When Impulse Stops Being Automatic: An Operator Brief on Marketing's Shifting Baseline

I. The Unnoticed Shift

The cart stays full. The checkout never happens.

Traffic holds steady. Engagement looks healthy. The email gets opened. The product page loads. The item gets added. Then nothing. Not abandoned in the traditional sense: no error message, no competitive comparison, no price shock. The session just ends. The customer leaves. They don't come back for it.

This has been happening since late 2022. Not everywhere. Not all the time. But consistently enough that the pattern shows up in quarterly reviews, post-mortems, and the private conversations operators have when the dashboards stop making sense. The metrics that used to correlate with revenue no longer do. The tactics that used to recover a stalled funnel don't. The promotional calendar that used to create urgency generates clicks but not purchases.

Operators noticed this first as a feeling. Something off in the conversion data. A disconnect between effort and outcome. Then they began running the explanations: consumer confidence declining for five consecutive months. Tariff uncertainty. Post-pandemic normalization. A "choiceful" customer being more selective with discretionary spend. Promotional intensity in the marketplace making it harder to break through. Inflation forcing trade-downs. The usual catalog of external pressures.

Each explanation has data behind it. Each is partially true. Each is insufficient.

Between 2023 and 2025, major CPG brands reduced their product assortments by double digits. Unilever cut 17% of its SKUs. Coca-Cola discontinued flavors after only months on shelves, citing "changing taste preferences." The official reason was always cost efficiency, margin protection, or operational streamlining. But brands don't quietly shrink their product lines and reduce package sizes when demand is strong. They do it when consumption patterns shift beneath the level of articulated strategy.

At the same time, digital platforms began adding friction to experiences that had been engineered for frictionlessness. DoorDash and Uber Eats moved tipping prompts away from checkout. Instagram introduced "Take a Break" reminders. TikTok added screen time limits. Amazon started charging for returns at certain locations. Retailers shortened return windows from 60 days to 30, then to 14. These weren't presented as responses to changing user behavior. They were framed as policy updates, feature experiments, operational adjustments.

No press releases connected these changes. No earnings calls drew the line between reduced SKUs, added interface friction, and declining conversion rates. The adjustments happened in parallel, across categories, without coordination. Brands don't simultaneously reduce options, add friction, and shrink portions unless the behavioral substrate has changed.

Something in the consumer has changed. Not their preferences. Not their awareness. Not their intent. The gap between seeing and wanting has widened. The path from interest to purchase has lengthened. The triggers that used to collapse that gap (urgency, scarcity, social proof, ease) are firing without effect.

By mid-2025, the language in earnings calls had calcified into a stable set of narratives. Consumers described as "cautious," "value-seeking," "trading down while still splurging." Macro conditions cited as the primary headwind. Internal execution acknowledged as a secondary factor. The framework never allowed for the possibility that the consumer's internal response to stimulus had fundamentally shifted.

Add to cart still happens, but it no longer commits the buyer to anything.

II. What Marketing Quietly Assumed About Desire

The techniques worked long before anyone agreed on why they worked.

Countdown timers. Limited stock warnings. One-click checkout. Free trials that roll into paid plans. These patterns spread because they produced results. They became standard because repeating them felt safer than questioning them.

What went largely unexamined was what had to be true for those techniques to function.

They required desire to be interruptible. They assumed urgency could be manufactured. That hesitation was temporary. That friction, not judgment, was the primary obstacle between interest and purchase.

The direct response model that emerged in the late twentieth century operated on this logic. Attention was interrupted. Scarcity was implied. Value was stacked. The path forward was simplified. The goal was completion.

Late-night television made this visible. Viewers at the end of the day. Low stimulation. Fewer competing demands. The environment mattered as much as the message. By the time the call to action arrived, the effort required to keep evaluating outweighed the effort required to comply.

When commerce moved online, the same assumptions traveled with it.

Checkout flows were shortened. Forms were compressed. Steps were removed. Any pause was treated as a leak. Any delay as a failure. The funnel was not designed to accommodate reconsideration. It was designed to minimize it.

Scarcity was rarely literal. Timers reset. Inventory warnings persisted. Time pressure existed to shorten the space between noticing and deciding. Not to inform, but to hurry.

Subscriptions refined the pattern further. Entry was made effortless. Exit was made inconvenient. Inaction became the default state of continued payment. The system did not depend on ongoing satisfaction. It depended on tolerance for interruption.

None of this was framed as pressure. It was described as efficiency. As experience. As removing friction in service of the user.

The commercial system did not assume a reflective buyer. It assumed a constrained one. A buyer for whom delay was costly, attention was fragile, and stopping required more effort than continuing.

III. GLP-1 Adoption as Clarifier

A portion of the population stopped responding to urgency the way they used to.

The intrusive thoughts about food quieted. The compulsion to keep scrolling weakened. Shopping carts filled and sat. Subscriptions were canceled. Pantries no longer needed to be fully stocked to feel safe.

These reports appeared first in patient forums and social media in 2023. By 2024, they were common enough to quantify. By 2025, they were showing up in aggregate conversion data.

GLP-1 receptor agonists, prescribed for weight management, reduced appetite. They also appeared to reduce wanting more broadly. Users described it as volume adjustment. The craving didn't disappear. It stopped being loud.

This created contrast. The portion of consumers whose reward response remained elevated continued converting at historical rates. The portion whose response had normalized required more time, more proof, more actual value to close. The tactics that worked by triggering urgency in elevated states began failing in normalized ones.

The system had been calibrated to a specific baseline. That baseline was moving in a subset large enough to register in quarterly metrics. Operators saw softness. Hesitation. Resistance. Consumers who used to convert in one session now required three. Funnels that closed in minutes now stalled for days.

The urgency that used to compress decision-making was being ignored. The friction removal that used to guarantee completion was no longer sufficient. The promotional intensity that drove immediate action was generating interest without commitment.

By late 2025, GLP-1 adoption in North America had reached levels where population-level effects became visible in category data. Not everywhere. Not uniformly. But in enough segments that the aggregate conversion baseline was detectably lower than 2022.

Operators running attribution models saw the pattern but misread it. Traffic intent looked fine. Engagement looked healthy. What had changed was the threshold at which interest became action. That threshold had risen. The metrics designed to predict conversion at the old threshold continued reporting success while conversion declined.

IV. Why Trusted Metrics No Longer Correlate

Traffic metrics, engagement rates, email opens, product page views, cart additions: these remain stable or improve while purchases decline. The problem is not data quality. The problem is that these metrics were designed in an environment where interest reliably converted to action. That correlation has weakened.

High engagement, low conversion

A social media account grows from 3,500 to 39,000 followers with 9.6% engagement. Strong saves, shares, consistent link clicks. By every platform metric, the account is successful. But sales don't follow. The content entertains or educates. It doesn't create purchase intent. Engagement measures whether people interact with content. It doesn't measure whether they want to buy the product.

This mismatch becomes invisible if engagement is the primary KPI. The disconnect persists until someone manually checks whether engaged followers are actually converting.

Viral traffic, inverse conversion

Traffic surges after content goes viral. Conversions plummet. On quiet days with minimal traffic, sales appear randomly. Viral traffic is curiosity-driven, high volume but low intent. These visitors browse and leave. Sales on quiet days come from high-intent sources: direct searches, referrals from people actively seeking solutions.

Stable traffic, declining conversion

Web traffic holds steady year-over-year. Cart additions look healthy. But fewer carts complete checkout. This pattern appeared across alcohol e-commerce in 2024 and 2025. Operators initially treated it as a brand-specific execution problem. Broader analysis revealed it as category-wide. Consumers were browsing more but buying less. Adding to cart but not completing purchase. Expressing interest without commitment.

The funnel performs as designed. What changed is the consumer's willingness to complete at the moment of decision. Historical conversion rate assumptions no longer hold.

Subscriber volume masking churn

Subscription services report stable or growing subscriber counts while revenue per subscriber declines. Gross adds look healthy. Net retention deteriorates. By late 2024, video-on-demand churn reached 44%, an all-time high. Apple TV+ had 45 million subscribers and over $1 billion in annual losses due to low engagement.

The subscriber count metric reports scale. The retention metric reports that subscribers aren't staying long enough or engaging deeply enough to justify content costs. If monthly recurring revenue growth is the success metric, the business looks healthy. If lifetime value and engagement depth are the metrics, the model is broken.

Email opens without clicks

Open rates remain solid. Click-through rates flatten. Subject lines drive opens. Opens are healthy. The problem is that email content or calls-to-action aren't compelling action. People open the email (curiosity is satisfied) but don't click through to the site.

This becomes invisible if open rate is the primary metric.

What this means for measurement

When the gap widens (when attention no longer implies near-term action) these metrics decouple from revenue. They continue measuring the top of the funnel accurately. But top-of-funnel health no longer predicts bottom-of-funnel outcomes.

V. Where Impulse-Dependent Models Break

Certain business models depend structurally on impulse converting quickly and reliably. When impulse dampens, the unit economics degrade in predictable ways.

Subscription retention

The subscription model assumed inertia would keep users paying even when engagement dropped. The effort required to cancel (navigating settings, confirming choices, resisting retention offers) was supposed to exceed the mental cost of continued payment. This worked when users lacked the bandwidth to actively manage subscriptions.

When impulse control improves, tolerance for unused subscriptions drops. Users become more willing to cancel. Fintech tools make subscription management easier. Churn accelerates. The model depended on behavioral friction, not product value. When behavioral friction stops working, churn reflects actual engagement levels.

Net revenue retention falls. Customer acquisition costs rise because more touches are required to convert a prospect. The math that made the model viable (acquire cheaply, retain through inertia, monetize through habitual usage) inverts. CAC rises while LTV falls.

Urgency-based conversion

Countdown timers, limited stock warnings, "last chance" messaging: these tactics trigger loss aversion. They work by making the fear of missing out stronger than the impulse to delay and evaluate. When impulse response dampens, loss aversion weakens. The urgency still registers. It doesn't compel immediate action.

Click-through rates remain acceptable (the urgency gets attention). Conversion rates decline (the urgency doesn't close). Attribution models credit the tactic with engagement but miss that engagement no longer predicts purchase.

Operators see urgency tactics "working" in dashboards while revenue from those tactics drops. The tactic performs as designed at the attention layer. It fails at the conversion layer. The system keeps running urgency because the metrics say it works.

Novelty cycling

SKU proliferation and seasonal launches depend on novelty triggering a dopamine response strong enough to justify the purchase. Limited editions, collaborations, new flavors: these command a premium when consumers are actively seeking the stimulation of something new.

When reward response to novelty weakens, the conversion premium disappears. Consumers default to core products or delay purchase entirely. The operational complexity of managing multiple SKUs stops justifying itself. Brands that expanded assortments to capture novelty-seeking behavior begin cutting SKUs when that behavior recedes.

Unilever's 17% SKU reduction and Coca-Cola's flavor discontinuations weren't cost optimization. They were responses to declining novelty-driven conversion. The consumer wasn't chasing new variants. The brand was carrying inventory for a behavior that had weakened.

Friction removal as conversion driver

One-click checkout, saved payment methods, streamlined forms: these remove decision points between intent and purchase. They work by minimizing the opportunity for reconsideration. The fewer steps, the fewer chances to abandon.

When deliberation increases, friction removal stops guaranteeing completion. The consumer still appreciates the streamlined flow. They use the pause to evaluate whether they actually need the item. A frictionless checkout that used to convert in 30 seconds now gives the user 30 seconds to reconsider.

Attribution model breakdown

Last-touch attribution credits the final interaction before conversion. This worked when urgency tactics reliably closed sales. A countdown timer or promotional email triggered immediate purchase. Last-touch attribution credited that tactic.

When urgency stops converting immediately, last-touch attribution becomes unreliable. The customer sees the urgency, clicks through, adds to cart, then leaves. They return days later through a different channel and complete purchase. Last-touch credits the final session. The urgency that initiated consideration goes uncredited.

Multi-touch attribution attempts to solve this but introduces new problems. If the conversion window lengthens from hours to days, which touches actually influenced the decision? The urgency email that started consideration? The retargeting ad three days later? The organic search when they were finally ready to buy?

Attribution models built for short consideration cycles and impulse-driven conversion stop functioning reliably when consideration lengthens and impulse weakens.

Habitual overconsumption

Pantry loading, bulk buying, supersizing: these behaviors were built into demand forecasting. Consumers purchased more than immediately needed, driven by the same reward circuitry that governed impulse behavior generally. This created predictable revenue premiums in average order value and purchase frequency.

When impulse to overstock weakens, baskets shrink. Purchase frequency drops. The consumer buys what they need rather than what feels safe to have.

VI. What Quiet Adjustments Reveal

While earnings calls blamed macro conditions, operational changes throughout 2023 and 2025 revealed that brands were already adapting to reduced impulse without saying so.

SKU rationalization as demand signal

Unilever cut 17% of SKUs. Coca-Cola discontinued new flavors after months. The stated reason was operational efficiency. The operational reality was that novelty-driven incremental sales had weakened to the point where SKU complexity couldn't justify itself.

When novelty-seeking behavior weakens, the long tail of variant SKUs becomes economically inefficient. The adjustment happened quietly because publicly acknowledging reduced novelty response would require explaining why consumer behavior changed.

Shrinkflation as portion adjustment

Package sizes shrank by an average of 16.2% across major grocery brands in 2024. Some products reduced by over 30%. The explanation was input cost management. The behavioral reality was that consumers were accepting smaller portions without abandoning brands.

Brands don't reduce package sizes unless they've tested that demand is elastic enough to tolerate it. Historically, portion reductions triggered competitive losses. When consumers began accepting smaller sizes, brands adapted.

Added friction as tolerance test

DoorDash and Uber Eats moved tipping to post-delivery. Amazon added return fees at select locations. Retailers shortened return windows. Social platforms added usage interruption features. These were framed as policy updates or user experience improvements.

Operationally, they were friction tolerance tests. Brands don't add friction unless they've observed that customer tolerance for friction has increased.

Instagram and TikTok adding "Take a Break" features is particularly revealing. Platforms that spent a decade engineering for maximum engagement began adding tools that interrupt usage. This doesn't happen unless engagement patterns have shifted enough that interruption no longer threatens retention.

Pattern of accommodation

The pattern across adjustments: accommodation to less novelty-seeking, smaller consumption volumes, higher friction tolerance, and more deliberate purchasing.

The gap between public narratives and private actions widened throughout 2024 and 2025. Earnings calls blamed macro factors and execution. Operational decisions accommodated a fundamentally different consumer baseline.

VII. Planning for 2026

Marketing teams should plan for a lower-impulse baseline in 2026. This is not a temporary condition. GLP-1 adoption continues accelerating. The cohort of consumers with normalized reward response is growing, not shrinking.

Expect longer decision cycles

Impulse can no longer be assumed as a bridge between interest and purchase. Historical time-to-conversion assumptions need adjustment. What closed in one session may now require three. What closed in 24 hours may now take 72. Funnels calibrated to impulse-driven speed will underperform.

If a campaign drives strong top-of-funnel engagement but conversion lags, the campaign may still be effective. The conversion is just happening later. Attribution windows need to lengthen. Success metrics need to account for delayed conversion.

Urgency tactics decay faster

Countdown timers and scarcity messaging will continue generating clicks. Conversion lift from these tactics will decline. The tactics appear to work in engagement metrics while failing in revenue metrics.

Teams should test urgency tactics against longer-term value delivery. If a product's actual value can't justify purchase without urgency, the urgency is compensating for weak product-market fit. When urgency stops compensating, the weakness becomes visible.

Frictionless flows no longer outperform reliably

Removing friction from checkout improves user experience. It no longer guarantees conversion lift. When consumers are deliberating more, a streamlined checkout gives them a comfortable environment to reconsider rather than a rushed path to completion.

Friction removal should be evaluated for user experience benefit, not conversion lift. The conversion lift that used to justify optimization investment may no longer materialize.

Retention math requires actual engagement

Subscription models cannot depend on inertia. Users will cancel unused subscriptions more readily. Churn will reflect actual engagement rather than cancellation friction. LTV assumptions based on historical retention rates need downward revision.

New subscriber acquisition should be evaluated against realistic retention curves, not historical ones. If churn is accelerating, CAC must drop proportionally or the model becomes unprofitable. Teams should assume higher churn and plan accordingly rather than treating current churn as temporary.

Dashboards will continue looking healthy

Top-of-funnel metrics will remain stable or improve while bottom-of-funnel conversion declines. This will persist. Teams need secondary metrics that track conversion speed, completion rates, and time-to-purchase. Engagement alone no longer predicts revenue.

Attribution models built for impulse-driven conversion will misreport campaign effectiveness. Multi-touch attribution with extended windows becomes necessary. Last-touch attribution will increasingly credit channels that happened to be present when deliberation finally concluded, not channels that actually influenced the decision.

What not to do

Do not interpret declining conversion as creative failure or targeting problems if top-of-funnel engagement is healthy. Do not optimize urgency tactics harder when they stop working. The substrate has changed, not the execution. Do not assume that current softness is temporary and that historical conversion rates will return.

Do not build 2026 forecasts on 2023 conversion assumptions. The baseline has shifted. Models calibrated to impulse-driven behavior will systematically overestimate performance.

What to test

Test whether longer consideration cycles are accommodated by current funnel design. Test whether value delivery can sustain conversion without urgency. Test whether actual product differentiation is strong enough to close sales in a higher-deliberation environment.

Test attribution models with extended windows. Test whether engagement metrics still correlate with revenue at historical rates. Test whether friction removal still delivers conversion lift or just improves experience.

If this shift shows up in your data and you want a second set of eyes, you can reach me here.