationalisations.

The reasonable expectation is that these proxies are noisy indicators of the decisions they are intended to measure, that they covary with several confounds, and that they will sometimes mislead. A practitioner using them should treat the diagnostic identification as a working hypothesis to be tested by the response to the intervention, not as a settled finding to be acted upon without verification. This is consistent with the diagnostic discipline the framework's commercial application already imposes - the diagnostic is followed by an intervention whose effect is observed, and the diagnosis is revised if the intervention does not produce the expected response.

The phase architecture adds a useful constraint to the diagnostic task. Rather than asking which of eight decisions is failing (a question with eight roughly equally probable answers in the absence of further information), the framework asks which of three phases is failing, and within that phase which of two or three decisions. This sequential narrowing is consistent with established diagnostic reasoning in medicine and engineering, where the localisation of fault to a subsystem precedes the identification of the specific failing component.

4. The Framework Against the Methodological-Critical Literature

The applied behavioural sciences have, over the past decade, undergone substantial methodological reform. The Open Science Collaboration's (2015) Reproducibility Project found that only 36 to 47 per cent of high-prestige psychology findings replicated under direct replication conditions. Camerer et al. (2018) extended the finding to Nature and Science publications, with 13 of 21 social-science studies replicating successfully. Maier et al. (2022), applying publication-bias-correction methods to a large meta-analysis of nudging interventions (Mertens et al., 2022), found that the bias-corrected average effect size was an order of magnitude smaller than the headline figure. DellaVigna and Linos (2022), comparing academic-journal nudge effect sizes against the in-field effect sizes obtained by two large government nudge units, identified a roughly six-fold inflation factor. Protzko et al. (2023) demonstrated that when rigour-enhancing practices are adopted as an integrated system, replicability rates approaching 97 per cent of the original confirmatory effect are achievable.

The 8D model operates against this evidentiary backdrop. Three implications follow.

4.1 Effect-size calibration

The first implication is that the framework's expected effect sizes should be calibrated against the in-field benchmark rather than the academic-journal benchmark. A practitioner expecting a D3 reframe intervention to produce the dramatic effects implied by the popular-press literature on reframing will be disappointed; a practitioner expecting a 1- to 3-percentage-point improvement on the relevant conversion outcome will be approximately correctly calibrated. The framework's commercial track record is consistent with this calibration: documented engagement outcomes (a 55:1 average return on paid acquisition over four months, lifts of 312 per cent in qualified conversion against same paid spend) are large relative to the in-field nudge-unit benchmark, but the comparison is misleading because the framework's interventions typically combine multiple decisions across multiple phases and operate against funnels with substantial pre-existing inefficiency. The composite-intervention, high-baseline-inefficiency conditions are precisely the conditions under which behavioural intervention produces larger effects than the average in-field nudge.

The Close phase introduces an additional calibration consideration. Post-conversion interventions - the D7 nurture loops and D8 advocacy triggers - operate on populations who have already self-selected as receptive to the brand. Effect sizes in this population are systematically larger than effect sizes in cold or partially qualified populations, both because the population is more receptive and because the marginal cost of intervention is lower. Practitioner expectations for D7 and D8 interventions should accordingly be calibrated upward from the cold-traffic benchmark, while remaining bounded by the bias-corrected ceiling.

4.2 Mechanism-specific reliability

The second implication is that the eight decisions are not equally well-supported by the post-2020 evidence base. D1 (belonging recognition), D2 (problem recognition specificity), D4 (proof), D6 (proportionate ask), and D7 (timing, particularly the post-conversion window) draw on literatures that have largely survived the replication reform. D3 (causal reframe) and D8 (identity and social risk) draw substantially on the priming and identity-priming literatures that have not. D5 (source credibility) is shifting under the impact of algorithmic personalisation (Salvi et al., 2024) and warrants ongoing literature review.

The practitioner using the framework should weight intervention investment accordingly. Interventions targeting D3 should be designed with explicit a-priori effect-size expectations and explicit measurement, so that the failure of the intervention to produce the expected response is detected quickly. Interventions targeting D8 in the lifetime-value extension should be paired with the more empirically robust advocacy and referral literatures (Berger, 2014; Reichheld, 2003) rather than relying solely on the contested identity-priming work. Interventions targeting D1, D2, D4, D6, and D7 can be designed with more confidence, though still against in-field rather than academic benchmarks.

4.3 Implementation discipline

The third implication is that the framework's practical value lies less in the eight decisions themselves than in the diagnostic discipline the phase architecture imposes at the point of intervention. The dominant failure mode in applied conversion work is not the application of the wrong mechanism but the application of the right mechanism to the wrong problem - a creative refresh applied to a D3 failure, a CTA test applied to a D1 failure, an additional follow-up sequence applied to a D8 failure, a loyalty programme applied to a D7 failure. The 8D framework's contribution is to force the diagnostic question - which decision is failing, in which phase? - to be answered before the intervention is designed. This is operationally consistent with the implementation-science orientation that has come to dominate the post-2020 behavioural-policy literature (Michie, van Stralen and West, 2011; OECD, 2024).

The phase architecture sharpens this implementation discipline by reducing the diagnostic search space. A practitioner who can confidently localise the failure to one of three phases has already narrowed the relevant intervention literature substantially. A practitioner who must search across eight equally-weighted decisions faces a more diffuse diagnostic task and a correspondingly higher risk of misattribution.

This implementation-discipline framing also accommodates the framework's methodological limits. The eight decisions need not be the unique correct decomposition, and the measurement proxies need not be the unique correct operationalisations, for the framework to add value over the dominant alternative (the continuous-funnel model that terminates at conversion). The relevant comparison is not against an ideal framework but against the standard practice the framework displaces.

5. Limits

Four limits warrant explicit statement.

The first limit is empirical. The framework's structural claims - phased sequentiality, decomposition, and diagnostic identifiability - are working hypotheses supported by practitioner regularity rather than by formal controlled evaluation. A direct empirical test of the framework against alternative decompositions of the buyer-action and buyer-lifecycle sequence has not been conducted. Until such an evaluation is conducted, the framework should be understood as an operationally useful structuring device whose claim to be the correct decomposition remains open.

The second limit is contextual. The framework was developed in commercial and digital-conversion contexts and is most directly applicable to those contexts. Its application to charitable, governmental, and other settings is plausible but unevaluated. The eight decisions may map differently - and may not exhaust the relevant decisions - in contexts where the buyer-action relationship is structurally different from the commercial digital case. The Close phase, in particular, may translate poorly to one-time-transaction contexts (such as much of the charitable sector) where lifetime value and advocacy operate on different timescales and through different mechanisms.

The third limit is methodological. The framework, like all applied behavioural-science frameworks, inherits the calibration problems of the underlying literature. The publication-bias-corrected evidence base supports the framework's mechanisms but not at the effect sizes implied by the popular-press synthesis from which much practitioner training currently derives. The practitioner using the framework should pair it with methodological-critical reading (Maier et al., 2022; DellaVigna and Linos, 2022; Protzko et al., 2023) so that intervention forecasts are calibrated to the in-field benchmark rather than the inflated academic benchmark.

The fourth limit is temporal. The framework was developed before the maturation of large-language-model-mediated persuasion (Salvi et al., 2024) and before the wide deployment of algorithmic personalisation at the level of the individual conversion path. Both developments may alter the relative weight of the eight decisions - particularly D5 (credibility), where AI-mediated personalisation appears to substitute partially for human-credibility signals, and D3 (reframe), where AI systems may produce reframes at greater scale and personalisation than human practitioners. The framework's structure accommodates these developments, but its specific operational guidance may require revision as the algorithmic-persuasion literature matures.

6. Conclusion

The NOVI 8-Decision Growth Model operationalises a sequence of well-established psychological mechanisms in a form designed to produce diagnostic discipline across the entire customer relationship - from first contact through conversion to lifetime value and advocacy. Its central contribution is not the proposition that any one of the eight decisions matters - each is grounded in a substantial underlying literature - but the operational claim that they must be addressed in phased sequence, that the primary blocking decision can be empirically identified within a specific phase, and that interventions designed against the identified blocker will outperform interventions designed against execution-level symptoms. The three-phase architecture extends the framework's reach beyond the conventional conversion-funnel terminus to encompass the post-conversion lifecycle, where two further decisions - about timing and identity - continue to determine the value the firm extracts from the relationship.

The framework is internally coherent, operationally useful, and theoretically defensible. It is also empirically open. The structural claims are working hypotheses whose ultimate evaluation will require controlled comparison against alternative decompositions, and the framework's effect-size expectations should be calibrated against the in-field benchmark rather than the academic-journal benchmark.

The practitioner using the framework should treat it as a disciplined working model rather than as a settled theory. The diagnostic discipline it imposes - the requirement to identify the failing decision in the failing phase before designing the intervention - is the framework's central contribution to applied practice. The mechanisms it invokes are well-established but their magnitudes are smaller in the field than the popular literature suggests. The combination of operational discipline and calibrated expectations is what allows the framework to produce reliable outcomes across the variety of contexts in which it is applied.

The eight decisions are not the final word on what determines buyer action and ongoing buyer behaviour. They are a working decomposition that has earned its place in applied practice through producing consistent improvements over the standard alternatives. The intellectual honesty appropriate to the framework is to describe it on those terms - useful, principled, operationally disciplined, but provisional.

References

Aaker, D. A. (1991). Managing brand equity: Capitalizing on the value of a brand name. New York: Free Press.

Akerlof, G. A., and Shiller, R. J. (2009). Animal spirits: How human psychology drives the economy, and why it matters for global capitalism. Princeton: Princeton University Press.

Berger, J. (2014). Word of mouth and interpersonal communication: A review and directions for future research. Journal of Consumer Psychology, 24(4), 586-607.

Brehm, J. W. (1956). Postdecision changes in the desirability of alternatives. Journal of Abnormal and Social Psychology, 52(3), 384-389.

Burger, J. M. (1999). The foot-in-the-door compliance procedure: A multiple-process analysis and review. Personality and Social Psychology Review, 3(4), 303-325.

Camerer, C. F., Dreber, A., Holzmeister, F., Ho, T. H., Huber, J., Johannesson, M., et al. (2018). Evaluating the replicability of social science experiments in Nature and Science between 2010 and 2015. Nature Human Behaviour, 2(9), 637-644.

Carver, C. S., and Scheier, M. F. (1998). On the self-regulation of behavior. Cambridge: Cambridge University Press.

Chance, Z. (2022). Influence is your superpower: The science of winning hearts, sparking change, and making good things happen. New York: Random House.

Cialdini, R. B. (2007). Influence: The psychology of persuasion (Revised edition). New York: Harper Business.

Cialdini, R. B. (2016). Pre-suasion: A revolutionary way to influence and persuade. New York: Simon and Schuster.

DellaVigna, S., and Linos, E. (2022). RCTs to scale: Comprehensive evidence from two nudge units. Econometrica, 90(1), 81-116.

Doyen, S., Klein, O., Pichon, C.-L., and Cleeremans, A. (2012). Behavioral priming: It's all in the mind, but whose mind? PLOS ONE, 7(1), e29081.

Eyal, N. (2014). Hooked: How to build habit-forming products. New York: Portfolio.

Festinger, L. (1957). A theory of cognitive dissonance. Stanford: Stanford University Press.

Fournier, S. (1998). Consumers and their brands: Developing relationship theory in consumer research. Journal of Consumer Research, 24(4), 343-373.

Freedman, J. L., and Fraser, S. C. (1966). Compliance without pressure: The foot-in-the-door technique. Journal of Personality and Social Psychology, 4(2), 195-202.

Goldstein, N. J., Cialdini, R. B., and Griskevicius, V. (2008). A room with a viewpoint: Using social norms to motivate environmental conservation in hotels. Journal of Consumer Research, 35(3), 472-482.

Hallsworth, M., List, J. A., Metcalfe, R. D., and Vlaev, I. (2017). The behavioralist as tax collector: Using natural field experiments to enhance tax compliance. Journal of Public Economics, 148, 14-31.

Heider, F. (1958). The psychology of interpersonal relations. New York: John Wiley.

Hoeken, H., and Hustinx, L. (2009). When is statistical evidence superior to anecdotal evidence in supporting probability claims? The role of argument type. Human Communication Research, 35(4), 491-510.

Hovland, C. I., Janis, I. L., and Kelley, H. H. (1953). Communication and persuasion: Psychological studies of opinion change. New Haven: Yale University Press.

Johnson, E. J., and Goldstein, D. (2003). Do defaults save lives? Science, 302(5649), 1338-1339.

Kahneman, D. (2011). Thinking, fast and slow. New York: Farrar, Straus and Giroux.

Kahneman, D., and Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263-291.

Kelley, H. H. (1973). The processes of causal attribution. American Psychologist, 28(2), 107-128.

Laibson, D. (1997). Golden eggs and hyperbolic discounting. Quarterly Journal of Economics, 112(2), 443-478.

Lerner, J. S., Li, Y., Valdesolo, P., and Kassam, K. S. (2015). Emotion and decision making. Annual Review of Psychology, 66, 799-823.

Macrae, C. N., and Bodenhausen, G. V. (2000). Social cognition: Thinking categorically about others. Annual Review of Psychology, 51, 93-120.

Maier, M., Bartoš, F., Stanley, T. D., Shanks, D. R., Harris, A. J. L., and Wagenmakers, E.-J. (2022). No evidence for nudging after adjusting for publication bias. Proceedings of the National Academy of Sciences, 119(31), e2200300119.

McGuire, W. J. (1968). Personality and attitude change: An information-processing theory. In A. G. Greenwald, T. C. Brock, and T. M. Ostrom (Eds.), Psychological foundations of attitudes (pp. 171-196). New York: Academic Press.

McGuire, W. J. (1985). Attitudes and attitude change. In G. Lindzey and E. Aronson (Eds.), Handbook of social psychology (3rd ed., Vol. 2, pp. 233-346). New York: Random House.

Mertens, S., Herberz, M., Hahnel, U. J. J., and Brosch, T. (2022). The effectiveness of nudging: A meta-analysis of choice architecture interventions across behavioral domains. Proceedings of the National Academy of Sciences, 119(1), e2107346118.

Metzger, M. J., and Flanagin, A. J. (2013). Credibility and trust of information in online environments: The use of cognitive heuristics. Journal of Pragmatics, 59, 210-220.

Michie, S., van Stralen, M. M., and West, R. (2011). The Behaviour Change Wheel: A new method for characterising and designing behaviour change interventions. Implementation Science, 6, 42.

OECD. (2024). LOGIC: Good practice principles for mainstreaming behavioural public policy. Paris: OECD Publishing.

Open Science Collaboration. (2015). Estimating the reproducibility of psychological science. Science, 349(6251), aac4716.

Petty, R. E., and Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. Advances in Experimental Social Psychology, 19, 123-205.

Protzko, J., Krosnick, J., Nelson, L., Nosek, B. A., Axt, J., Berent, M., et al. (2023). High replicability of newly discovered social-behavioural findings is achievable. Nature Human Behaviour, advance online publication.

Reichheld, F. F. (2003). The one number you need to grow. Harvard Business Review, 81(12), 46-54.

Reinartz, W. J., and Kumar, V. (2003). The impact of customer relationship characteristics on profitable lifetime duration. Journal of Marketing, 67(1), 77-99.

Salvi, F., Horta Ribeiro, M., Gallotti, R., and West, R. (2024). On the conversational persuasiveness of large language models: A randomized controlled trial. arXiv preprint, arXiv:2403.14380.

Schlenker, B. R. (1980). Impression management: The self-concept, social identity, and interpersonal relations. Monterey: Brooks/Cole.

Sharot, T. (2017). The influential mind: What the brain reveals about our power to change others. New York: Henry Holt.

Soderberg, C. K., Errington, T. M., Schiavone, S. R., Bottesini, J., Thorn, F. S., Vazire, S., et al. (2015). Initial evidence of research quality of registered reports compared with the standard publishing model. Nature Human Behaviour, 5, 990-997.

Steel, P. (2007). The nature of procrastination: A meta-analytic and theoretical review of quintessential self-regulatory failure. Psychological Bulletin, 133(1), 65-94.

Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12(2), 257-285.

Thaler, R. H., and Benartzi, S. (2004). Save More Tomorrow: Using behavioral economics to increase employee saving. Journal of Political Economy, 112(S1), S164-S187.

Thaler, R. H., and Sunstein, C. R. (2008). Nudge: Improving decisions about health, wealth, and happiness. New Haven: Yale University Press.

Trope, Y., and Liberman, N. (2010). Construal-level theory of psychological distance. Psychological Review, 117(2), 440-463.

Wood, W., and Rünger, D. (2016). Psychology of habit. Annual Review of Psychology, 67, 289-314.

Word count: approximately 5,800 words (main text, excluding abstract, table, and references).