In my latest Lead with AI episode, I sat down with Nathan Werst, founder of Shoptimized AI, a visionary who's uncovering a shocking truth in retail marketing: millions of dollars are being wasted on ads that never drive purchases. As a data scientist who previously worked with retail giants like DICK's Sporting Goods and Farfetch, Werst noticed a disturbing pattern across the industry - retailers were blindly trusting algorithms that were burning through their budgets with minimal returns.

What began as an insight during his agency days blossomed into Shoptimized, a company using neural networks to predict with astonishing accuracy which products will actually sell. Since launching in October 2023, Werst has been showing retailers something that makes their jaws drop: up to 40% of their ad spend goes to products that don't drive a single purchase. Not just products that don't sell directly, but products that don't even influence a sale elsewhere on their website.

The most fascinating aspect of our conversation wasn't just the waste Werst is uncovering, but the solution he's built - a system that can predict with 92.3% accuracy whether a product will sell in a given timeframe. This level of predictive power is transforming how retailers approach their marketing strategy, focusing dollars where they'll actually drive revenue instead of disappearing into the digital void.

The Jaw-Drop Moment  

When new clients first work with Shoptimized, they experience what Werst calls the "holy smokes moment" - but it happens before they even start using the platform. The real revelation comes when Shoptimized pulls their advertising data and shows them exactly how much money they're wasting. For major retailers, this can amount to millions of dollars spent promoting products that never drove a single purchase.

This waste occurs because typical marketing algorithms are designed to optimize for clicks and general engagement rather than actual sales. A customer might click on an advertised product, look at it, then navigate to something completely different and make a purchase. The original ad gets credit in traditional attribution models, but Shoptimized's deeper analysis reveals these false positives. About 40% of ad spend typically goes to products that don't even influence purchases, making the problem far bigger than most marketers realize.

What makes this revelation so powerful is its immediacy and clarity. There's no complex explanation needed - just a straightforward breakdown of money spent versus sales generated that makes the waste undeniable. This moment of truth changes how retailers think about their entire marketing approach, shifting from blind trust in black-box algorithms to demanding evidence of actual sales impact.

Neural Pattern Recognition  

Under the hood, Shoptimized's technology is powered by sophisticated neural networks analyzing patterns that humans simply could never detect. The system examines hundreds of thousands, if not millions, of variables - everything from product descriptions and keywords to metadata and timing factors that influence purchasing behavior.

This pattern recognition can identify counterintuitive insights: blue items might sell exceptionally well on Sundays at 3pm, while red items consistently underperform regardless of timing. These patterns emerge from analyzing massive datasets across product catalogs, revealing connections that would be impossible for human marketers to spot. The system continuously learns and adapts, becoming more accurate with each interaction.

What truly blew Werst's mind was seeing his neural network achieve 92.3% accuracy in predicting whether a product would be purchased in a specific timeframe. Watching the system process testing data in real-time - seeing the confusion matrix update with true positives, false negatives, and watching the accuracy climb higher - was a transcendent moment. It proved that AI could do something humans fundamentally cannot: predict with near-certainty which specific products from a vast catalog will actually sell.

Ideal Customer Profile  

While Shoptimized can work with businesses of various sizes, the technology delivers its most impressive results with certain types of retailers. The ideal client has a substantial product catalog with at least 100 SKUs and enough historical data for the neural networks to analyze patterns effectively. A year's worth of data provides a good foundation, but more is always better for training the sophisticated models.

The dream clients for Shoptimized are retail giants like LVMH, Macy's, and Home Depot - companies with enormous product catalogs and mountains of data that can fully leverage the pattern recognition capabilities. Additionally, the platform works best with higher-priced products (in the $100-$1,000 range) because of the economics of digital advertising. With typical cost-per-click rates hovering around $1-3 nationally, low-priced consumer goods have inherently limited optimization potential.

This focus on data-rich environments is why Shoptimized targets marketing managers at larger retail operations or agencies handling substantial e-commerce accounts. For these professionals, the platform can deliver returns on ad spend improvements of up to 32% by redirecting budgets from non-performing products to those with proven sales potential. The more data and higher the product value, the more dramatic the potential savings and revenue growth.

Future Shopping Experience  

Looking toward 2030, Werst's vision extends beyond simply improving retailers' return on investment - he aims to transform the shopping experience itself fundamentally. The current disconnected experience where consumers click on ads for products they like only to be led to something different or unavailable is frustrating for everyone involved. Shoptimized's technology addresses this by ensuring advertisers promote products that actual customers want to buy.

This approach benefits both retailers and consumers in equal measure. Shoppers waste less time clicking on irrelevant products, while retailers stop wasting money advertising items that won't convert. Werst pointed to a common consumer frustration: seeing an appealing outfit on Instagram, clicking through, and finding either a completely different product or nothing at all. This disjointed experience drives customers away and wastes marketing dollars.

By 2030, Werst envisions shopping experiences that feel almost telepathic - where the products you see advertised are precisely those you're likely to purchase. This isn't about manipulative targeting but about eliminating friction and waste from the process. The future of retail marketing isn't just about better algorithms; it's about creating shopping journeys that respect both the consumer's time and the retailer's budget.

Take Action Today  

After my conversation with Nathan Werst, I'm convinced every retail marketer needs to take a hard look at their advertising efficiency. The good news is you don't need to wait to start improving your marketing ROI - Werst offered a simple exercise anyone can try right now to glimpse the waste in their current strategy.

Here's how to run your own quick audit:

1. Go to Google and search for one of your products that's currently out of season

2. Keep the search generic - use descriptive terms like "boots" or "tank top" rather than your brand name

3. Look at how many results appear from your company and competitors

4. Notice how many out-of-season products are being actively advertised

5. Consider how likely these ads are to drive actual purchases

This simple exercise reveals how much money is being wasted across the industry, including potentially by your own company. The results are often eye-opening, showing just how much of the advertising budget is being directed to products unlikely to sell due to seasonality alone.

If you discover this waste in your own marketing, it might be time to explore more sophisticated approaches. Whether you work with a specialized platform like Shoptimized or implement more rigorous analytics on your own, the key is moving beyond basic click attribution to truly understanding which products drive purchases. For marketers managing substantial product catalogs, the potential savings can reach into the millions.

As AI continues transforming retail marketing, the divide will grow between companies still wasting 40% of their budget on non-performing products and those directing their spend with surgical precision. The future belongs to retailers who can identify exactly which products will sell before spending a dollar on promotion. With neural networks achieving over 90% accuracy in sales prediction, that future isn't some distant possibility - it's already here for those ready to embrace it.