How to Write a Strong Introduction for a Top-Journal Paper

A practical framework for PhD students, built from real examples

academic writing
PhD advice

Most PhD students know their research better than anyone — but struggle to communicate why it matters. This post breaks down the anatomy of a top-journal introduction into 9 concrete steps, with skeleton sentences and a fully annotated real-world example.

Author

Bahman Rostami-Tabar

Published

March 16, 2026

“A good introduction does not summarise the paper. It builds the case that the paper needed to exist.”

Writing an introduction for a peer-reviewed journal article is one of the hardest things a PhD student has to learn — and almost nobody teaches it explicitly. Most of us learn by osmosis: reading dozens of papers and slowly internalising what “works,” without ever being able to articulate why.

This post makes that implicit knowledge explicit. I analysed four published introductions from top operations management journals — Decision Sciences, MSOM, JOM, and POM — and extracted the common structure underlying all of them. The result is a 9-step framework organised into three zones, complete with skeleton sentences you can adapt, and a fully annotated real-world example.

TipWho is this for?

This guide is aimed at PhD students writing their first journal paper, but it is equally useful for experienced researchers who want to sharpen their introductions before submission.

The Big Picture: Three Zones, Nine Steps

Before diving into individual steps, it helps to understand the macro-structure. Every strong introduction moves through three distinct zones:

Zone Purpose Steps
🟣 Motivation Convince the reader the topic matters 1 → 3
🟢 Positioning Show what gap this paper fills 4 → 5
🟡 Contribution Explain what this paper does and why it matters 6 → 9

The movement across zones follows a funnel logic: you start broad (the domain), narrow to a specific problem and tension, identify the gap, and zoom back out to the contribution. If your introduction feels flat, it is almost always because one of these zones is missing, underdeveloped, or out of order.

Here is the full 9-step sequence:

flowchart LR
    subgraph M["🟣 Zone 1 — Motivation"]
        direction TB
        S1["① Context & importance\nEstablish domain relevance\nwith quantified stakes"]
        S2["② Problem or tension\nIntroduce the friction,\ncost, or inefficiency"]
        S3["③ Non-triviality / paradox\nShow why the obvious\nsolution does not work"]
        S1 --> S2 --> S3
    end

    subgraph P["🟢 Zone 2 — Positioning"]
        direction TB
        S4["④ Literature positioning & gap\nWhat is known —\nand what is missing"]
        S5["⑤ Research question / objective\nState clearly what\nthis paper does"]
        S4 --> S5
    end

    subgraph C["🟡 Zone 3 — Contribution"]
        direction TB
        S6["⑥ Conceptual framing\nName and justify\nthe theoretical lens"]
        S7["⑦ Method overview\nSignal empirical\ncredibility"]
        S8["⑧ Key findings\nPreview results at direction\n+ mechanism level"]
        S9["⑨ Contributions & implications\nPosition value for\ntheory and practice"]
        S6 --> S7 --> S8 --> S9
    end

    M --> P --> C

    style M fill:#EEEDFE,stroke:#7F77DD,color:#3C3489
    style P fill:#E1F5EE,stroke:#1D9E75,color:#0F6E56
    style C fill:#FDF3E3,stroke:#BA7517,color:#5A3A08

Zone 1 — Motivation

The goal of this zone is simple: make the reader care. You are answering the question “why does this domain matter?” before you have even introduced the problem.

Step 1 — Context & Importance

Purpose: Open broadly with the domain’s scale and societal relevance. Quantify stakes immediately — numbers do the persuading, not adjectives.

This is your opening paragraph. Its only job is to make the reader care about the topic before you explain the problem. Think of it as answering: “Why should anyone — academic or practitioner — pay attention to this domain?”

The most effective technique is to anchor your opening in a concrete, memorable statistic. Numbers do the persuading; you do not need to write “this is an important topic.”

WarningCommon mistake

Starting with a vague claim like “Supply chains play an important role in modern society.” This tells the reader nothing specific and signals a weak introduction. Always anchor your opening in a real number, a real organisation, or a real observable trend.

Skeleton sentences

“[Domain] plays a critical role in [broader system], accounting for [statistic].”

“In [year], [X%] of [population] [did/experienced X] ([citation]).”

“[Domain] contributes to [stakeholder] by generating [benefit 1], [benefit 2], and [benefit 3].”

“Despite [challenging context], [organisation] reported [impressive metric], representing [X%] of [total].”

Examples from top journals

In-kind donations constitute a substantial and economically significant portion of the supply flowing to charities and humanitarian organisations worldwide. In 2017 alone, 52% of Americans donated clothing, food, or other personal items to such organisations (Non Profit Source 2018). These donations contribute to charities’ triple bottom line: they generate additional revenue, advance social welfare, and reduce environmental waste by rechanneling used goods away from landfills (Montgomery and Mitchell 2014). In 2020, despite the disruptions of the COVID-19 pandemic, the Salvation Army reported $598 million in revenue from 1,116 thrift stores across the United States — representing 18% of its total organisational revenue (Salvation Army 2021). Goodwill alone diverted 3.3 billion pounds of usable goods from landfills in the same year (Goodwill 2021).

Common seasonal illnesses such as influenza continue to affect the lives of people around the world. Global death rates from seasonal influenza are estimated to be around 300,000 to 646,000, with a majority of these numbers coming from countries situated in the tropical and subtropical regions. Flu outbreak patterns in tropical and subtropical regions tend to differ from those found in more temperate climates. As discussed by Hirve et al. (2016), many countries in these regions experience more than one outbreak or peak, generally showing two distinct peaks of influenza in a season.

In the United States, as many as 28.9 million individuals do not have basic healthcare insurance as of 2019. In 1986, Congress enacted the Emergency Medical Treatment and Liability Act (EMTALA), a federal law requiring all federally funded hospitals to provide basic medical examination and treatment to any patient who seeks care in the emergency department (ED), regardless of ability to pay. This allowed an uninsured patient who would have difficulty obtaining an outpatient or inpatient hospital admission to instead seek care at the nearest ED. For millions of uninsured Americans today, the hospital’s ED is often the only recourse for obtaining much-needed medical care.

Healthcare is a significant part of the economy — for example, accounting for 17.3% of U.S. gross domestic product (GDP), healthcare spending grew 4.1% in 2022, reaching $4.5 trillion, about $13,493 per capita (CMS.gov). Given its importance, healthcare operations have received attention from both practitioners and scholars. A fundamental characteristic of healthcare, posing unique challenges to providers (hospitals and clinics), is the uncertainty in care delivery. Seminal work (e.g., Balsa et al., 2003; Wennberg, 1985) contends that uncertainty is the single most important factor affecting physician behaviour.

Step 2 — Problem or Tension

Purpose: Introduce the friction. Contrast the positive picture of Step 1 with a concrete, costly problem.

After establishing the domain’s value, you reveal the complication. The classic pivot word is “However” — it signals a contrast and prepares the reader for bad news. The problem should be just as quantified as the hook in Step 1.

WarningCommon mistake

Writing “Charities face challenges” rather than “$1 million annually to dispose of 13 million pounds from just 30 stores.” When you name a specific organisation and a specific number, the problem becomes real and verifiable — not just an assertion.

Skeleton sentences

“However, not all [X] are [desirable outcome]. [Specific type] can neither [use 1] nor [use 2].”

“Instead, [problem] imposes significant [type of cost] on the very [stakeholder] it was meant to benefit.”

“[Organisation], for instance, spends [amount] annually on [problem activity] ([citation]).”

“While [X] offers clear benefits, it also creates [type of problem] that [who] must absorb.”

Examples from top journals

However, not all donated goods are useful. Low-quality items — stained clothing, torn blankets, broken furniture — can neither be resold in thrift stores nor distributed to beneficiaries. Instead, these inappropriate donations impose significant operational costs on the very organisations they are intended to support. Goodwill Northern New England, for instance, spends over $1 million annually to dispose of 13 million pounds of unsuitable items from only 30 thrift stores (Bookman 2021). Scaling this figure to the more than 3,000 Goodwill stores and 25,000 nonprofit resale shops across the United States (U.S. Census Bureau 2021) reveals a staggering systemic cost.

Similar to what is observed in temperate regions, there is a time gap between the identification of the circulating strain by the World Health Organization (WHO) and the onset of the first peak. Due to the long production lead time of the influenza vaccine (Anupindi et al., 2021; Deo & Corbett, 2009) and the short time window between strain declaration and the start of the first peak, vaccine supply uncertainty is typically present particularly for the first peak, potentially resulting in late vaccine deliveries. Vaccine demand uncertainty in subtropical developing nations is also commonly observed (Gong et al., 2021; Tapia-Conyer et al., 2021), as individuals may decide to vaccinate or not depending on factors such as the severity of infection, the incidence of the circulating strain, and perceived vaccine efficacy. The possibility of late vaccine delivery can further exacerbate demand uncertainty as patients that seek vaccination before or during the first peak may not return if vaccines are unavailable.

However, the use of the ED as the major provider of healthcare for the uninsured raises significant public health concerns. The ED is not designed for routine, preventative, or chronic care treatment. Rather, it is organized around its primary mission of providing round-the-clock emergency care for treating patients with conditions such as accident injuries, heart attacks, and strokes. Consequently, the substitution of long-term care with ED care may lead to recurring ED visits, further contributing to the spiraling cost of healthcare, and without any improvement in patient health. For example, a patient with a chronic condition such as chronic obstructive pulmonary disease (COPD) can be treated with appropriate medication or proper treatments such as home oxygen therapy if the patient has health insurance. Instead, the lack of insurance may lead the patient to continue making suboptimal, costly, and repeated visits to the ED.

From an operations management (OM) perspective, the many sources of uncertainty in healthcare include demand-side uncertainty (originating from a patient’s need for care), supply-side uncertainty (originating from physician and hospital characteristics), and clinical uncertainty (originating from technology, knowledge, and policy), to name a few. We focus on mix uncertainty, the uncertainty stemming from variation in the complexity of care needs of patients in a hospital department, and volume uncertainty, the uncertainty due to variation in the volume of patients served. While an extensive body of research pertains to streamlining healthcare operations (e.g., Devaraj et al., 2013), little research directly addresses the uncertainties in care needs or patient volumes, their impact on the care-delivery process, and mitigating mechanisms.

Step 3 — Non-Triviality / Paradox

Purpose: Show why the obvious solution does not work — and therefore why the paper is intellectually necessary.

ImportantThe most commonly missing step

This is the step most often absent from PhD student introductions — and the one that most separates good papers from great ones. A reviewer’s immediate instinct after Step 2 is: “Why can’t someone just fix this in the obvious way?” Step 3 answers that question.

Ask yourself: “What would a smart, experienced manager or policymaker try first — and why would that fail?” The answer to that question is your Step 3. Without it, your paper fills a gap nobody noticed. With it, your paper solves a puzzle that has stumped practitioners and academics alike.

Two to four sentences are sufficient. The key signal words are “yet”, “however”, or “but.”

Skeleton sentences

“This creates a striking paradox: [party A’s action] ends up harming [party B] they intended to help.”

“Yet [stakeholder] cannot simply [obvious solution], because doing so risks [unintended consequence].”

“Although one might expect [intuitive outcome], in practice [counterintuitive reality] ([citation]).”

“[Stakeholder] therefore faces a dilemma: [option A] imposes [cost A], but [option B] risks [cost B].”

Examples from top journals

This creates a striking paradox: donors’ good intentions end up harming the organisations they intend to support. Yet charities find themselves unable to simply reject low-quality donations. Declining a goodwill offer risks damaging the donor relationship and discouraging future support — a serious concern given that recurring donors are estimated to give 440% more over their lifetime than one-time donors (Classy 2018). As Daniels and Valdés (2021) demonstrate, donors use rejection experiences as self-serving justification to disengage from future giving. In practice, therefore, charities face an uncomfortable dilemma: accepting inappropriate donations imposes direct operational costs, but refusing them may undermine long-term financial sustainability.

The decision on how many vaccine units to order and when to receive the order is complicated by factors such as the presence of multiple peaks, demand uncertainty, as well as the possibility of late deliveries before the first peak. In order to deliver orders on time, manufacturers need to start vaccine production before the official strain declaration by the WHO (Dai, 2015). Such a decision is risky and can lead to production loss or delays. In order to align incentives in the vaccine supply chain, procurement contracts are negotiated seeking to share risks and rewards — yet practitioners emphasised that negotiating contractual parameters is typically a challenging endeavour due to the uncertainty that originates from both the supply and demand sides.

Universal healthcare is motivated by the pressing need to solve these problems of timely access to care and overall population health. Yet economic theory and the existing literature do not provide a clear answer when examining various forms of health insurance expansions in the United States. For example, Finkelstein et al. (2016) and Dresden et al. (2017) find increased ED use following Medicaid expansion, while Chen et al. (2011) and Gotanda et al. (2019) find no significant change, and Kolstad and Kowalski (2012) and Sommers et al. (2016) find decreased ED usage. This contradictory evidence makes it genuinely unclear whether expanding insurance will reduce or worsen the very ED burden it aims to address.

Variation in patient volume may considerably disrupt process flows and operational planning, thus making it a key source of uncertainty in the care-delivery process (Devaraj et al., 2013; Jacobs & Swink, 2011). Understanding the implications of volume uncertainty in healthcare operations is therefore critical as hospitals struggle to cope with increasing demand for access to care. Yet unlike in a factory setting, individual patients have complex care needs, manifest in a multitude of chronic and comorbid conditions — making it impossible to simply standardise the response to uncertainty the way a manufacturer might. These conditions of varied severity interact to create considerable challenges in both the process and outcomes of care, for which no simple operational fix exists.

Zone 2 — Positioning

The goal of this zone is to show the reader exactly where your paper sits in the literature — and why that space has not been filled before.

Step 4 — Literature Positioning & Gap

Purpose: Show what is known, and then state precisely what is missing.

This step has two parts that must work together:

  1. A brief, targeted review of what prior research has established
  2. A precise gap statement — what those papers did not examine

Think of it as: “The literature knows A, B, and C — but it does not know D. D is what this paper studies.”

WarningCommon mistake

Writing an exhaustive literature review here. Every citation in the introduction should serve one purpose: building the logical path toward your gap. If removing a citation does not weaken your argument, it probably does not belong here.

Test your gap with this question: “Could a reviewer confirm that this specific thing has not been studied before?” If yes, your gap is precise enough. If not, sharpen it.

Skeleton sentences

“A growing body of research has examined [broad topic], focusing on [aspect 1], [aspect 2], and [aspect 3] ([citations]).”

“Prior research has focused predominantly on [what is known]. However, little is known about [the gap].”

“While [related construct] has received significant attention ([citation]), the [specific aspect] and its implications for [outcome] remain unexplored.”

“[Phenomenon] has heretofore not been examined in the context of [your setting] — a gap this paper addresses directly.”

Examples from top journals

A growing body of research has examined donor behaviour and the effectiveness of interventions designed to increase charitable giving (Croson and Treich 2014; Shang and Croson 2009; Martin and Randal 2008). Within this literature, information disclosure and social norm interventions have received particular attention, with demonstrated effectiveness across domains from nutrition labelling to energy conservation (Thaler and Sunstein 2009; Goldstein et al. 2008). However, this literature has focused predominantly on monetary donations and on increasing the quantity of giving. Little is known about how behavioural interventions can improve the quality of in-kind donations — a fundamentally different behavioural target with distinct operational consequences for recipient organisations.

The existing literature on contracts considers that both VMs and VPAs have profit-maximizing objectives. Research studies on contracts typically consider that both the VM and VPAs have profit-maximizing objectives. While all VMs focus on maximizing profit, the objective for a VPA will change depending on whether they are for-profit or not-for-profit providers. Given that practitioners are familiar with the use of late penalty and buyback options, the suitability of proposed hybrid contracts must be examined for both for-profit as well as not-for-profit VPAs — a distinction the existing contracts literature has not addressed.

The existing work has primarily focused on the impact of universal healthcare on the total ED volume. Also, many studies such as Miller (2012) use ED data only and categorise them into “inpatient ED visits” and “outpatient ED visits.” Although the aggregate ED visit is an important dimension of public health performance, it is also crucial to examine how the same set of individuals choose among different types of hospital resources once they gain access to health insurance. This form of examination necessitates following individual patients over time, and examining their choices in not only the ED but also other hospital settings both before and after the implementation of the policy — something the existing literature has not done.

A growing body of work has examined the impact of demand volume and workload on operational planning and service-delivery outcomes (Powell et al., 2012; Tom & Serguei, 2014; Kuntz et al., 2015; Berry Jaeker and Tucker, 2017). Similarly, a rich body of prior research has examined the complexity of patient-care needs and its implications for patient routing, care outcomes, and hospital portfolio strategy (Kc and Terwiesch, 2011; Clark and Huckman, 2012; Kuntz et al., 2019; Thirumalai and Devaraj, 2024). However, while this body of research has largely focused on the level of complexity and volume of care, the variation in the complexity of patient-care needs and its impact on performance have heretofore not been examined — a gap this study addresses directly.

Step 5 — Research Question / Objective

Purpose: State clearly and concisely what this paper does.

After building the problem and identifying the gap, the reader is waiting for one thing: what exactly does this paper do? Answer that in one or two sentences.

Too vague Too technical Just right
“We study charitable giving.” “We estimate a probit model with facility fixed effects…” “The goal of this paper is to find a practical solution that reduces inappropriate donations without losing donors.”

If you have more than one research question, number them. Never list more questions than you can fully address.

Skeleton sentences

“The goal of this paper is to [verb: identify / develop / examine / quantify] [what] in the context of [setting].”

“Specifically, we ask: (i) [RQ1]? (ii) [RQ2]?”

“This paper investigates [phenomenon] and its implications for [stakeholder / outcome].”

“We study whether [intervention / factor] can [outcome] without [undesirable side effect].”

Examples from top journals

The goal of this paper is to find a practical solution that reduces the number of inappropriate in-kind donations a charity receives without losing donors. Specifically, we ask: can low-cost behavioural interventions — embedded within existing charity communication channels — nudge donors to voluntarily improve the quality of their donations, while preserving long-term donor retention?

Motivated by the above examples and context, we first study whether all vaccine units should be obtained before the start of the first peak (known as Obtain All or OA policy) or if some should be obtained before each of the two peaks (known as Obtain Some or OS policy). Next, we examine the coordination mechanism of the new proposed hybrid contracts and how they apply to both private for-profit and public not-for-profit supply chains. Specifically, we seek answers to the following research questions: (R1) What should be the optimal vaccine procurement policy for the public and the privately owned VPA? (R2) Under what conditions will the hybrid contracts coordinate the private structure supply chain? Will the same contracts coordinate the public structure supply chain? (R3) Do VPA and VM always prefer one hybrid contract over the other?

In this paper, we address the following fundamental questions concerning universal healthcare: (1) Does universal healthcare lead previously uninsured individuals to avoid the ED in favor of alternative care settings? (2) Do different groups of patients respond differently to this availability of choice? (3) Does universal healthcare improve public health outcomes among the previously uninsured?

Using a large dataset of 830,853 patient discharges in 26 clinical departments across 731 hospitals in five U.S. Midwest states, we empirically examine the impact of mix and volume uncertainties on the care-delivery process — specifically the number of procedures (NPR) offered to patients and how long patients spend in the system (length of stay, LOS). We also examine the mitigating effects of department utilisation levels and the operational decision to focus in related areas on addressing uncertainty in healthcare.

Zone 3 — Contribution

The goal of this zone is to show the reader what your paper does, how it does it, what it finds, and why all of this matters.

Step 6 — Conceptual Framing

Purpose: Name and justify the theoretical lens or conceptual framework guiding the analysis.

This step is often labelled “optional” in writing guides. In most OM, management, and social science journals it is effectively expected — especially if your paper makes a theoretical contribution.

Two things are required:

  1. Name your theory or framework
  2. Justify why it applies to this specific setting

Simply writing “we draw on social norm theory” is not enough. You must explain why that theory is the right lens for this problem. One sentence of justification transforms the theory choice from arbitrary to deliberate.

Skeleton sentences

“We draw on [theory] ([citation]), which [core argument of theory].”

“This framework is particularly well-suited to [your setting] because [reason it fits].”

“We conceptualise [construct] as [definition], which operates through [mechanism].”

“We focus on [X] and [Y], which influence [outcome] through [mechanism 1] and [mechanism 2] respectively.”

Examples from top journals

We draw on behavioural economics and the nudge framework (Thaler and Sunstein 2009), which argues that individuals can be steered toward desired behaviours through low-cost, voluntary interventions that preserve freedom of choice. This approach is particularly well-suited to the charity context: unlike harder policy instruments such as taxes or regulatory bans, nudges create less friction between the organisation and its donors. We focus on two mechanisms — information disclosure and descriptive social norms — which influence behaviour through distinct psychological pathways: the former by making consequences more salient (Loewenstein et al. 2014), the latter by leveraging individuals’ desire to conform to their reference group (Cialdini et al. 1990).

In order to align incentives in the vaccine supply chain, procurement contracts are negotiated seeking to share risks and rewards with the purpose of better aligning divergent interests of healthcare providers and vaccine manufacturers (VMs). A “flexible” contract is one that offers sufficient flexibility to allow for any division of supply chain profits between VM and VPA by adjusting parameter values (Cachon, 2003, p. 230), providing greater ease of implementation by giving stakeholders more flexibility to negotiate contract parameters and reach a contractual agreement. We study two new hybrid contracts — the 2-Wholesale Buyback (2WB) and the Penalty-1 and Buyback (P1B) — that utilise various combinations of the wholesale price, buyback price, and late penalty, because practitioners indicated familiarity with these mechanisms.

Our paper is motivated by the desire to understand how patient choice is impacted by universal healthcare. The relative preference for the type of healthcare service is also likely to vary according to patient characteristics. We postulate that, by allowing individuals more choices, universal healthcare can lower the likelihood of ED visits across various patient segments through a substitution mechanism: patients who previously had no alternative to the ED can now seek more appropriate care in outpatient or inpatient settings. Moreover, allowing patients to seek more appropriate care could even improve patient health through better match between condition and care setting.

Our study builds on theoretical perspectives from Ramasesh and Browning’s (2014) uncertainty framework (hereinafter referred to as the R&B framework) and organisational information processing theory (OIPT) (e.g., Daft & Lengel, 1986; Galbraith, 1977). The R&B framework provides the conceptual foundation to study the sources and measurement of uncertainty and is well-suited to the healthcare setting because each hospital–patient encounter may be viewed as a project — that is, temporary work with unique outcomes. OIPT then offers an organisation-design perspective for understanding how hospitals might mitigate the effects of uncertainty on performance: specifically, by reducing the need for information processing and/or increasing information-processing capabilities through related focus and operational slack.

Step 7 — Method Overview

Purpose: Signal that your paper has the empirical infrastructure to answer the research questions posed in Step 5.

Think of this as the credibility check. Three things should appear:

  • The type of study (field experiment, survey, archival data, simulation)
  • The scale of the data (N = 763 households; 5 countries; 830,000 discharges)
  • The key design feature that makes the study credible (random assignment, panel structure, matching technique)
WarningCommon mistake

Over-detailing the method in the introduction. If you spend three paragraphs on methodology here, the reader loses the narrative thread. One clear paragraph is enough.

Skeleton sentences

“We conduct a [type of study] in collaboration with [organisation/setting] between [time period].”

“We collect [type of data] from [N units] across [setting / countries / time period].”

“Using a [between/within]-subjects design, we compare [condition A] against [condition B] and a control group.”

“To address [identification concern], we employ [technique] ([citation]).”

Examples from top journals

We test these interventions in a field experiment conducted in collaboration with the Society of St. Vincent de Paul of Arizona (SVdP) between October 31 and November 11, 2020. We collected a panel dataset from 763 households in a between-subjects design with three groups: social norm message, information disclosure message, and control. We subsequently tracked donor retention across all groups over 12 months to capture potential long-term spillover effects.

In order to gain a better understanding of the vaccine procurement contracts used in practice, we conducted interviews with the COO of AIG multispecialty hospital (an 800-bed facility in India), a Global Sales Manager at a major global vaccine manufacturer, and a Senior Manager of a global biopharmaceutical company with presence in India. We develop analytical models for both private for-profit and public not-for-profit supply chain structures, examining the 2WB and P1B hybrid contracts under conditions of demand uncertainty and possible late delivery before the first influenza peak.

To answer our research questions, we study a natural experiment from the U.S. state of Massachusetts, which in April 2006 became the first state to pass a comprehensive healthcare bill requiring every legal resident to obtain basic health insurance. We examine the revealed healthcare choices of individuals both before and after the implementation of the law over a 6-year period between 2004 and 2010, observing all visits to all major healthcare facilities across the entire state — including EDs, inpatient units, and outpatient units — and tracking insurance status at each encounter.

Using a large dataset of 830,853 patient discharges in 26 clinical departments across 731 hospitals in five U.S. Midwest states, we empirically examine the impact of mix and volume uncertainties on the care-delivery process. We use the mean-adjusted variation in the complexity of patients served within a primary diagnostic category as a measure of mix uncertainty, and the variation in patient volume across time periods as a measure of volume uncertainty within a department.

Step 8 — Key Findings

Purpose: Preview results at direction + mechanism level — not statistical detail.

The right level of abstraction is: direction + mechanism, not regression coefficients. The reader has invested attention through seven steps — now you reward them.

TipThe power of a surprising result

Include at least one counterintuitive finding. In the example below, the failure of information disclosure — one of the most commonly used charity communication tools — goes against conventional wisdom and is the single most memorable sentence in the introduction.

The best findings previews also address a second-order outcome that directly resolves the concern raised in Step 3. The introduction and findings should answer each other.

Skeleton sentences

“Our results show that [intervention/factor] significantly [increases/decreases] [outcome], suggesting that [mechanism].”

“Contrary to conventional wisdom, [expected result] did not [expected effect].”

“We find that [finding], with the effect being stronger for [subgroup/condition].”

“Importantly, [secondary finding], indicating that [concern raised in Step 3] does not materialise in practice.”

Examples from top journals

Our results show that the social norm intervention significantly improved the quality of in-kind donations — an effect that remained stable over time. Contrary to conventional wisdom, information disclosure did not alter donor behaviour. While both interventions produced an initial, temporary decline in return rates, this disparity converged fully at 12 months, indicating no lasting harm to donor retention. A conservative estimate suggests the intervention reduced inappropriate donations by approximately 50%, at zero direct operating cost to the charity.

Our results show that contingent on the various costs and the likelihood of the VM delivering the vaccine on time before the first peak, the OS policy is optimal for both the public and the private VPA. We find that hybrid contracts coordinate the supply chain within a wide range of wholesale prices, and that they tend to coordinate the private structure supply chain under more scenarios than the public structure supply chain. While it is intuitive to expect that the public VPA will procure the highest number of vaccines, our results suggest that under certain conditions the privately owned VPA will achieve greater vaccine coverage.

We find that universal healthcare has a marked impact on individual healthcare choice: upon obtaining insurance, a previously uninsured patient is significantly less likely to continue choosing the ED, with approximately a 6.3% reduction in the probability of ED visits relative to other hospital care. We further find heterogeneity across patient segments — patients with chronic conditions are more likely to eschew the ED in favour of outpatient and inpatient care, whereas frequent ED users are less likely to avoid the ED post-policy. Importantly, frequent fliers and chronic patients see a reduction in short-term ED revisits following universal healthcare — a sign of improved healthcare quality for the patient segments that matter most.

Our analysis indicates that mix uncertainty in patient-care needs triggers more procedures (higher NPR), suggestive of the information needs of providers. On the other hand, volume uncertainty drives providers to reduce procedures (lower NPR) and cause delays (longer LOS), suggestive of bottlenecks from variable process flow. We find that breadth of expertise in relevant areas can lower patients’ LOS and reduce the need for more procedures. We confirm that higher utilisation levels motivate a speed-up in healthcare operations, consistent with prior observations in the literature. Lastly, we find considerable heterogeneity in the effects of uncertainty across hospital types and departments.

Step 9 — Contributions & Implications

Purpose: Separate from findings — explain why those findings matter for theory and practice.

ImportantDo not merge this with Step 8

Step 8 answers “what did you find?” Step 9 answers “why does that matter?” These require different sentences. Merging them is the single most common structural weakness in student introductions.

A good contributions paragraph has two parts:

  • Theoretical contribution — which literature stream does the paper extend? What is new?
  • Practical implication — what should managers, policymakers, or practitioners do differently?

Be specific. “This contributes to the operations management literature” is too vague. “This extends the behavioural OM literature to the domain of in-kind supply quality” is specific enough for a reviewer to evaluate. Write with declarative confidence: “This study contributes…” — not “We hope this might contribute…”

Skeleton sentences

“This study contributes to the [specific literature stream] by [what it does that is new].”

“We extend [prior work / theory] to [new domain/context], demonstrating that [finding].”

“Our findings have direct implications for [practitioner type] who [face the problem described in Steps 1–3].”

“By showing that [finding], this paper provides [stakeholder] with an actionable [tool/guideline] to [desired outcome].”

Examples from top journals

This study makes several contributions. Theoretically, it extends the behavioural operations management literature to a new domain — the quality of in-kind supply — and demonstrates that behavioural interventions can shape not only the volume but the composition of charitable supply flows. It also contributes to the literature on behavioural spillovers by tracking how an initial intervention affects subsequent donor behaviour over a 12-month horizon (Dolan and Galizzi 2015). Practically, the findings offer an immediately actionable, resource-free tool for charities seeking to improve operational efficiency without jeopardising donor relationships — a particularly valuable insight for organisations that must do more with less.

We extend our analysis to consider both public and private VPAs existing simultaneously in the market, analysing vaccine allocation under conditions of limited budget and limited vaccine production capacity. Interestingly, with the manufacturer’s capacity becoming restrictive, the private supply chain achieves higher vaccine coverage than the public supply chain. These findings have important policy implications, particularly for developing nations with limited public funds, by showing that contract design — not just ownership structure — can drive vaccine access, and that hybrid contracts offer a practical, flexible tool for supply chain coordination in settings with both supply and demand uncertainty.

To the best of our knowledge, ours is the first paper to utilise a panel of individual patients observed over a long period of time to assess the impact of universal healthcare on the choices of individuals, and the first to find a positive quality of care effect for certain patient segments. These findings have important public health and socioeconomic implications: they show that universal healthcare can help guide individuals toward appropriate care settings and improve health quality for the groups — chronic care patients and frequent ED users — that contribute disproportionately to spiraling healthcare costs. These findings can inform the ongoing nationwide public policy debate around the costs and benefits of universal healthcare.

Overall, our study provides further insight into the forces underlying mix and volume uncertainties in healthcare, empirically exploring how uncertainty affects healthcare delivery and outcomes. Theoretically, we extend the R&B uncertainty framework and OIPT to the healthcare operations context, demonstrating that different types of demand-side uncertainty have distinct and sometimes opposing effects on care-delivery processes. Practically, our study highlights the role of two operational levers — related focus and utilisation management — in mitigating the negative effects of uncertainty on patient-care delivery, providing hospital administrators with actionable guidance for managing uncertain care environments.

Putting It All Together

Here is how all nine steps come together as a single, flowing introduction — exactly as it would appear in a published paper. The step labels disappear; the structure is felt, not announced.

[Step 1] In-kind donations constitute a substantial and economically significant portion of the supply flowing to charities and humanitarian organisations worldwide. In 2017 alone, 52% of Americans donated clothing, food, or other personal items to such organisations (Non Profit Source 2018). These donations contribute to charities’ triple bottom line: additional revenue, social welfare, and environmental sustainability (Montgomery and Mitchell 2014). In 2020, the Salvation Army reported $598 million in revenue from 1,116 thrift stores — representing 18% of its total revenue (Salvation Army 2021).

[Step 2] However, not all donated goods are useful. Low-quality items — stained clothing, torn blankets, broken furniture — can neither be resold nor distributed to beneficiaries. Goodwill Northern New England spends over $1 million annually to dispose of 13 million pounds of unsuitable items from only 30 thrift stores (Bookman 2021), and the systemic cost across 25,000 nonprofit resale shops in the United States (U.S. Census Bureau 2021) is staggering.

[Step 3] This creates a striking paradox: donors’ good intentions end up harming the organisations they intend to support. Yet charities cannot simply reject low-quality donations — declining an offer risks damaging the donor relationship, and recurring donors give 440% more over their lifetime than one-time donors (Classy 2018). Charities therefore face an uncomfortable dilemma: accepting inappropriate donations imposes direct costs, but refusing them risks long-term financial sustainability.

[Step 4] A growing body of research has examined interventions designed to increase charitable giving (Croson and Treich 2014; Shang and Croson 2009), with particular attention to information disclosure and social norm mechanisms (Thaler and Sunstein 2009; Goldstein et al. 2008). However, this literature has focused on the quantity of giving. Little is known about how behavioural interventions can improve the quality of in-kind donations — a different behavioural target with distinct operational consequences.

[Step 5] The goal of this paper is to find a practical solution that reduces the number of inappropriate in-kind donations a charity receives without losing donors. Specifically: can low-cost behavioural interventions nudge donors to voluntarily improve donation quality while preserving long-term retention?

[Step 6] We draw on the nudge framework (Thaler and Sunstein 2009), particularly well-suited to the charity context because, unlike taxes or bans, nudges create minimal friction with donors. We focus on information disclosure — which makes consequences more salient (Loewenstein et al. 2014) — and descriptive social norms — which leverage conformity to peer behaviour (Cialdini et al. 1990).

[Step 7] We conduct a field experiment with the Society of St. Vincent de Paul of Arizona across 763 households in a between-subjects design (October–November 2020), with social norm, information disclosure, and control conditions, tracking donor retention over 12 months.

[Step 8] The social norm intervention significantly improved donation quality — an effect stable over time. Contrary to conventional wisdom, information disclosure had no effect. The initial decline in return rates fully converged at 12 months, confirming no long-term harm to donor retention. The intervention reduced inappropriate donations by approximately 50% at zero direct operating cost.

[Step 9] Theoretically, this study extends the behavioural OM literature to the quality of in-kind supply, showing that interventions can shape supply composition — not just volume. Practically, it provides charities with an immediately actionable, cost-free tool to improve operational efficiency without jeopardising donor relationships (Dolan and Galizzi 2015).

[Step 1] Common seasonal illnesses such as influenza continue to affect the lives of people around the world. Global death rates from seasonal influenza are estimated to be around 300,000 to 646,000, with a majority of these numbers coming from countries situated in the tropical and subtropical regions. Many countries in these regions experience more than one outbreak or peak, generally showing two distinct peaks of influenza in a season (Hirve et al., 2016).

[Step 2] Due to the long production lead time of the influenza vaccine and the short time window between strain declaration and the start of the first peak, vaccine supply uncertainty is typically present particularly for the first peak, potentially resulting in late vaccine deliveries. Vaccine demand uncertainty in subtropical developing nations is also commonly observed, as individuals may decide to vaccinate or not depending on factors such as the severity of infection, the incidence of the circulating strain, and perceived vaccine efficacy. The possibility of late vaccine delivery can further exacerbate demand uncertainty as patients that seek vaccination before or during the first peak may not return if vaccines are unavailable.

[Step 3] The decision on how many vaccine units to order and when to receive the order is complicated by factors such as the presence of multiple peaks, demand uncertainty, as well as the possibility of late deliveries before the first peak. In order to deliver orders on time, manufacturers need to start vaccine production before the official strain declaration by the WHO — a risky decision that can lead to production loss or delays. Practitioners confirmed that negotiating contractual parameters is a challenging endeavour due to uncertainty originating from both the supply and demand sides.

[Step 4] The existing literature on contracts considers that both VMs and VPAs have profit-maximizing objectives. Research studies on contracts typically consider that both the VM and VPAs have profit-maximizing objectives. However, the suitability of contracts for not-for-profit VPAs — whose objective is to maximise vaccine coverage rather than profit — has not been examined, representing a critical gap given the dual public-private structure of healthcare in developing economies.

[Step 5] Motivated by the above context, we first study whether all vaccine units should be obtained before the start of the first peak (Obtain All or OA policy) or if some should be obtained before each of the two peaks (Obtain Some or OS policy). Specifically: (R1) What should be the optimal vaccine procurement policy for the public and the privately owned VPA? (R2) Under what conditions will the hybrid contracts coordinate the supply chain? (R3) Do VPA and VM always prefer one hybrid contract over the other?

[Step 6] We draw on the supply chain contracts literature (Cachon, 2003), focusing on “flexible” contracts that offer sufficient flexibility to allow for any division of supply chain profits between VM and VPA by adjusting parameter values. We study two new hybrid contracts — the 2WB and the P1B — that utilise combinations of wholesale price, buyback price, and late penalty, grounded in practitioners’ familiarity with these mechanisms.

[Step 7] We conducted interviews with the COO of AIG multispecialty hospital (an 800-bed facility in India), a Global Sales Manager at a major global vaccine manufacturer, and a Senior Manager of a global biopharmaceutical company. We develop analytical models for both private for-profit and public not-for-profit supply chain structures under conditions of demand uncertainty and possible late delivery before the first influenza peak.

[Step 8] Our results show that contingent on the various costs and the likelihood of on-time delivery, the OS policy is optimal for both the public and the private VPA. Hybrid contracts coordinate the supply chain within a wide range of wholesale prices, and tend to coordinate the private structure under more scenarios than the public structure. Surprisingly, under certain conditions the privately owned VPA achieves greater vaccine coverage than the public VPA.

[Step 9] We extend our analysis to consider both public and private VPAs simultaneously in the market. These findings have important policy implications, particularly for developing nations with limited public funds, by showing that contract design — not just ownership structure — can drive vaccine access. The hybrid contracts offer a practical, flexible coordination tool for supply chains facing both supply and demand uncertainty.

[Step 1] In the United States, as many as 28.9 million individuals do not have basic healthcare insurance as of 2019. In 1986, Congress enacted EMTALA, requiring all federally funded hospitals to provide basic medical examination and treatment to any patient seeking care in the ED, regardless of ability to pay. For millions of uninsured Americans today, the hospital’s ED is often the only recourse for obtaining much-needed medical care.

[Step 2] However, the use of the ED as the major provider of healthcare for the uninsured raises significant public health concerns. The ED is not designed for routine, preventative, or chronic care. Consequently, the substitution of long-term care with ED care may lead to recurring ED visits, further contributing to the spiraling cost of healthcare — without any improvement in patient health.

[Step 3] Universal healthcare is motivated by the pressing need to solve these problems. Yet economic theory and the existing literature do not provide a clear answer: some studies find increased ED use following insurance expansion (Finkelstein et al., 2016), others find no change (Chen et al., 2011), and still others find decreased ED usage (Kolstad and Kowalski, 2012). This contradictory evidence makes it genuinely unclear whether expanding insurance will reduce or worsen the ED burden it aims to address.

[Step 4] The existing work has primarily focused on the impact of universal healthcare on the total ED volume. However, it is also crucial to examine how the same set of individuals choose among different types of hospital resources once they gain access to health insurance — a form of examination that necessitates following individual patients over time across ED, inpatient, and outpatient settings, which the existing literature has not done.

[Step 5] In this paper, we address the following fundamental questions: (1) Does universal healthcare lead previously uninsured individuals to avoid the ED in favour of alternative care settings? (2) Do different groups of patients respond differently? (3) Does universal healthcare improve public health outcomes among the previously uninsured?

[Step 6] Our paper is motivated by the desire to understand how patient choice is impacted by universal healthcare. We postulate that, by allowing individuals more choices, universal healthcare can lower the likelihood of ED visits across various patient segments through a substitution mechanism — allowing patients to seek more appropriate care and potentially improving health outcomes.

[Step 7] To answer our research questions, we study a natural experiment from Massachusetts, which in April 2006 became the first state to require every legal resident to obtain basic health insurance. We examine the revealed healthcare choices of individuals over a 6-year period between 2004 and 2010, observing all visits to all major healthcare facilities across the entire state.

[Step 8] We find that upon obtaining insurance, a previously uninsured patient is significantly less likely to choose the ED — approximately a 6.3% reduction in the probability of ED visits relative to other hospital care. Patients with chronic conditions are more likely to eschew the ED in favour of outpatient care, whereas frequent ED users are less likely to avoid the ED. Importantly, frequent fliers and chronic patients see a reduction in short-term ED revisits — a sign of improved healthcare quality for the highest-cost patient segments.

[Step 9] Ours is the first paper to utilise a long panel of individual patients to assess the impact of universal healthcare on individual healthcare choices, and the first to find a positive quality-of-care effect for certain patient segments. These findings have important public health and socioeconomic implications — particularly for chronic care patients and frequent ED users, who contribute disproportionately to spiraling healthcare costs — and can inform the ongoing national debate around the costs and benefits of universal healthcare.

[Step 1] Healthcare is a significant part of the economy — accounting for 17.3% of U.S. GDP, healthcare spending grew 4.1% in 2022, reaching $4.5 trillion, about $13,493 per capita (CMS.gov). Given its importance, healthcare operations have received attention from both practitioners and scholars. A fundamental characteristic of healthcare, posing unique challenges to providers, is the uncertainty in care delivery — seminal work contends that uncertainty is the single most important factor affecting physician behaviour (Balsa et al., 2003; Wennberg, 1985).

[Step 2] From an OM perspective, the many sources of uncertainty in healthcare include demand-side, supply-side, and clinical uncertainty. We focus on mix uncertainty — the uncertainty stemming from variation in the complexity of care needs of patients in a hospital department — and volume uncertainty — the uncertainty due to variation in the volume of patients served. While an extensive body of research pertains to streamlining healthcare operations, little research directly addresses these uncertainties, their impact on the care-delivery process, and mitigating mechanisms.

[Step 3] Variation in patient volume may considerably disrupt process flows and operational planning. Yet unlike in a factory setting, individual patients have complex care needs — chronic and comorbid conditions of varied severity — making it impossible to simply standardise the response to uncertainty. These conditions interact to create considerable challenges in both the process and outcomes of care, for which no simple operational fix exists.

[Step 4] A growing body of work has examined the impact of demand volume on operational outcomes (Powell et al., 2012; Kuntz et al., 2015; Berry Jaeker and Tucker, 2017) and a rich body has examined the complexity of care needs (Kc and Terwiesch, 2011; Clark and Huckman, 2012; Thirumalai and Devaraj, 2024). However, while this research has focused on the level of complexity and volume, the variation in these dimensions and their impact on performance have heretofore not been examined.

[Step 5] Using a dataset of 830,853 patient discharges across 731 hospitals, we empirically examine the impact of mix and volume uncertainties on the care-delivery process — specifically the number of procedures (NPR) and length of stay (LOS) — and the mitigating effects of related focus and utilisation levels.

[Step 6] Our study builds on Ramasesh and Browning’s (2014) uncertainty framework and organisational information processing theory (OIPT) (Daft & Lengel, 1986; Galbraith, 1977). The R&B framework provides the conceptual foundation to study the sources and measurement of uncertainty, while OIPT offers an organisation-design perspective for understanding how hospitals might mitigate uncertainty effects — by reducing the need for information processing and/or increasing information-processing capabilities.

[Step 7] We use a large dataset of 830,853 patient discharges in 26 clinical departments across 731 hospitals in five U.S. Midwest states. We measure mix uncertainty as the mean-adjusted variation in the complexity of patients within a primary diagnostic category, and volume uncertainty as the variation in patient volume across time periods within a department.

[Step 8] Our analysis indicates that mix uncertainty triggers more procedures (higher NPR), suggestive of the information needs of providers, while volume uncertainty drives providers to reduce procedures (lower NPR) and cause delays (longer LOS), suggestive of bottlenecks from variable process flow. Breadth of expertise in relevant areas lowers patients’ LOS and reduces the need for more procedures. We find considerable heterogeneity in the effects of uncertainty across hospital types.

[Step 9] Theoretically, we extend the R&B framework and OIPT to the healthcare operations context, demonstrating that different types of demand-side uncertainty have distinct and opposing effects on care-delivery processes. Practically, our study highlights the role of related focus and utilisation management as operational levers for mitigating uncertainty — providing hospital administrators with actionable guidance for managing uncertain care environments.

Quick Diagnostic Checklist

Use this when reviewing any draft introduction — yours or a student’s.

# Check Zone
1 Does the opening paragraph include at least one concrete statistic? Motivation
2 Is the focal construct defined precisely within the first two paragraphs? Motivation
3 Is there a clear “however” or “yet” pivot that reveals a problem? Motivation
4 Does the paper explain why the obvious solution fails? (Step 3) Motivation
5 Is the literature review purposeful — building toward the gap, not exhaustive? Positioning
6 Is the gap precise enough for a reviewer to verify? Positioning
7 Are research questions stated explicitly? Positioning
8 Is the theoretical lens both named and justified for this setting? Contribution
9 Does the method preview include study type, data scale, and key design feature? Contribution
10 Are findings previewed at direction + mechanism level — not statistics? Contribution
11 Are contributions in a separate paragraph from findings? Contribution
12 Is the voice declarative? (“we find”, not “we hope to show”) Craft

Closing Thoughts

The nine-step framework is not a formula. Real introductions rarely have nine cleanly separated paragraphs — some steps are merged, some span several paragraphs, and the transitions between them are invisible by design. What the framework gives you is a diagnostic lens: when an introduction feels weak, you can usually trace it to a missing step, an underdeveloped zone, or a gap statement that is not precise enough.

The three questions worth asking about any introduction, at any stage of revision, are:

  1. After reading the first paragraph, does the reader care? (Motivation zone)
  2. After the gap statement, does the reader understand exactly what is missing? (Positioning zone)
  3. After reading the whole introduction, does the reader know what the paper does, how, and why it matters? (Contribution zone)

If the answer to all three is yes, you have a strong introduction.

This post is part of an ongoing series on academic writing for PhD students in Management Science, Operations Research, Operations Management, and related fields.

References

De La Torre Pacheco, S., Eftekhar, M., & Wu, C. (2023). Improving the quality of in-kind donations: A field experiment. Manufacturing & Service Operations Management, 25(5), 1677-1691.

Joshi, A. M., et al. (2024). Influenza vaccine contracts in developing nations. Decision Sciences, 55, 436–455.

KC, D. S., & Kim, S. (2022). The impact of universal healthcare on emergency department visits and patient health. Production and Operations Management, 31, 2167–2184.

Thirumalai, S., et al. (2024). Mix and volume uncertainty in healthcare operations. Journal of Operations Management.