• Zero Clicks
  • Posts
  • Zero Clicks #22: Product search is an impossible problem

Zero Clicks #22: Product search is an impossible problem

Or the bull case for commerce creators

At AS, one thing is clear—affiliate marketing is evolving.

The biggest brands, agencies, and creators are adapting. Are you?

Amazon drives 37% of U.S. eCommerce sales, yet most affiliate strategies still overlook it. That’s because, until now, brands on Amazon didn’t have the right tools to scale.

Levanta changes that.

💰 Affiliates – Earn higher commissions by working directly with Amazon brands that value your audience.

🛍️ Brands – Drive external traffic and sales while improving organic rankings.

🏢 Agencies – Create a new revenue stream by offering affiliate marketing for your clients’ Amazon stores.

The best Amazon brands, agencies, and creators are already using Levanta—what’s your move?

Every two weeks in Zero Clicks, we explore the interplay of AI, media, and commerce. In each edition, we explore…

  • The big picture: Creators, product search, and trillion-dollar problems  

  • Job posts: Help Wanted: The AI needs to sound more human 

  • The meme is the message: A vibes guy is talking   

  • Cocktail hour: The official cocktail of the Consumer Reports revival era  

Many of you are out in Vegas this week for Creator Economy Live and/or Affiliate Summit West. Alas, I am not. I prefer the company of my keyboard, office plant, and dark thoughts to $19/night hotel rooms at the Harrah’s on the Strip.

That said, I’m particularly interested in the session that friend of Zero Clicks Blake Saunders is leading with Perplexity’s Head of Publisher Partnerships and two of the best commerce media minds in the game on “AI in Content Commerce.”  This one has the potential to get saucy so please report back. Now, let’s get to it.

Product search is an impossible problem

For the past several weeks, I’ve been in the market for a nonstick pan. I’m generally a stainless steel diehard but several times per week, a toddler stands on my barstool and demands EGGIES. I am quite frankly running out of butter and patience here.

My only real criteria is that I want a pan with some durability that doesn’t have the cancer chemicals in it. Should be simple, right? Nope. Enshittification of search notwithstanding, nearly every marquee commerce media brand disagrees on which nonstick pan is best… or if even using nonstick at all makes you a Bush league idiot. Furthermore, I’ve worked in this business for a decade and even I’ve mostly lost track of how you can clearly discern if something is quiet pay-to-play. On Reddit, the dynamic is even worse, with scores of dissenting opinions and the chance that chef42069 might say I’m committing child abuse by even considering nonstick.

No problem, I’m an eCommerce professional– let me just poll a community of practitioners for their recommendations in an industry Slack. How’d that go, you ask? Seven different brands were recommended (Caraway, Hexclad, Our Place, some dude’s to-be-launched brand, MadeIn, Tramontina, and some British guy saying “just use an egg cooker mate”) with each name brand recommended triggering a subsequent dissenting opinion.

On the surface, this looks like a ringing endorsement of LLMs and agentic commerce. Powered by semantic search, experiences like Amazon’s Rufus will tame the hopeless abyss of the conventional product search grid into a streamlined set of 1-3 product recommendations, perfectly tailored to your browsing interests. Finally, AI will solve the paradox of choice.

The fundamental flaw in this thinking is that AI is explicitly trained by the same morass of competing information about products and services that makes conventional product search impossible. In the aggregate, enough review data should help the “best” product bubble up to the top. But again, even a detailed query like “best cookware for an intermediate chef that costs less than $100” has layers of subjectivity.

Let’s think back to my nonstick pan conundrum.

At the risk of being reductionist, an answer engine is going to try and find the best pan for me by essentially cross-referencing a set of user and expert reviews (that often contradict each other) against my previous buying history and price sensitivity. There are infinite problems here– whether or not you like a given pan depends on a litany of minute details in your cooking style and several other irrational heuristic details that will always be beyond the reach of AI.

As a small aside, this is why highly personalized recommendations still lose half of their A/B tests to the most basic common sense suggestions. We are both complex and simple creatures who generally want french fries with our burgers.

Individualized commerce has and always will be a pipe dream (see the meme below). Humans are just too complex, irrational, and rapidly context-switching to ever really know anything about them to drive better conversion rates than basic behavioral economics can. So who or what can shoppers trust?

In the face of cosmic uncertainty, people just want a single entity that they vibe with to tell them what to do. This–at the most fundamental level– is the function of a brand in media and commerce.

This is why Consumer Reports is still the pre-eminent consigliere in its space for boomer dads, and I predict soon for Gen Z and Gen Alpha as well.  This is why Wirecutter has only tapped a fraction of its potential value to the New York Times. This is why Marques Brownlee can single-handedly determine the fate of a new tech product launch.

But more than ever, the prevailing vibes favor individuals over institutions. This is radically underrealized in the market today. In plain English, there is room for scores of additional Marques Brownlees to be minted in the commerce creator space. Product search is a large enough market and it damn well sucks enough. Algorithmic personalization had its shot, let’s give people a chance.

However, the rate-limiting step to realizing the potential of commerce creators is the current set of business models for monetizing creator commerce. Historically, most of the product review and commerce content ecosystem has been monetized via simple affiliate marketing, a business model that worked exceptionally well when Google and Facebook had the traffic spigots opened to eleven.

Today’s best creators– commerce or otherwise– are essentially niche vertical media brands that need to find a plethora of revenue streams to capture the inherent value they provide to their audiences. But we’re still mostly applying yesterday’s business models to this industry. By and large, creators are still left to negotiate cumbersome flat fee deals with brands or use affiliate tech that is designed for very different purchase journeys. And the traffic they can realistically expect to achieve is a fraction of what the platforms sent five years ago.

For those of you who are in Vegas and beyond, this column essentially is a call to think first principles about the set of technology and business model innovation that is needed to help more middle-class commerce creators break out. I say with no small sense of pride that a lot of smart people read this newsletter; a lot of you are in the same place this week and I hope you make some magic.

The bull case for the creator commerce economy is simple– product search is a multi-trillion dollar market and it still doesn’t really work.

All the big data, AI, and recommendation algorithms in the world won’t meaningfully improve product search one iota. People with a near maniacal obsession for finding the best things to buy will.  

Job Posts: Each week we feature job postings that we believe are microcosmic of larger corporate strategies and broader trends in the zeitgeist.

Creative Director, Consumer Reports

In the age of the tradwife renaissance, Consumer Reports appears poised to become the latest slice of old Americana that becomes cool for Gen Z and Gen Alpha’s soon-to-be trad dads. And hey, maybe you’re the person who imagines how the brand does it.

Consumer Reports will celebrate its 90th birthday next year and brings in a tidy $250M/yr in revenue as a nonprofit so as far as job stability goes in the commerce media world, hard to do better.

Wake up, human. Your commerce LLMs need training data.

Facetiousness aside, the somewhat ironic thing is that more than anything, this job exists to give Rufus a conversational tone or “authentic” voice. On the one hand, farming out a dash of your humanity for an entry levelish salary might not feel great. On the other, as an early in career writer/editor, there are few opportunities where your work can have this scale of impact.

“In this role, you will sit at the intersection of two incredibly important and fast-moving spaces - Trust & AI. You will be responsible for deeply understanding how evolving policy and regulation affect how LinkedIn builds and launches AI products. You will play a critical role in launching new AI features across the consumer and customer audiences.”

All I ask is that one of my readers get this job to become our inside man to push back against AI-generated slop comments that are enshittifying LinkedIn at an exponential rate.

(DJ Khaled Voice)...ANOTHER ONE!  

While not yet a household name, OfferFit is a highly intriguing enterprise SaaS company that more than others, seems to be skating to where the puck is headed in marketing tech by building “personalization” that is grounded in rapid experimentation and testing above all else. The company also pays really nicely for Series B software.

This one needs no introduction. Apple is looking for a new content strategy leader for its consumer marketing function. Godspeed.

New to Zero Clicks?

If you'd like to get Zero Clicks in your inbox every week, you can subscribe here for free.

The meme is the message

I’m breaking my own rule and including a nonoriginal meme this week because this one is just too on point.

I’m not even sure who the original creator is in the game of internet telephone but massive thanks to my friend and Zero Clicks reader, Jonathan Senin, for this slice of content perfection. 

Cocktail hour

Amidst the grand revival of pre-war cocktails, the prettiest pony of the bunch has yet to have her revival tour. My dear friends, it’s up to you. Let’s make 2025 the year of the hot boy aviation summer.

My riff on the classic cocktail uses the usual ingredients but in slightly different proportions than the barkeep’s manual.

  • 2 oz. Dorothy Parker Gin

  • 1/2 oz. Luxardo Maraschino Liqueur

  • 2/3 oz. fresh squeezed lemon

  • 1/2 oz. Creme de Violette

Per tradition, Dorothy Parker is my choice of gin but any classically botanical forward swill will do. No substitute for Luxardo and fresh squeezed lemon here- do the right thing. Of course, the star of the show is that bizarre purple floral mixture known as Creme de Violette. Give it a chance to shine.

In general, I’m absolutely militant about cutting back the sweetness in classic cocktails. In general, you’ll find me slashing simple syrup and sugary tinctures with nearly the same reckless abandon as any savvy cook has when they triple the garlic or vanilla in a recipe. But here, I actually amp up the Creme de Violette above the classic proportions. You’re making a purple drink— if it isn’t Smoke on the Water deep purple and doesn’t immediately evoke running through a field of violet petals, what’s the point? You wanna hear Tom Petty sing the chorus of Wildflowers when you lift it to your lips.

As always, shake like your life depends on it and enjoy.

Thanks for reading. Drop me a note at [email protected] with any feedback or with topics you’d like to see us explore.