Why We Took a Contrarian Approach to E-Commerce

Retail is sophisticated. We’ve got more SaaS tools than ever, more dashboards, more automation. And yet CAC keeps rising, cart abandonment remains stubbornly high and margins are under pressure.

Maybe it’s time to stop playing the same game harder. Maybe it’s time for a contrarian playbook.

At behamics, we stopped chasing shoppers after they’d left the store, stopped layering personalisation on emails, and stopped plugging Gen AI trained on the entire internet into our customer journeys.

Instead, we built our strategy around one simple truth: people drive retail. Technology should amplify human behaviour, not replace it.

Here’s how we rethought three areas of ecommerce.

1. Cart Abandonment: Stop Chasing, Start Guiding

Retargeting is the common playbook when a customer abandons their cart. Blast customers with emails, ads, promotions, anything to lure them back.

We flipped that thinking. Why wait until they leave? Why not help them decide while they’re still in the store?

Our Nudge product does that. Grounded in behavioural science, it operates in real-time to help customers make confident shopping decisions.

Today’s shoppers are bombarded with social feeds, search results, chatbots, offers. Attention is fragmented with research showing consumers make subconscious decisions in milliseconds. And while these decisions happen fast, they’re not random. Every choice is shaped by intent, emotion and mental shortcuts like loss aversion, cognitive dissonance or social proof.

Our people, with academic degrees in behavioural science, design the Nudges. AI is then deployed to analyse human behaviour at a scale and speed that humans can’t match, allowing Nudge to intervene in real-time 24/7. behamics’ AI models, trained on causality, make sense of the complex dynamics behind purchasing behaviour, uncovering not just what customers do, but why.

The result: a 10% conversion uplift. 3% reduction in returns rates due to customers making better decisions. And a shopping journey customers want to come back to.

2. Personalisation: Stop Selling, Start Supporting

For years, “personalisation” has meant a first name in an email or an ad for the shoes you already decided not to buy.

But what about the site itself? Most websites remain hard coded. AB tested to perfection perhaps, but ultimately designed for the common denominator. The problem with this is it leaves revenue on the table.

At behamics we looked beyond marketing and merchandising, developing 1:1 personalisation at discovery (PLP), the product page (PDP), and checkout.

Why does this matter? Each visit brings slightly different behaviours. How your customer behaves on site today can differ to how they showed up last week or last month. Customers need different flags, signals and reassurances to feel confident in a purchasing decision.

This is personalisation where it helps most: in the decision-making moment, not in the inbox.

3. AI: Stop Layering, Start Building

The industry rush has been to layer LLMs on top of e-commerce. But here’s the problem: those models are trained on the entire internet, not your customers’ unique behaviours.

At behamics, we went contrarian again. We built our own models that focus first on your customer’s actions and your ecommerce metrics, and then added AI that operates in real-time and reduces mundane tasks for teams. 

That’s how Diagnostic catches revenue losses caused by UX issues as they happen. How we fix site speed before a customer bounces. How User Insights takes hundreds of recordings of actual customer behaviour and lets ecommerce teams “talk” to that data through an LLM layer — analysing real shopper behaviour at a scale that humans can’t process manually.

And with Organic, we flipped SEO content strategy too. Instead of using LLMs to generate SEO content that machines ignore, we first model how Google, Gemini, and ChatGPT rank your site. Then we create content that both humans and machines recognise.

This is AI built based on your customers and site, not borrowed from the internet at large.

People First, Machines Second

We named our company behamics because we’re interested in the interplay between human behaviour and tech.

At the end of the day, it’s not dashboards or algorithms that win customers. It’s people — making decisions, building trust, and shaping experiences. Technology just needs to help them do it better.

---

Brooke Eichhorn is Country Manager, Australia & NZ, at behamics.

behamics is a behavioural-science and causal-AI platform helping retailers like Douglas, Nespresso, and Volvo optimise conversion, increase traffic and reduce ad costs.

Drop Brooke a line at brooke.eichhorn@behamics.com or visit www.behamics.com