It’s a familiar moment. You shop for shoes on one site, and within minutes, they follow you across the internet like a well-meaning stalker. Are these ads helpful? Annoying? Creepy? Or – as one study suggests – maybe even kind of cool?

According to research from Marigold, 84% of consumers say it's "cool" to receive a personalized birthday offer, and 80% appreciate recommendations based on their past purchases. But when the personalization crosses into unexpected territory – like being targeted based on a hyper-specific physical location or a site visit you'd rather forget – the cool factor quickly chills into discomfort.

This shifting perception raises an important challenge for marketers: how to navigate the line between personalization that delights and personalization that disturbs. This tension – between relevance and surveillance – is especially acute when it comes to visual experiences.

Understanding the shift: from creep factor to conditional acceptance

For years, the phrase "creepy marketing" was practically synonymous with retargeting. Consumers didn’t know how brands acquired their data, and the lack of transparency made every well-timed ad feel like it came with an invisible camera crew.

But something is changing. The same Marigold study suggests that attitudes are evolving. People are more tolerant of personalized content – when it feels appropriate and beneficial. Personalized birthday discounts? Yes, please. A promo for a brand you've never heard of after passing their storefront once? Not so much.

As Chris Porteous on Entrepreneur points out, the key to overcoming this "personalization paradox" lies in consent and transparency. People aren’t inherently against personalization. They just want to know what data you're using, why you're using it, and how it improves their experience.

The personalization paradox

The paradox is this: 72% of consumers engage only with highly personalized content. And yet 86% are concerned about how brands handle their data. That’s a tightrope walk for any marketing team.

Porteous lays out five stages of personalization, from broad-based messages ("business to all") to the holy grail: "business to one." But the closer you get to individual targeting, the higher the risk of creeping people out – especially if you can't explain where the data came from.

So how do you keep your visual personalization sharp – without triggering the digital equivalent of an eyebrow raise?

How to navigate the threshold: five visual personalization principles

Visual content plays a massive role in perception. A smart visual strategy can feel intuitive and even delightful. A clumsy one? Instant distrust. Here are five principles to get it right.

1. Design for context, not intrusion
Use non-invasive data like device type, time of day, or general location to tailor your visuals. This form of "ambient personalization" avoids the need for deeply personal data while still delivering relevant imagery.

2. Respect the line between audience and individual
Don’t chase hyper-specificity if you can’t do it transparently. Focus instead on segment-based optimization – serving images designed for runners, not just one particular runner.

3. Build in consent and transparency
Be upfront about what you personalize and why. If visuals shift based on behavior, say so. Transparency builds trust, and trust is the only real foundation for long-term engagement.

4. Let the user lead
Offer options for image preferences, like themes or display styles. Allow them to opt in to personalization features. Visual customization can feel empowering – as long as it doesn’t feel forced.

5. Optimize for delight, not just conversion
The best visual experiences spark joy. Birthday offers are a hit because they feel thoughtful, not targeted. When optimizing images, ask: Will this visual surprise and delight, or will it make someone squint at their webcam in suspicion?

What smart image optimization enables

You don’t need to compromise user trust to create relevance. Smart image optimization lets you adapt visuals in real time based on contextual signals, not creepy data trails.

It enables:

• Instant format and resolution adjustments for different devices and screen sizes
• A/B testing and performance feedback for visual assets
• Dynamic visuals that respond to time of day or campaign without tracking individuals

This technology puts the control back in marketers' hands, allowing them to design visuals that are engaging and performant – without crossing the line.

Final thoughts: the new relevance playbook

Consumers aren’t allergic to personalization. They’re allergic to being manipulated. Visuals, done well, can feel like magic. Done poorly, they feel like surveillance. Marketers must understand the threshold, not just for legal compliance, but for brand credibility.

Want to see smart image optimization in action? Watch our webinar, How Greetings Island Enhanced User Experience with Image Optimization. You’ll learn how they accelerated page load speeds, deployed dynamic visuals, and reduced dev time by 40 hours – all while improving the user experience.