Consumer behavior, particularly spending behavior, is becoming more difficult to predict. Not long ago, professional marketers used ‘personas’ to characterize entire audience segments, assuming groups moved through the funnel in similar ways.
This assumption is breaking down. Recent consumer data from Ipsos highlights that widening financial gaps across different markets and generations are creating a highly fractured landscape— some economies are climbing, while others are contracting, often in the same month.
Financial confidence is also diverging across generational divides, not just borders. A younger, more cautious segment in one market can be more financially confident than an older, historically stable segment in another. Geography alone no longer predicts customer behavior.
Why Does One Retargeting Algorithm Fail Today’s Fractured Buyer?
For any business selling online, the ‘ideal customer’ no longer behaves or spends uniformly. Yet many brands still run retargeting through a single engine – one algorithm, reengaging every past visitor the same way.
The problem is that one technology and one underlying dataset produce one model for predicting intent. When you apply a lone model to a customer base that is anything but uniform, it will always underperform for the segments it wasn’t built to see.
Myth 1: A Single Retargeting Partner Is Simpler To Manage
Digital marketers are usually encouraged to simplify their processes, for example, by consolidating their tech stack or reducing the number of vendors they’re working with. Similarly, a common argument for using a single retargeting provider is that it’s ‘simpler to manage.’
The reality is that while adding another partner requires initial work around setup, measurement, and governance, once that infrastructure is in place, the ongoing demand on a marketing team is small.
Performance marketers need to ask whether the convenience of working with one retargeting partner is worth limiting their vision and intelligence in understanding the customer journey.
Myth 2: Multiple Vendors Cannibalize Each Other
Another common argument for sticking with one retargeting solution is that multiple retargeting engines overlap and result in waste. This is far less of an issue than many think: different-in-kind algorithms find distinct micro-segments and moments rather than competing for the same impression. A shopper flagged as high-intent by a rules-based model and one flagged by a Deep Learning model are often not the same shopper.
Take a returning visitor who viewed a product three times but never converted. A static, rules-based engine may read that repetition as strong intent and bid accordingly. A Deep Learning model, trained on a wider set of behavioral signals, might instead flag a different visitor entirely – one who browsed only once but showed the pattern of signals that more reliably precedes a purchase.
Both are legitimate reads of intent. Neither is wrong. But a single engine can return only one of them.
Myth 3: The Incremental Gain Isn’t Worth The Complexity
Recent research found that AI tools are extending, not shortening, the path to purchase: 42% of US consumers and 33% of non-US consumers say AI now increases the time it takes them to reach a final decision, by surfacing more brands and alternatives to weigh up. More than half of consumers globally say AI is good at surfacing brands they didn’t previously know.
That longer window changes what retargeting needs to do. A shopper who takes two weeks to decide, rather than two days, is being exposed to more competing offers throughout – which means their intent is shifting, not static. A lone engine, trained once on one dataset, reads that shopper the same way on day one and day twelve. A second, differently trained model is more likely to catch the shift as it happens, rather than repeating an early read that’s since gone stale.
Acknowledging The Fractured Market
This is not an argument for adding vendors for their own sake. It’s an argument that a fractured spending landscape – different confidence levels, different urgency, different willingness to act, often within the same customer list – needs more than one lens to read it accurately.
Newer, Deep Learning-driven approaches pick up real-time signals that older, more static intent models systematically underweight, particularly among hesitant or price-sensitive buyers.
An overdependence on a single-source critical supplier wouldn’t make sense in any other business function: it’s time for retargeting to be treated with the same logic. In a world where no two customers or countries spend the same way, a single model was never built to see the whole picture.
Jaysen Gillespie is a seasoned product and analytics leader with over 15 years in Adtech and data science. As VP of Global Product Marketing and Analytics at RTB House, he translates complex insights into clear, actionable narratives marketers can use. Whether guiding global teams or speaking on stage, Jaysen focuses on making performance results immediately relevant, helping brands connect data to smarter decisions, measurable impact, and real business outcomes.
Related stories:
- Making The Most Out Of Marketing Budgets With The Power Of Robots
- Apollo Introduces New AI-Powered GTM Platform for Revenue Teams
