Refurbished electronics have officially crossed a psychological—and commercial—threshold. Back Market closed 2025 with more than $3.5 billion in global gross merchandise value (GMV), up 32% year over year, marking one of the strongest signals yet that recommerce is no longer a side bet in consumer tech.

The company also delivered its largest Black Friday ever, posting 41% year-over-year growth during the period, further underscoring that refurbished devices are becoming a default consideration, not a last resort.

“These numbers matter because they show refurbished is no longer a fringe or experimental category,” said Thibaud Hug de Larauze, co-founder and CEO of Back Market. “It works as a disciplined business model, at scale, across markets.”

That statement carries weight in a consumer electronics industry long built on rapid upgrade cycles and planned obsolescence. Back Market’s results suggest the model is bending—and possibly breaking.

From Bargain Hunting to Deliberate Choice

Back Market’s 2025 growth wasn’t driven by a single breakout product or market. Instead, it reflects a deeper behavioral shift: repeat purchasing, broader category adoption, and rising trust in refurbished devices beyond smartphones.

Laptops, tablets, gaming consoles, and audio products all gained traction as consumers increasingly opted for last-generation, proven devices instead of the latest releases. The decision is less about compromise and more about value optimization—paying less for hardware that’s already validated in the real world.

This trend mirrors a broader change in how people evaluate technology. Performance improvements between hardware generations have slowed, while price increases have not. In that context, refurbished becomes a rational default rather than a budget fallback.

Europe Shows What “Mature” Looks Like

Europe remains Back Market’s most mature region—and its most instructive one.

In markets like France, where Back Market launched more than a decade ago, refurbished electronics are already normalized. Device replacement cycles are longer, consumer expectations are clearer, and trust infrastructure—warranties, grading standards, returns—has been fully established.

France delivered 35% EBITDA margins, and Back Market has now reached EBITDA break-even at the global level, a milestone that distinguishes it from many growth-first marketplaces still chasing profitability.

Europe, in effect, functions as a living case study for how the refurbished category behaves once it exits its adolescence. Growth becomes steadier, margins improve, and refurbished becomes less about thrift and more about system efficiency.

The U.S.: Earlier, Faster, and Catching Up

The United States tells a different story—one that may be even more consequential.

While still earlier in its adoption curve, the U.S. is now emerging as one of Back Market’s largest markets by GMV. In 2025, core U.S. test markets grew more than 40 percentage points faster than the company’s broader average, signaling the start of an acceleration phase.

That gap suggests pent-up demand rather than temporary tailwinds. American consumers have historically been slower to embrace refurbished electronics, often equating “new” with “better.” That assumption is eroding.

Back Market believes this shift is structural, not cyclical—and the company’s data increasingly supports that view.

Software Eats Hardware (Again)

One of the most compelling explanations for refurbished’s rise has little to do with sustainability messaging or inflation—and everything to do with how technology now evolves.

As AI systems, cloud platforms, and software-defined features become the primary engines of innovation, hardware is playing a different role. Devices are less about raw compute leaps and more about serving as durable access points to intelligence delivered elsewhere.

“The argument is not that devices matter less,” said Joy Howard, Chief Marketing Officer at Back Market. “They matter differently.”

When performance, security, and intelligence are delivered through the cloud, longer device lifespans become a feature, not a flaw. Frequent hardware replacement starts to look inefficient rather than aspirational.

This reframing quietly undermines the logic of annual upgrade cycles—and creates fertile ground for refurbished marketplaces.

Changing U.S. Consumer Behavior, by the Numbers

Back Market’s internal data shows that this shift is no longer theoretical in the U.S.:

• Older, proven models consistently outperform newer releases on the platform

• Non-smartphone categories now account for ~40% of U.S. GMV

• Nearly 50% of Gen Z consumers say their next smartphone will be refurbished

That last data point is particularly telling. Gen Z’s relationship with technology is pragmatic rather than status-driven. Performance, price, and longevity matter more than novelty—and refurbished fits neatly into that value system.

For brands and manufacturers, this raises uncomfortable questions about future demand curves.

Implications for the Device Economy

Back Market’s growth doesn’t exist in isolation. It reflects mounting pressure across the entire device ecosystem:

• OEMs face longer replacement cycles and weaker incentives for marginal hardware upgrades

• Carriers must rethink subsidy and trade-in strategies

• Retailers are forced to acknowledge recommerce as a parallel, not secondary, channel

• Policymakers increasingly view refurbishment as both an economic and environmental lever

The refurbished market challenges the idea that growth must come from producing more devices. Instead, it suggests growth can come from using existing devices better, longer, and more efficiently.

What’s Next: From Marketplace to Movement

Looking ahead to 2026, Back Market plans to deepen its engagement with industry leaders, partners, and policymakers, positioning itself not just as a marketplace but as a convener in the evolving device economy.

This includes participation in global forums like Mobile World Congress, where discussions will focus on how AI, cloud infrastructure, and durability are reshaping assumptions about hardware value.

“Globally, refurbished already works,” Hug de Larauze said. “The next chapter is about how quickly the U.S. catches up.”

If 2025’s numbers are any indication, that catch-up phase may already be underway.

The Bigger Takeaway

Back Market’s $3.5B year isn’t just a financial milestone—it’s a signal.

Refurbished electronics have crossed from alternative to inevitable. As innovation migrates up the stack and economic logic favors durability over novelty, recommerce is becoming a core pillar of the tech industry’s future.

For a sector long addicted to the new, that may be the most disruptive development of all.

Get in touch with our MarTech Experts.

Websites were once the center of digital marketing. Then came ad platforms, personalization tools, CDPs, and AI-driven targeting—while the website itself largely stayed frozen in time. Fibr AI wants to change that, and investors are betting it’s overdue.

The company announced $7.5 million in Seed funding, led by Accel, with participation from WillowTree Ventures and MVP Ventures, along with angel investors and advisors drawn from Fortune 100 operators. The funding positions Fibr AI to push forward what it calls the Agentic Web Experience Layer—a system designed to make websites adaptive, context-aware, and revenue-driving in real time.

It’s an ambitious pitch, but one that lands squarely in the middle of a growing problem for modern marketers: traffic has become intelligent, but websites haven’t.

The Website Is the Weakest Link in the Growth Stack

Marketing today is dynamic by default. Ads adjust in milliseconds. Recommendations personalize continuously. AI-driven discovery—through search engines, chat interfaces, and LLM-powered assistants—routes users to brands with increasingly specific intent.

Yet most websites still assume a one-size-fits-all experience.

Whether a visitor arrives from a branded search, a performance ad, a product comparison thread, or an AI-generated answer, they’re often dropped onto the same static page. The result is familiar: broken journeys, leaky funnels, wasted spend, and revenue left on the table.

Fibr AI was founded to close that gap.

“Marketing has become intelligent everywhere except the website,” said Ankur ‘AJ’ Goyal, CEO and co-founder of Fibr AI. “We’re building the Agentic Web, where every URL operates as a living experience system that understands context and responds in real time—for humans, cohorts, and even AI agents.”

In other words, the website stops being a destination and starts acting like software.

What Fibr AI Is Actually Building

At its core, Fibr AI reimagines high-traffic, consumer-facing websites as adaptive systems rather than static endpoints.

Instead of publishing a fixed experience and layering optimization tools around it, Fibr embeds AI agents directly into the web experience layer. These agents interpret context—traffic source, intent signals, user behavior, and increasingly, non-human actors like AI crawlers or agents—and dynamically adjust the experience in real time.

The goal is not just personalization in the traditional sense, but continuous optimization at the URL level, tailored to why someone (or something) arrived in the first place.

This approach reflects a broader shift in how discovery works:

• Users increasingly arrive with high intent, pre-informed by AI summaries and recommendations

• Conversion windows are shorter, not longer

• The cost of mismatch between intent and experience is rising

Fibr’s bet is that future-ready websites must behave less like content repositories and more like intelligent systems that react instantly.

Why Accel Is Leaning In

For Accel, the investment thesis hinges on timing as much as technology.

“Most websites today still run on infrastructure built years ago,” said Prayank Swaroop, Partner at Accel. “CMS platforms are effective at publishing content, but not at understanding context or adapting in real time.”

That limitation is becoming increasingly visible as conversational discovery takes hold. When users arrive from ChatGPT, LLM-driven ads, or AI-powered search, they’re often ready to act immediately. A static page isn’t just inefficient—it’s a conversion killer.

What differentiates Fibr, according to Accel, is its focus on embedding AI directly into the experience layer, rather than bolting it on through plugins, scripts, or fragmented tooling.

The implication is consolidation. What previously required a stack of personalization tools, experimentation platforms, analytics systems, agencies, and manual workflows could be handled within a single adaptive system.

That’s a compelling narrative for CMOs facing stack sprawl, rising costs, and pressure to show ROI.

From Personalization to Agency

Personalization isn’t new. What’s changing is agency.

Traditional website optimization relies on rules, segments, and experiments that marketers configure in advance. Fibr’s agentic approach flips that model. Instead of asking marketers to anticipate every scenario, AI agents continuously interpret context and decide how the experience should adapt.

This matters in a world where traffic sources are fragmenting and evolving faster than teams can keep up.

AI agents don’t just change how users search—they change how they arrive, what they expect, and how quickly they decide. A website that can’t respond in real time risks becoming the slowest part of the growth loop.

The Rise of Non-Human Visitors

One of the more forward-looking aspects of Fibr’s positioning is its acknowledgment that not all visitors are human anymore.

As AI agents increasingly browse, summarize, and recommend content on behalf of users, websites must serve dual audiences: people and machines. That requires clarity, adaptability, and structured intelligence at the experience layer—not just SEO markup and fast load times.

Fibr’s platform is designed to optimize for both, treating AI-driven discovery as a first-class channel rather than an afterthought.

This aligns with a broader trend across MarTech and SEO, where visibility increasingly depends on how well systems interpret and surface content—not just how users click on it.

A New Layer in the MarTech Stack

Fibr AI’s emergence points to a growing category gap.

CMS platforms manage content. CDPs manage data. Experimentation tools test variants. But none of them fully own the real-time experience layer—especially one that adapts autonomously based on live context.

That gap is becoming more painful as expectations rise. CMOs are being asked to deliver one-to-one experiences at scale, with fewer resources and tighter budgets. Static infrastructure simply doesn’t support that mandate.

By positioning itself as an Agentic Web Experience Layer, Fibr is effectively proposing a new foundation—one that sits between traffic acquisition and conversion, orchestrating experiences dynamically rather than statically.

Founders With a Clear Point of View

Founded by Ankur Goyal and Pritam Roy, Fibr AI reflects a clear frustration with how disconnected modern marketing systems have become from the website itself.

While tools around the site have evolved rapidly, the site has been treated as immutable. Fibr challenges that assumption, arguing that the website should be as intelligent and adaptive as the channels feeding it traffic.

That philosophy resonates at a moment when marketers are rethinking fundamentals, not just optimizing at the margins.

What This Means for Marketers and CMOs

If Fibr succeeds, it could reshape how teams think about web optimization entirely.

Instead of asking:

• “Which variant should we test?”

• “Which segment should see which page?”

Teams might ask:

• “What intent is arriving right now?”

• “How should the experience respond instantly?”

That’s a shift from configuration to orchestration—and from static journeys to adaptive systems.

It also reframes the website as an active revenue driver, not a passive conversion endpoint.

The Bigger Picture: The Agentic Web Is Coming

Fibr AI’s funding round isn’t just a startup milestone. It’s a signal.

As AI reshapes discovery, advertising, and decision-making, the web itself must evolve. Static pages built for generic users are increasingly misaligned with a world of contextual, intent-rich interactions.

The next generation of digital experiences will be agent-driven, adaptive, and responsive by default. Fibr AI is betting that the companies who modernize their websites accordingly will capture disproportionate value.

With Accel and a roster of experienced operators backing the vision, Fibr now has the runway to test whether the Agentic Web is not just possible—but inevitable.

Get in touch with our MarTech Experts.

Retail execution has never suffered from a lack of data. What it has lacked—until now—is confidence. GOcxm is betting that artificial intelligence, applied at the right layer, can close that gap.

The retail execution and consumer engagement platform announced it has acquired Emotive Technologies, the company behind Apex, an AI-powered consumer insights and decision intelligence platform. The move adds predictive modeling, simulation, and behavioral-science-trained AI directly into GOcxm’s Shopper Engagement and Execution Platform, signaling a clear push beyond reporting toward real-time decisioning.

In short: GOcxm wants brands to stop reacting to data after the fact—and start acting on it before execution even begins.

From Data-Rich to Decision-Ready

For global CPG brands, the challenge is no longer collecting data. Between loyalty programs, digital activations, promotions, and retail touchpoints, brands are awash in zero- and first-party signals. The harder problem is translating that information into clear, confident actions that drive performance at shelf and across channels.

“Brands have more data than ever but often struggle to translate it into confident action and workflows,” said Gary Kalk, CEO of GOcxm. “Apex adds a powerful intelligence layer that allows teams to model outcomes, validate strategies, and optimize performance using predictive intelligence and AI-driven decisioning.”

That framing reflects a broader industry shift. Retail is increasingly viewed not just as a point of sale, but as a continuous feedback loop—one that can inform strategy if insights are surfaced quickly enough and tied directly to execution.

What Apex Adds to the GOcxm Platform

Before the acquisition, GOcxm’s platform already covered a wide span of the retail lifecycle. Brands use it to:

• Activate consumers and shopper campaigns

• Capture zero- and first-party data

• Validate initiatives in-market

• Gather shopper sentiment

• Verify proof of purchase

Apex extends this foundation upstream.

The AI platform combines proprietary datasets with applied data science to simulate outcomes and generate predictive recommendations. Rather than telling teams what happened, Apex focuses on what is likely to happen—and what actions are most likely to work.

Embedded into GOcxm’s execution layer, that intelligence allows brands to design, test, measure, and continuously optimize initiatives inside a single system, rather than bouncing between analytics tools, dashboards, and agencies.

Predictive Intelligence Meets Retail Reality

What makes the acquisition notable is how tightly Apex is being integrated into execution, not just analytics.

In many CPG organizations, insights live in one place while activation happens somewhere else—often weeks later. That lag undermines the value of even the best data.

GOcxm’s strategy is to collapse that distance. By pairing predictive modeling with real-world retail execution, teams can:

• Validate strategies before rollout

• Model different scenarios and trade-offs

• Adjust programs during execution, not after

• Tie outcomes directly back to consumer behavior

That shift reflects a growing expectation among enterprise marketers: intelligence should be operational, not observational.

A Behavioral Science Angle

Another differentiator in Apex’s approach is its emphasis on behavioral-science-trained AI. Rather than relying solely on historical correlations, the platform is designed to account for how consumers actually make decisions—particularly in retail environments where context, incentives, and timing matter as much as messaging.

This is increasingly important as CPG brands face tighter margins and higher execution costs. Incremental gains now depend on making fewer, better decisions, not just more campaigns.

“Apex was built to help organizations make better decisions with greater confidence,” said Jordan Van Schyndel, Founder of Emotive Technologies. “By joining GOcxm, we’re able to extend our mission at a greater scale and help brands learn faster, prioritize actions, and turn insight into confident, measurable decisions.”

Consolidation in Retail Tech Continues

The acquisition also fits into a broader consolidation trend across retail and shopper marketing technology.

Brands are pushing back against fragmented stacks that separate insights, activation, and measurement. The demand is shifting toward unified platforms that can close the loop—from data capture to execution to performance validation.

By absorbing Apex, GOcxm is positioning itself less as a point solution and more as an intelligence-driven operating system for retail performance. That’s a competitive stance in a market where differentiation increasingly comes from how well platforms drive action, not how many metrics they expose.

What This Means for CPG Brands

For CPG teams under pressure to justify spend and prove ROI, the implications are significant.

Instead of relying on static reports or post-mortem analysis, brands gain the ability to:

• Predict outcomes before committing budget

• Optimize execution while campaigns are live

• Treat retail touchpoints as a continuous insight engine

• Align teams around a single source of truth

That alignment is especially valuable as retailers demand more precision and accountability from brand partners.

Looking Ahead: Intelligence as the Core of Execution

GOcxm says Apex’s technology and team will be fully integrated into its product roadmap, with continued investment in AI and data science. The goal is to ensure that insight, activation, and measurement operate as one connected system, rather than as loosely coupled functions.

The larger signal is clear: retail execution platforms are evolving beyond task management and reporting into decision intelligence systems.

As data volumes grow and execution windows shrink, brands that can simulate, predict, and adapt in real time will hold the advantage. With this acquisition, GOcxm is making a decisive move to compete in that future.

Get in touch with our MarTech Experts.

Performance advertising is entering a decisively algorithmic phase—and Cognitiv is riding the center of that wave. The advanced performance partner reported strong business momentum in 2025, led by 388% growth in ContextGPT™ and accelerating adoption of its deep learning advertising platform across brands and agencies.

The results reflect more than a good year. They signal a structural shift in how performance marketing is executed as marketers contend with signal loss, media fragmentation, and rising expectations for measurable outcomes. In that environment, Cognitiv’s pitch—custom deep learning algorithms operating in real time—has found a receptive market.

Industry forecasts help explain why. Algorithm-driven spending is expected to account for 71.6% of global ad spend in 2026, rising to 76% by 2028, placing Cognitiv squarely in the fastest-growing segment of digital advertising.

From Self-Serve Optimization to Deep Learning Systems

For years, performance marketing has relied on incremental optimization inside self-serve platforms. That model is starting to show strain.

“Our growth reflects a clear shift in how brands and agencies are approaching performance marketing,” said Jeremy Fain, CEO of Cognitiv. “Advertisers are moving beyond self-serve optimization toward automated deep learning systems that can understand context, interpret intent, and drive outcomes at scale.”

Cognitiv’s differentiation lies in how those systems are deployed. Rather than running containerized models that process data after the fact, Cognitiv co-locates its infrastructure with leading SSPs, allowing algorithms to make bidding decisions in milliseconds.

According to the company, that architecture delivers up to 10 times the computing power of many competing solutions, enabling real-time optimization of metrics like viewability, video completion rate (VCR), and clicks—while impressions are still being auctioned.

In an environment where milliseconds increasingly determine outcomes, that speed advantage matters.

The Curated Contextual Era Takes Shape

At the heart of Cognitiv’s growth is what it calls the Curated Contextual Era—a move away from static audience segments toward real-time, algorithmic decisioning driven by content and context.

This strategy centers on ContextGPT, which Cognitiv positions as the industry’s first real-time contextual targeting intelligence platform. Unlike legacy contextual tools that rely on keyword matching or rigid taxonomies, ContextGPT uses deep learning and advanced language models to interpret content with human-like nuance.

The platform allows advertisers to activate performance without cookies, pixels, or user IDs, a capability that has become table stakes as privacy restrictions and signal loss reshape digital advertising.

In 2025, ContextGPT customer adoption grew more than 44%, fueled by continued product innovation. New features such as Interactive Audience Exploration and an upgraded Relevance Engine delivered up to 40% greater accuracy in connecting brands with custom-defined audiences.

The result is advertising that is not only more privacy-resilient, but often more relevant—because it’s grounded in real-time intent rather than inferred identity.

Performance Gains Drive Customer Expansion

Cognitiv’s momentum isn’t limited to adoption metrics. Customer performance improved meaningfully over the course of the year.

In 2025, the company saw:

• A 67% increase in new clients year over year

• A 29% improvement in average ROAS from the first half to the second half of the year

Those gains translated into deeper customer relationships. Existing clients expanded spend and launched more campaigns as results improved, particularly across verticals like CPG, pharma, and travel, where contextual relevance and brand suitability are critical.

This performance-driven expansion highlights a broader trend: advertisers are increasingly willing to consolidate spend with partners that can prove incremental value, not just incremental reach.

Talent and Leadership Signal a Scaling Phase

Cognitiv’s growth in 2025 wasn’t purely technical. The company also invested heavily in organizational readiness as demand increased.

Over the year, Cognitiv added new talent and promoted dozens of team members across engineering, data science, product, customer success, and go-to-market functions—a sign the company is building for scale rather than chasing short-term growth.

A notable milestone was the promotion of Justine Frostad to Chief Marketing Officer, marking the company’s first CMO appointment. The move signals a more deliberate investment in brand, narrative, and executive visibility as Cognitiv enters a new phase of maturity.

The company also added four new VP and SVP leaders, alongside multiple senior-level promotions, reinforcing a stated commitment to developing talent from within.

At the board level, Cognitiv strengthened its strategic perspective with the addition of Michael Kassan, a longtime marketing and media industry veteran—an appointment that underscores the company’s long-term ambitions.

Why This Matters for the Ad Industry

Cognitiv’s results arrive at a pivotal moment for digital advertising.

As identifiers disappear and platforms fragment, performance advantage is shifting toward infrastructure, algorithms, and execution speed, not audience ownership. Contextual intelligence—once viewed as a fallback—has become a primary strategy, especially when paired with deep learning that can operate in real time.

Cognitiv’s growth suggests that advertisers are no longer experimenting at the edges. They are committing budget to systems that can autonomously interpret context, optimize bids, and adapt continuously.

If forecasts around algorithm-driven spend hold, companies built natively around deep learning—not retrofitted onto legacy stacks—will shape the next phase of performance marketing.

The Takeaway

Cognitiv’s 2025 momentum tells a clear story: performance advertising is becoming a machine-led discipline, and contextual intelligence is no longer optional.

With ContextGPT growing nearly fourfold, improving ROAS, and a platform designed for speed at scale, Cognitiv is positioning itself not just as a tool provider, but as an infrastructure partner for the algorithmic future of advertising.

As the industry moves toward 2026 and beyond, the question may not be whether deep learning drives performance marketing—but which platforms are fast enough, smart enough, and close enough to the exchange to matter.

Get in touch with our MarTech Experts.

As tracking weakens and retargeting becomes less reliable, opt-in has quietly become one of the most valuable moments in marketing. Intuit is making that case forcefully with a new global report that suggests list building isn’t the finish line—it’s where the real work begins.

Intuit Inc. (Nasdaq: INTU) has released The Art of the Opt-In: Why List Building is Only the Beginning, a research report developed by Intuit Mailchimp in partnership with Ascend2. Drawing on insights from thousands of marketers and consumers across the U.S., Canada, the U.K., and Australia/New Zealand, the report maps a widening gap between what brands ask for at opt-in—and what consumers are actually willing to give.

The conclusion is blunt: brands are still optimizing for volume, while consumers are optimizing for trust.

Opt-In Is the Last Clean Signal

In a marketing ecosystem shaped by privacy regulation, signal loss, and platform fragmentation, opt-in has become something rare—a moment of explicit permission.

“As tracking and re-targeting become more complex, the opt-in stands out as one of the few moments when a brand can earn a direct relationship,” said Matt Cimino, product manager at Intuit Mailchimp. “Sign-up is the first signal that someone is willing to engage, and what a brand does in that moment sets expectations for everything that comes next.”

That framing reflects a broader shift underway in MarTech. As third-party data erodes, first-party relationships are no longer just valuable—they’re fragile. The inbox and SMS channel are increasingly guarded spaces, and consumers are far more selective about who they let in.

Marketers Have Lists—But Not Confidence

The data reveals a paradox at the heart of modern marketing execution.

Nearly all marketers maintain email and SMS lists. Yet:

• Fewer than one-third consider their lists “very high quality”

• Only 8% report conversion rates above 20%

• Just 21% have fully automated email and SMS campaigns

• Only one-third feel very confident about which channels drive opt-ins

In other words, most brands are collecting contacts—but few are convinced those contacts will actually convert.

That lack of confidence has consequences. Without clarity on source, intent, or preference, marketers struggle to personalize responsibly. The result is more messaging, not better messaging—fueling consumer fatigue.

Consumers Are Paying Less Attention, Not More

On the other side of the relationship, consumers are noticing the increase in marketing messages—but that doesn’t mean it’s working.

According to the report:

• Only 40% of consumers say they’re paying more attention to marketing emails and texts

• About 25% say they’re tuning these channels out more than a year ago

For those who remain engaged, the expectations are clear:

• 56% want content that genuinely adds value

• 40% want messaging frequency that doesn’t feel like spam

The problem isn’t the channel. It’s relevance.

Asking Too Much, Too Soon

One of the report’s most telling findings highlights how early missteps at opt-in can erode trust before a relationship even starts.

For example:

• 65% of brands ask for a phone number in popup forms

• Only 28% of consumers are willing to provide it

That gap signals a trust mismatch. Brands are optimizing for downstream value, while consumers are still deciding whether the relationship is worth starting.

“Most opt-ins come up short because they’re created only thinking about what the business needs, not what the customer actually wants,” Cimino explained.

The data suggests brands perform better when they focus on high-intent moments rather than intrusive interruptions:

• 50% of consumers are more likely to opt in after browsing

• 39% are more likely during checkout

Timing, context, and restraint matter more than aggressive data capture.

Trust Is Generational—and Visual

The report also reveals sharp generational differences in how trust is formed.

• 39% of Gen Z assume brands will follow privacy laws

• That number drops to just 19% among Baby Boomers

For younger consumers, trust is often immediate and design-driven:

• 43% of Gen Z say clean, simple design makes them more comfortable completing opt-in forms

• Only 29% of Boomers+ say the same

This suggests that for Gen Z, trust is communicated visually and experientially, not through fine print or lengthy explanations. For older audiences, skepticism remains higher—and harder to overcome.

Automation Separates List Builders From Leaders

One of the report’s clearest signals is the role of automation in closing the relevance gap.

Brands that consider their contact lists best-in-class—dubbed “List Quality Leaders”—are:

• 3x more likely to have fully automated email and SMS programs

• More likely to run structured welcome series (64% vs. 53%)

• More likely to deploy cross-sell and upsell flows (45% vs. 36%)

Automation, in this context, isn’t about sending more messages. It’s about responding appropriately, consistently, and at scale.

Omnichannel Orchestration Beats Channel Optimization

The research also challenges channel-by-channel thinking.

Brands with highly aligned omnichannel messaging and timing report significantly higher value across nearly every channel:

• Organic social: 62% vs. 43%

• Paid social: 56% vs. 40%

• Emerging channels like generative engines: 10% vs. 5%

The implication is clear: when channels reinforce each other, each one performs better. When they operate in silos, even strong channels underperform.

This is especially relevant as AI-driven discovery reshapes how consumers find and evaluate brands. Opt-in no longer sits downstream of awareness—it’s part of a continuous, cross-channel conversation.

Data Isn’t the Problem—Fragmentation Is

Despite access to more data than ever, marketers struggle to turn signals into relevance.

Only:

• 30% use preference or frequency data

• 29% use browsing behavior

These are among the strongest predictors of engagement, yet they remain underutilized—largely because data lives in disconnected systems.

“Relevance comes from clarity, not volume,” said Diana Williams, Vice President of Product, Intuit Mailchimp. “When data is fragmented, even the best intentions fall short.”

Mailchimp’s strategy, as outlined in the report, centers on unifying behavioral signals, automations, and omnichannel insights so marketers can act with confidence—not guesswork.

Why This Matters Now

The timing of the report is notable.

As privacy regulations tighten and AI reshapes discovery, permission-based marketing is becoming the foundation of sustainable growth. Opt-in is no longer a mechanical step—it’s a value exchange that sets the tone for the entire relationship.

Brands that treat opt-in as a checkbox will continue to see list fatigue and declining engagement. Those that treat it as a moment of trust—earned through relevance, restraint, and clarity—stand to build relationships that compound over time.

The Takeaway

The Art of the Opt-In makes a simple but uncomfortable point: most brands are still designing opt-ins for themselves, not for their customers.

In a world where attention is scarce and permission is precious, list building is no longer the goal. Earning the right to stay relevant is.

For marketers navigating signal loss, AI disruption, and rising consumer skepticism, the opt-in moment may be the most strategic lever left—and one too often mishandled.

Get in touch with our MarTech Experts.

Connected TV has officially taken the lead—and it’s rewriting the rules of media buying. As streaming surpassed broadcast and cable combined for the first time in 2025, advertisers are facing a tougher challenge than ever: how to reach audiences whose attention is fragmented across apps, screens, and formats.

Nexxen (NASDAQ: NEXN) and independent agency H/L believe they have an answer. By pairing Nexxen’s demand-side platform (DSP) and advanced data insights with H/L’s performance-driven media strategy, the partners report up to a 14x lift in conversion outcomes for clients across multiple verticals. The results highlight a broader shift underway in CTV—from reach-first buying to signal-backed, outcome-oriented execution.

Streaming Wins, Attention Fragments

The backdrop to this partnership is a historic inflection point. In 2025, streaming accounted for 44.8% of total TV usage, overtaking broadcast and cable combined for the first time. At the same time, free ad-supported streaming TV (FAST) channels are surging, with monthly viewership up roughly 12% year over year and session lengths continuing to grow.

For advertisers, that growth is a double-edged sword. CTV now offers access to highly engaged audiences—but those audiences are scattered across platforms, apps, and content types. Add in second-screen behavior, and the old playbook of broad, undifferentiated media buys starts to fall apart.

The implication is clear: attention, not impressions, is the scarce resource.

From CPMs to Cost-Per-Unique Reach

H/L’s approach with Nexxen reflects this reality. Instead of optimizing around traditional CPMs, the agency has shifted focus to cost-per-unique reach, blended frequency management by app, and optimization tied directly to down-funnel business outcomes.

“At H/L, we’ve strategically embraced this evolving CTV landscape,” said Jeremy Cobb, Vice President of Digital Platforms at H/L. “By tapping into Nexxen’s advanced insights, we’ve crafted a strategy that blends premium inventories and custom placements for initial viewer activation, with cost-effective long-tail.”

That balance matters. Premium placements help capture attention early, while long-tail inventory extends reach efficiently—without oversaturating the same viewers. The result is frequency discipline in a medium where overexposure is increasingly easy and increasingly expensive.

Measurable Results, Not Just Promising Signals

The payoff has been tangible. For H/L’s clients in sectors like automotive and insurance, these signal-backed strategies have driven up to fourteen times higher conversion outcomes compared with more traditional CTV programs.

Crucially, those gains weren’t self-attributed. They were validated by Marketing Mix Modeling (MMM) partners, giving advertisers greater confidence that CTV isn’t just delivering awareness—but measurable performance.

This kind of validation is becoming table stakes as CFOs scrutinize media spend more closely and demand proof that CTV can compete with, or complement, lower-funnel digital channels.

Nexxen’s Bet: Know Who’s Actually Watching

Nexxen’s role in this equation centers on insight density and execution speed. Its DSP combines buying capabilities with supply-side intelligence, allowing advertisers to understand not just where ads appear, but what content viewers are actively engaged with.

“Navigating the new CTV landscape requires knowing who’s truly paying attention and managing ad frequency with precision,” said Kara Puccinelli, Chief Customer Officer at Nexxen. “That’s why agencies are leaning into advanced measurement, balancing high-attention placements with cost-efficient reach.”

Rather than treating CTV inventory as interchangeable, Nexxen emphasizes contextual and engagement signals—an approach that aligns with how advertisers now evaluate performance in a multi-screen world.

Publishers Lean In With Smarter Supply

The push toward signal-backed buying isn’t limited to DSPs and agencies. Media owners are also adapting, surfacing richer data and more flexible formats to help buyers transact more intelligently.

Platforms like Philo and DIRECTV Advertising are working with Nexxen to expose high-value inventory—particularly in premium and live content—paired with granular audience and contextual signals.

“CTV remains one of the most impactful ways to connect with audiences, particularly while watching content they’re passionate about,” said Aulden Kaye Yi, Head of Advertising Partnerships at Philo. “Working with Nexxen allows us to surface our inventory with granular audience and contextual signaling.”

That capability becomes even more important as live sports continues its migration to CTV, a trend DIRECTV Advertising is actively capitalizing on.

“As live sports programming continues to shift to CTV, Nexxen allows us to surface and package this high-value ad inventory with contextual and audience data signals,” said Edmund Jules, Senior Director of Ad Sales Partnerships at DIRECTV Advertising.

Why This Matters for Advertisers Now

The Nexxen–H/L results point to a broader industry lesson: CTV performance hinges less on scale and more on signal quality.

As streaming dominates viewing time, advertisers can no longer afford to treat CTV as a blunt instrument. Success depends on:

• Understanding where attention is concentrated

• Managing frequency with discipline across apps

• Optimizing toward outcomes, not just exposure

In that sense, CTV is starting to look less like traditional TV and more like performance media—with all the accountability that implies.

The Bigger Takeaway

CTV’s growth story is no longer about adoption—it’s about execution. As the market matures, the winners will be those who can navigate fragmentation with precision, not those who simply buy more impressions.

The partnership between Nexxen and H/L shows what’s possible when advanced data, flexible buying, and outcome-driven strategy converge. In a world where attention is splintered and budgets are scrutinized, a reported 14x lift in conversions is more than a case study—it’s a signpost for where CTV buying is headed next.

Get in touch with our MarTech Experts.

Artificial intelligence is now deeply embedded in everyday marketing workflows—but new research suggests accuracy hasn’t kept pace with adoption.

According to NP Digital’s AI Hallucinations and Accuracy Report, AI-generated errors are not only common, they’re increasingly slipping into live campaigns. Nearly half of marketers (47.1%) encounter AI inaccuracies several times per week, and 36.5% report that hallucinated or incorrect AI content has already gone public.

The findings underscore a growing tension in modern marketing: AI delivers speed and scale, but without sufficient oversight, that efficiency can introduce serious brand risk.

AI Errors Are Frequent—and Time-Consuming

The report combines two data sources:

• An accuracy analysis of 600 prompts tested across six major large language models (LLMs), including ChatGPT, Claude, and Gemini

• A survey of 565 U.S.-based digital marketers

Together, the data paints a picture of widespread friction between AI output and real-world accuracy.

More than 70% of marketers say they spend one to five hours each week fact-checking AI-generated content, eroding some of the productivity gains AI is supposed to deliver. Despite this effort, errors still escape into production.

“AI has become an incredible tool to accelerate efficiencies, but speed without accuracy creates real risk,” said Chad Gilbert, Vice President of Content at NP Digital. “What makes AI hallucinations especially dangerous is that many of them look believable at first glance.”

When AI Mistakes Go Live

Among marketers who reported publishing inaccurate AI-generated content, the most common issues included:

• False or fabricated facts

• Broken or nonexistent citations

• Brand-unsafe or misleading language

These errors often appear polished and confident, making them harder to detect without careful review. Once published, they can damage credibility, confuse audiences, or expose brands to compliance and reputational risks.

Yet despite these dangers, 23% of marketers say they are comfortable using AI output without human review, a gap between awareness and behavior that the report flags as particularly concerning.

No Model Is Error-Free

NP Digital’s accuracy testing also evaluated how different LLMs perform under scrutiny.

• ChatGPT delivered the highest rate of fully correct responses at 59.7%

• No model consistently avoided hallucinations

• Error rates increased sharply for:

  • Multi-part questions

  • Niche or specialized topics

  • Real-time or time-sensitive queries

The most common hallucination types across all models included:

• Omissions

• Outdated information

• Fabrication

• Misclassification

Crucially, these errors were often delivered with high confidence—making them more persuasive and more dangerous.

Where AI Breaks Down Most Often

The report found that AI struggles most with tasks requiring precision, structure, or technical rigor, including:

• HTML or schema creation

• Full long-form content development

• Reporting and data-driven summaries

These are also the areas where marketers are most likely to trust AI to “just handle it,” increasing the likelihood of mistakes slipping through.

The Real Takeaway: AI Needs Guardrails

The data points to a clear conclusion: AI works best as an assistant, not an authority.

Strong prompts, defined review processes, and human oversight consistently reduce risk. With no single LLM emerging as reliably accurate across use cases, marketers can’t solve the hallucination problem by switching tools alone.

Instead, the report reinforces a mindset shift:

• Treat AI output as a draft, not a final answer

• Match AI tasks to its strengths, not its hype

• Keep humans accountable for what goes live

As AI becomes standard infrastructure in marketing, accuracy—not speed—may be the new competitive advantage.

Get in touch with our MarTech Experts.

The Shelf, a performance-first influencer marketing agency, has earned multiple top industry honors in 2025, reinforcing its position at the intersection of creator-led storytelling and measurable growth.

The agency was named Best Activation in Support of Sustainability at the Adweek Experiential Awards for its work with Natura, and received finalist recognition at the Drum Awards in Retail & Consumer Products for its campaign with Weekday, as well as for Micro-Influencer Campaigns with Gregory Mountain Products. The Shelf also earned a Platinum MarCom Award for its work with Self Financial.

Together, these accolades reflect more than creative excellence. They point to a structural shift underway in performance marketing—one in which creative has become the primary driver of optimization, learning, and scale.

When Targeting Fades, Creative Becomes the Signal

Major platform changes over the past year have fundamentally altered how performance is generated. With automated campaign structures expanding and granular targeting controls shrinking, platforms increasingly rely on creative behavior signals—how content is watched, saved, shared, and engaged with—to determine delivery and scale.

Updates like Meta’s Andromeda have accelerated this shift. As deterministic targeting weakens, platforms can only optimize against what creative reveals in real time. In this environment, creative is no longer just a message—it is the feedback loop.

Many brands have responded by increasing content volume, assuming scale would offset the loss of control. But results have often remained inconsistent. According to The Shelf, the issue isn’t quantity—it’s structure.

Without systems designed to test, interpret, and compound creative insight quickly, content fails to generate durable performance signals.

A System Built for Signal, Not Just Output

The Shelf’s operating model was designed for this reality before platforms forced the shift.

Rather than starting with creators or deliverables, the agency begins with social listening and market intelligence, mapping interest clusters that reveal how audiences actually think, speak, and make decisions. These insights shape creator selection, briefing, and narrative design—aligning brands with creators embedded in relevant, interest-based communities.

From there, creative variations are tested rapidly to surface early engagement signals—such as saves, shares, and watch time—that indicate downstream performance potential. Winning narratives are scaled, while underperforming ones are iterated or retired.

Creative is treated as a source of intelligence, not just an output.

Repeatable Performance at Scale

That systemized approach has delivered consistent results across categories and regions.

Over the past year, The Shelf reports:

• Sustained ROAS across multiple U.S. and European markets for fashion and retail brands

• More than 40 million organic impressions driven by creator-led campaigns at retail scale

• Performance lifts through iterative optimization, including ROAS up to 23x during peak seasonal windows, with efficiency sustained in the weeks that followed

Early engagement signals proved to be reliable predictors of downstream efficiency, allowing platforms to learn faster and scale more effectively.

“These platform changes didn’t make marketing harder, they made weak systems visible,” said Atul Singh, Co-Founder and CEO of The Shelf. “When targeting control fades, the only thing platforms can optimize against is signal. If your creative can’t generate meaningful signal, the platform can’t learn from you. We built our model to turn creative into intelligence long before platforms forced that shift.”

Structure Becomes the Advantage

As performance marketing becomes creative-led by design, the competitive advantage is no longer who produces the most content—but who learns the fastest.

In an environment where platforms optimize behavior, the brands that win will be those with systems built to:

• Identify the right creators

• Pair them with insight-driven narratives

• Interpret performance signals in motion

In that future, scale follows structure—not the other way around.

Get in touch with our MarTech Experts.