ArticleIssue #175 | The 2026 Top 5: A Holiday Reading List
These are the 5 issues that generated the most activity from the ~175 published so far. If you’re new here (and a lot of you are), this is the fastest way to get the jist of what this newsletter is about. If you’ve read every issue, well then – either (a) enjoy a few quiet minutes while pretending to read this newsletter or (b) re-read these ones, in case there’s something in there that can help you going forward.
1. Customer Service Is The Only Thing That Matters
The one that started with me trying to return a few bags of clothes that didn’t fit. Nordstrom: a breeze. Nordstrom Rack: a manager who flatly refused the refund and then told me that next time I should read the return policy. That single experience had me questioning a decade-long relationship with the brand, which is the whole point. Relationships and trust are built in drops, over hundreds of interactions, and lost in buckets, in a single bad one. The argument that follows: your customer service experience is your brand, far more than your logo or your ads, and the rush to automate it for cheapness rather than improve it is one of the great unforced errors of the AI era.
2. The AI Regulation Wars Begin
The issue I told people to forward to whoever runs their creative production, before June 9th. That was the day New York’s synthetic-performer disclosure law took effect: the first state law in the country requiring advertisers to disclose AI-generated humans in ads, with no cure period and penalties that stack per creative. Run 50 undisclosed AI video ads and you’re not looking at a $5k problem, you’re looking at a $250k+ problem. What’s worse – the burden runs backwards. The state doesn’t have to prove your ad used AI; you have to prove you didn’t. With 15+ states advancing their own versions, this was the issue that turned “AI compliance” from a someday-problem into a right-now problem.
PS.: The deadline has passed, so we’re all living in the synthetic disclosure world right now.
→ Read it here
3. Diminishing Returns Are Not Failure
The cognitive upgrade I wish every performance marketer would make. Every time a quarterly (or annual) review rolls around, brands/agencies look at a channel that used to print money, decide it’s broken, and blow it up – undoing months of learning to fix something that was never actually broken.
Why?
Because failure in the marketing world occurs when something stops being viable. Diminishing returns is when something works, so you do more of it, until you hit the natural ceiling of the current inputs. It’s the same 200-year-old law Malthus described: and it’s running in your ad account right now. The issue was never that the platform didn’t work – the issue is that the platform worked too well, so you blew past diminishing returns. You asked the platform/tactic to do something more than it was capable of doing given its current limitations (data, audience size, budget, etc.). The solution isn’t to throw out the proverbial baby with the bathwater – it’s to think at the margin, not in blended averages.
4. The Questions CRO Doesn’t Want To Answer (Parts I & II)
A two-parter that received a surprising amount of love (and some hate) from the CRO community. Part I opens on an 8-figure brand that ran 47 tests in a year (all documented, rigorous and maintained in a beautiful spreadsheet) and got a whopping 12% lift to show for it. Why? Because they didn’t have a test process or volume problem (i.e., what most brands *think* they have) – they had a test selection problem. Their team (and their leadership) had unknowingly created a set of incentives that rewarded stacking up safe, easy wins instead of taking bigger, bolder, shots that could have generalized across the business.
Part II, the one that really lit up my inbox – is where the conclusion from P1 goes from interesting to actionable. In it, I’ve shared my framework for classifying every test by which question it answered and at what magnitude. When we applied this to testing data from 100s of accounts, a pattern emerged: everything clusters in safe, incremental “10% cells” while the difficult-to-implement, high-leverage ones (recognition, trust, differentiation) are habitually ignored. The big-picture takeaway for any CMO or CRO person: page-level questions aren’t independent optimizations, they’re steps in a story – which means your conversion rate is a product (not a sum) of all of them. The right test isn’t the highest-expected-lift one. It’s the one that addresses the bottleneck.
→ Start with Part I, then read Part II
5. The Click Isn’t Dead
The most-replied-to issue of the year, and the one that kicked off everything we’ve been talking about since (including last week’s AI search to-do list). The issue’s core argument: AI didn’t eliminate the click, it relocated it. Platforms benefit enormously from convincing you that visibility and citations are the new finish line, because the moment you accept that, you’ve handed them the ability to grade their own homework. The antidote isn’t a clever new AEO/GEO/AIO tactic; it’s becoming a brand distinct enough that a customer asks for you by name, because if an AI’s answer about your category could be any brand, your problem was never visibility.
The Takeaway
If there’s a thread running through all 5, it’s the same one I keep coming back to: in a year where everything is getting automated, optimized & generated, the winning move is the human one: serve people like they matter. Tell the truth (and document it). Know the difference between a ceiling and a wall. Be uniquely yourself – and impossible to confuse with anyone else. The cost of creation is going to zero – which means the only thing with value left is curation. Taste. Knowing what’s worth creating in the first place.
That’s it from me this week. Back to our regularly scheduled programming next week. Until then, go enjoy the long weekend. The ad accounts will still be there Monday, I promise.
Cheers,
Sam
