Behavioral Design Under Pressure

This page has interactive footnotes: Hover over pink words to get a little more information. Like this! (On tablet and mobile, just tap to view.)

When crisis hits, the behavioral design principles that typically drive conversion can become ineffective and counterproductive. I learned this firsthand while co-leading the campaign for an online assessment tool in the early months of COVID-19, when our business insurance clients were facing unprecedented stress. What started as a straightforward application of proven ideas in UX psychology turned into a masterclass in context sensitivity—and a reminder that effective marketing requires deep empathy for users’ mental states.

Crisis Context & Strategic Imperative

By July 2020, COVID-19 shutdowns had been devastating businesses across the U.S. for months. As a growth marketer at a business insurance carrier specializing in small businesses, I witnessed our clients drowning in existential uncertainty about coverage, liability, and business survival. State regulations changed weekly, federal relief programs launched in confusing waves, and much of our clientele—typically employing fewer than nine¹ workers—operated under psychological stress far exceeding normal market conditions.

Sales and marketing leadership assembled a cross-functional task force to create an online assessment tool to help overwhelmed entrepreneurs understand their insurance options during the crisis. We adapted logic from an existing tool that had successfully helped policyholders evaluate coverage strength, creating a new 20-question version that would deliver personalized, pandemic-specific coverage recommendations.

We marketed the tool through paid ads, organic social posts, and emails to policyholders, directing them to a landing page that shared statistics about the COVID crisis. This page required users to input their policy number before starting the assessment, at which point a blend of standard questions were then asked for context: annual revenue, industry type, employee count, location, and property ownership status.

Deeper into the assessment, more specific questions were asked to help identify COVID’s potential impact on the policyholder’s business, as well as possible solutions based on their specific circumstances. “Did you make a profit last year?” was followed by, “Do you think you'll make a profit this year?” Next: “If no, will COVID be the main reason why?” Our curiosity was pointed, necessarily: “Do you have sufficient cash flow to meet payroll and rent obligations over the next 3-6 months?” “Have you defaulted or are you at risk of defaulting on loans or lease agreements?”

Framework: Three Core UX Principles

Our methodology centered on three behavioral design principles that had driven successful insurance campaigns:

• Loss aversion leveraged people’s tendency to feel losses more acutely than equivalent gains. We highlighted what users stood to lose without proper coverage: valuable insights in the short term, potentially their entire business in the long term.

• Authority-driven guidance tapped into users’ willingness to follow expert advice, positioning our century-plus of insurance experience as definitive guidance during uncertain times.

• Urgency-based engagement used crisis language to create immediate action, emphasizing that businesses under extreme pressure needed tailored recommendations quickly to maximize survival odds.

Original System Architecture

The assessment featured clean, intentionally minimal visual design—simple black and white styling with ample white space to create an uncluttered experience for those who engaged with it (whom we did anticipate were likely stressed). Content was personalized based on their state and business profile, with technical infrastructure that could branch them down different paths depending on their answers. All messaging positioned our recommendations as authoritative guidance.

Our methodology assumed linear user psychology during crisis: present authoritative risk information → user acknowledges gaps → urgency drives action → customer upgrades coverage or engages with our Customer Success or Sales teams. Each component reinforced the others in a systematic cascade toward deeper engagement, exactly as these psychological principles had worked in pre-pandemic campaigns.

We briefed Sales and Customer Success about the initiative, asking them to report any interactions, positive or negative, that referenced this new COVID-related tool we were marketing; beyond digital analytics alone, this would help the task force keeps its finger on the pulse of how the assessment was actually being received.

Where the Framework Collapsed

Our main structural failure was a fundamental misunderstanding of the psychological state of our policyholders. We treated crisis-affected business owners who were taking the assessment as simply “very concerned” versions of their normal selves, but they weren’t their normal selves at all. They were stressed out in an unprecedented way—juggling supply chain disruptions, revenue freefall, and complex relief applications simultaneously—which meant our architecture built on precedent was, essentially, doomed to fail.

This miscalculation caused each framework component to backfire.

Loss Aversion Backfire

• Design Intent: Using the Von Restorff effect, we made coverage gaps visually obvious through red callouts. When users from specific states and industries revealed gaps—like Texas event planners needing Coverage Plan B instead of Plan A—we highlighted these gaps prominently.

• Actual Result: Gap callouts amplified anxiety rather than motivating action.

• Evidence: We observed statistically significant hard exits on pages where gap callouts appeared most frequently, and Customer Success reported an influx of distressed calls from customers referencing these callouts—panicked about coverage issues the tool had flagged.

• Root Cause: We assumed policyholders would have the cognitive bandwidth to process risk information constructively—not recognizing that stressed decision-makers exhibit “tunneling” behavior that narrows rather than expands their processing capacity.

Authority-Driven Guidance Backfire

• Design Intent: Marketing emphasized our 100-plus years of experience and deep regulatory knowledge, positioning the assessment as essential given it was possible that people’s livelihoods were at stake.

• Actual Result: Heavy-handed authority positioning created unrealistic expectations about our ability to solve pandemic-related business problems beyond insurance.

• Evidence: Sales reported an increase in the number of users who asked questions far beyond our scope—cash flow management, supply chain alternatives, employee retention strategies—an issue that had never manifested significantly until the tool launched.

• Root Cause: Authority signals during crisis can often generate dependency rather than empowerment, which can cause users to expect comprehensive guidance—total, panacean support—that extends outside the core competency.

Urgency-Based Engagement Backfire

• Design Intent: We deployed time pressure through crisis language, citing devastating statistics like 68% of food-industry businesses completely stopping employee payments during the pandemic,² both in LinkedIn ads and customer communications that promoted the tool.

• Actual Result: We triggered decision paralysis in already overwhelmed users.

• Evidence: Contrary to our expectations of improved performance, average exit rates increased significantly compared to the original assessment we had adopted our logic from, with 95% and 85% confidence intervals respectively for reduced email and phone engagement. Exit spikes occurred at four particularly stress-inducing questions about business survival timelines, layoffs, cash flow, and loan default risk.
  
  The broader context made this outcome more understandable in hindsight. As Psychology Today reported, “The COVID-19 pandemic caused not only millions of hospitalizations and over a million deaths in the United States, but [also] a widespread and enduring mental health crisis.”³ While entrepreneurs are typically less prone to anxiety than the general population,⁴ everyone has a breaking point. Our assessment could trigger dozens of warnings that a business wasn’t properly situated to survive the pandemic, creating decision paralysis even in users who might normally handle complex decisions well.

• Root Cause: Fostering a sense of urgency in users who are already experiencing acute stress will usually decrease their cognitive bandwidth, which makes it harder for them to focus their attention and harder for us to inspire the desired action.

Diagnosis: Context Blindness

Our fundamental methodological error was context blindness—applying behavioral design principles calibrated for stable market conditions to users who were experiencing unprecedented disruption. Restaurant owners facing permanent closure, retail stores navigating capacity restrictions, and service businesses losing 80% of their revenue weren’t exhibiting standard insurance buying behavior.

The framework would likely have worked well for baseline customers, but it became actively counterproductive for cognitively depleted users. Cognitive depletion, as research shows, drains mental batteries and impacts judgment, impulse management, and information processing⁵—exactly the opposite conditions our framework required.

Rebuilt Architecture: Adaptive Design for Crisis

Learning from these failures, our task force rebuilt the assessment around adaptive design principles that acknowledged varying levels of cognitive depletion.

New Diagnostic Questions

We added five psychological screening questions at the start of the experience to evaluate people’s general crisis management state:

1. “Right now, how would you describe your general outlook on the world?”
   Options: Optimistic/Neutral/Pessimistic

2. “If your business were to fail, how much would it impact your personal or family finances?”
   Options: Minimal/Moderate/Severe

3. “Is this your primary source of income, or more of a side business?”
   Options: Primary/Supplemental/Side hustle

4. “Would you describe your industry as deeply impacted by COVID?”
   Options: Yes/Somewhat/Not much

5. “How confident do you feel in your ability to navigate your business through this crisis?”
   Options: Very/Somewhat/Not confident

There was a downside to this rebuild. Five new questions increased the total number of questions from 20 to 25, which meant users would need more time to complete the assessment, and longer digital experiences typically see more user drop-off. Also, we had to acknowledge an important limitation of our own: We were marketers, not psychologists (despite one team member’s psychology background), which we had an ethical responsibility to remember at all times.

But the potential advantage, as we explained to leadership, was singular: a dramatically improved user experience. We hoped this would fix the misfires of the first go-to-market attempt (which had driven statistically significant decreases in engagement) and help customers navigate the tool more productively while learning about coverage gaps relevant to their specific situation.

New Adaptive Messaging Paths

Based on the diagnostic responses, the assessment now automatically categorized each user into one of three personas:

• Survival Mode Sam (high cognitive depletion): Simplified explanations, minimal detail, maximum support messaging

• Pivot Mode Paul (moderate): Balanced detail level, strategic focus, moderate complexity

• Recovery Mode Rachel (low): Comprehensive explanations, detailed strategies, full complexity

The logic behind this segmentation reflected our growing understanding of crisis psychology. Pessimism might signal emotional exhaustion and cognitive depletion. A business owner who felt their financial success or failure was tied to the vicissitudes of a global crisis might, understandably, have more decision stress or depletion risk—as might someone whose business was more closely tied to their identity, their family’s wellbeing, or an industry that was hugely impacted.

These three messaging paths signaled our increased sensitivity to the fact that different people were psychologically managing their way through the COVID crisis in different ways, with different potential changes to their normal decision-making capacity.

Designing for the Majority

The average policyholder who engaged with the assessment would be categorized as a Survival Mode Sam based on their answers: an entrepreneur who might normally handle complex decisions very well but was now operating with limited cognitive capacity due to unprecedented, unplanned stressors that could prove fatal to their business. Most of our architectural changes focused on making the experience work—and feel—much better for this majority of users.

Each persona received the same core questions but with messaging calibrated to their processing capacity.

For example, where Survival Mode Sam might see “Right now, how are you feeling about managing your business?” with simple options like “I’m handling it” or “I’m really struggling,” Recovery Mode Rachel would encounter “How would you rate your current confidence level in adapting to or navigating this unprecedented economic disruption?” with more detailed response options. This represented our first systematic attempt at crisis-sensitive user experience design.

Framework Corrections

• Loss Aversion Refinement: We implemented two key changes here. First, we removed the assessment’s ability to reveal coverage gaps on individual pages, as we noticed this was causing hard exits—likely because this information was distressing for policyholders in real time. Instead, we saved all gap coverage feedback for the very end of the experience.
  
  Second, while it might seem small, we moved away from the Von Restorff effect. Originally, we had highlighted potential coverage gaps in red text to make the important advice unmissable. But we realized this was coming at a cost to users’ psychology. While research proves that red can encourage users to take desired action faster, it also shows that in Western culture, red is associated with danger or failure.⁶ We put our policyholders first by changing gap notifications to black text—the same as normal text in the assessment. This color choice might spur less immediate action, but it encouraged a healthier, less anxiety-inducing connection with the information.

• Authority Repositioning: We retained the mention of our credentials but softened the language we used to promote the tool, now explicitly positioning that it was just one option of many that we offered, including phone consultations and self-paced reading materials about coverage possibilities for small business owners during the ‘new normal.’ Also, we were careful to disclaim—not in legalese, but in the tone of an approachable specialist—that our expertise, while real and renowned, was in insurance and insurance only, and we could not aid our customers in exploring more general business strategy.

• Urgency De-escalation: We removed statistics about economic devastation from marketing promotion and the tool’s landing page, shifting our word choice from “Take action today to ensure your business is protected” to “When you’re ready to expand your protection...” It was OK if a policyholder who owned a small business chose not to engage with our COVID-specialized tool. It was OK if they ultimately did take it, but did so in their own time. Anyone investigating the tool was already sufficiently aware of the ongoing crisis; our own catastrophizing was unneeded and unhelpful when the goal wasn’t to reinforce, Think again about distressing things are right now, but to inspire proactive discovery about steps someone could take to insure their future.

New Features for User Support

Beyond correcting our original framework missteps, we also used this upgrade opportunity to implement all-new features that prioritized user wellbeing.

• Persistent Save Functionality: We added a “save and continue later” button at the bottom of every page of the assessment. This let customers leave and return to the experience at any point in the future if it became too challenging or stressful for them to finish it in one session. When users clicked this button, they were redirected to a page that told them it was safe to close the window: not only helping them piecemeal their way through the assessment if they needed to, but also enabling us to decrease the number of hard exits from the tool.

• Graduated Follow-up Sequences: We installed new workflows that would send deep-linked email reminders to everyone who used this new “save and continue later” feature. After three days of leaving the tool and not returning, all customers received a reminder. After eight days, only Pivot Mode Pauls and Recovery Mode Rachels—who had moderate and low cognitive depletion, respectively—received follow-ups (i.e., protecting Survival Mode Sams, who were the most psychologically impacted by the pandemic, from additional pressure). And after 14 days, only Recovery Mode Rachels received one final prompt.

We understood that both the save functionality and the persona-specific follow-ups could theoretically reduce overall completion rates. But this was about putting users first. We believed we’d get better, more accurate completions while doing right by our policyholders, and we hoped the save feature would reduce hard exits by providing a clear, guilt-free way to pause the experience.

Marketing Implications

This experience fundamentally changed how I think about behavioral design. I learned that standard conversion optimization assumes people are operating at their normal cognitive capacity—but when users are dealing with major disruption, whether it’s economic upheaval, personal challenges, or industry-wide chaos, they simply aren’t functioning as they normally would. I developed deep empathy for the small business owners who engaged with our tool: real people watching their life’s work hang in the balance, many facing potential financial ruin through no fault of their own.

As marketers, we’d felt the pressure to create something as fast as we could to not only help our policyholders but also serve our organizational goals. Our instinct and remit is to cut through the noise of a crisis and find a way to connect with people, which we absolutely did. But looking back, I can see the trap we fell into: We had a proven assessment framework and, to capitalize on an urgent market moment, we duplicated and fine-tuned it for crisis conditions. But we didn’t fine-tune it enough—and more importantly, we fine-tuned it for speed and market timing, not for the actual mindsets of the users who would be engaging with it.

COVID was unprecedented, so our approach to marketing during it needed to be unprecedented too.

Through the backfires came learnings, and those forced me to reevaluate assumptions I’d held about behavioral psychology. Authority can create dependency instead of confidence, in the right (or wrong, depending on how you’re looking at it) context. Urgency can trigger paralysis instead of action. Loss aversion can amplify anxiety instead of motivating change. And all of the personas that marketing departments painstakingly create usually do not account for these bizarro truths which tend to emerge in crisis conditions.

What does this mean for growth marketers like myself? We can’t move so fast to meet the market moment that we meet it poorly—in which case, it would probably be better not to try to meet it at all. We have to develop not only normative state personas but also crisis state personas. It’s important to understand the context of the moment we’re trying to engage with, and it’s important to understand when something is right for normative circumstances and when it’s not. Frameworks should be able to adapt. A crisis plan should be in place. And it’s important to remember that the same methods that drive conversion during stable periods may require complete inversion during crisis: emphasizing reassurance over urgency, simplification over sophistication, patience over pressure.

Marketing is built on optimization and acceleration. Speed is everything. Every marketer, usually through great personal pain and baptismal fire, knows this. But this project reminded me that sometimes the most effective strategy is slowing down to meet users where they are, psychologically speaking. That human-centered approach, ultimately, can better serve both the business outcomes and the user experience.

Works Cited

¹ “A look at small businesses in the U.S.” Pew Research Center.
² “Impact of the Coronavirus Pandemic on Businesses and Employees by Industry.” U.S. Bureau of Labor Statistics.
³ “COVID‑19 and Americans’ Mental Health.” Psychology Today.
⁴ “Personality Traits of Entrepreneurs: A Review of Recent Literature.” Harvard Business School.
⁵ “Cognitive Depletion.” Sustainability Directory.
⁶ “The color red attracts attention in an emotional context.” National Library of Medicine.