Happy Sunday!

If you're reading this, I hope you got a good sweat this weekend, you're hydrated, and you have a few minutes for something I haven't been able to stop thinking about since a call earlier this week.

I was talking to an established consumer brand... real team, real spend, years in market. Over the last three to six months, they'd built somewhere between 20 and 30 paid landing pages. Different products, different angles. That is a lot of work.

The PDP was still winning.

And look, that doesn't mean landing pages are dead. I've watched a landing page take a client from a $100 CPA to a $50 CPA. Landing pages work. But this brand's experience means something most teams don't want to hear... page count and learning are not the same thing. You can ship 30 pages and know nothing new about your customer.

A new URL is not a new test. That's what today is about. But, before we get into that...

Vendor of the Week

Audiense — Same checkout, six completely different reasons to buy.

Most BFCM plans are built around the promo hunter: bigger discount, louder countdown, one more urgent subject line. That shopper converts, then disappears. According to Audiense’s new ungated BFCM report (goes straight to the PDF), only 27% of first-time Black Friday buyers purchase again within the next 12 months.

The better customer already decided to buy. They’re waiting for one condition: permission, the right price, a reason to stock up, proof they can cancel, a real shipping deadline, or confidence that next year’s model won’t make today’s purchase obsolete.

That’s the idea behind “Six reasons your customers are waiting for BFCM”. The report breaks these shoppers into six behavioral patterns, then shows how the creative, offer, and message should change for each one.

The best line in the report: BFCM conversion is a release event, not a persuasion event. If the intent already exists, “25% off everything” is not the strategy. Recognizing why the customer waited is.

Audiense Action helps model these audiences before you spend. Audiense Online, powered by Elevar, helps recover the conversion data that pixels, cookies, and platform reporting miss after the campaign goes live. You get both sides: what to run, then what actually worked.

They made the 31-page report ungated. No form, no friction, just the playbook.

Read the free, ungated BFCM playbook.

This week on Limited Supply

This week's episode is basically the operational companion to today's email. I got into how the landing page production process should actually work now... what stays in-house, what goes to a vendor or freelancer (spoiler: more than you think), and the tools that make the whole thing run. If today's topic hits, queue it up.

Hit play on Apple Podcasts, Spotify, or YouTube. Just search "Limited Supply"... it's S17 E4.

Alright, let's get into it.

What are you actually testing?

Here's the thing. When a team tells me "we're testing," they usually mean one of three things, and the difference between them is the whole ballgame. Sometimes they're testing an element. Headline A versus headline B, green button versus black button, reviews above or below the fold. Fine. You'll learn which button is better. You will not learn why anyone wants your product.

Sometimes they're testing a page. PDP versus advertorial, old design versus redesign. Better. But the traffic hitting those two pages is usually carrying five different creative promises, the new page changed ten things at once, and nobody wrote down what they expected to happen before launch. So one URL wins, and the team still can't tell you why. I see this constantly.

What almost nobody tests is the thing the customer actually experiences... the whole path. The ad that made a promise. The page that continued it (or didn't). The product and offer sitting at the end of it. The checkout. The confirmation email. The customer it created.

I've started calling that a buying route (like a paper route). Ad → argument → offer → checkout → customer. The page is one surface inside the route.

Once you see it that way, you realize most landing page programs start in the middle. Someone grabs a template, rewrites the hero, moves the reviews up, adds a comparison chart, and calls it a challenger. But the customer didn't start on that page. They started with a problem, or a creator they trust, or a random scroll on the couch. The page only makes sense as one piece of that path.

The PDP doesn't need to be exciting. It needs to be beaten.

Back to the brand with the 20 to 30 pages. The easy takeaway would have been "your PDP is boring, keep swinging." Wrong takeaway.

Your PDP is probably the most battle-tested selling surface you own. It has the navigation, the reviews, the FAQs, and all the trust you've accumulated over years of being in business. It's been quietly converting traffic since day one. It's not the problem. It's the control.

So when I build a challenger, I start with wherever the majority of paid cold traffic is landing today, and I make the new page earn its spot. Same audience, same creative, same offer, same window. And the challenger has to have one advantage I can say out loud... a clearer argument, a stronger offer, a lower-friction way to buy, proof that actually matches the ad. If I can't name the advantage in one sentence, I'm not testing anything. I'm redesigning.

That reframes the question: was my idea strong enough to beat a page that already carries most of my brand’s trust?

20 that never answer that question is production, not learning.

The ad made a promise. The page has to keep it.

Nobody wakes up wanting to browse your site's navigation. Cold traffic clicked because one idea created a spark of interest, and your job is to keep that spark alive all the way to the purchase.

I think of it as a mental receipt. The ad made a promise, and the visitor lands holding that receipt. If the ad tells one story and the page opens with a different one, the customer has to restart the sale from zero. Most won't bother.

And no, this isn't "make the headline match the ad." It goes deeper than that. Someone who already knows they need an air purifier can handle a page that jumps straight into why yours is better. Someone who just learned that indoor air can be worse than outdoor air needs the problem built up first, before they ever see a product.

Here's a live one. I've been working with a brand in the water space on a page where someone in New York sees an ad about what's actually in their shower water, and the page opens with their local water data. It explains why it matters, then introduces the filter, with proof stacked against the exact worry the ad created. The clever part isn't the personalization. It's that the visitor's context decides which argument they hear before the product shows up.

That’s a real test... a button color swap can’t teach you any of it.

Give your best creators their own front door

The traffic source is part of the experience too. A good creator doesn't just generate awareness... they hand you a specific kind of trust, a vocabulary, a use case, a recommendation. And then most brands take that traffic and dump it onto a generic page, where the homepage awkwardly redoes the introduction the creator already nailed.

Look at Ridge. Their primary always-on offer is a 10% influencer code. The creator is literally the front door to the purchase.

On two separate calls recently, my instinct was the same both times: find the creator or affiliate driving unusual sales, and build the funnel behind that person. Continue their demonstration on the page. Show the configuration they recommend. Answer the objections sitting in their comment section. Use their clips as proof next to the brand's own evidence, and land on one recommended product... not the catalog. (One of those brands, by the way, was still sending affiliate traffic to a dead page. Go check yours. Tonight.)

Does every influencer deserve a custom funnel? No. Build one modular page where the hero, proof, and product swap per creator, and save it for the people already driving real volume. If a creator did the work of making the customer care, don't make the homepage do the introduction again.

Your buy box is trying to do too much

Now for the part nobody thinks of as "testing" at all... the offer itself.

A team I was working with recently had pricing, percentage savings, a free gift, and a one-time-to-subscription push all crammed into a single purchase section. Every lever sounded persuasive on its own. Together, they forced the customer to do math at the exact moment they should be saying yes.

Sorry, but that's not an offer strategy. That's homework. Respectfully.

An offer can be generous and still be hard to buy. If the offer needs a spreadsheet, you’ve handed the customer a menu.

The same goes for commitment. A one-time purchase, a 30-day subscription, and a 90-day subscription are not minor variants inside one buy box... they're completely different asks. A customer can want your product and still say no to the way you're asking them to buy it. When we launched IM8, the 90-day subscription doubled AOV from $110 to $220, but that offer could not have worked on day one. It needed 16,000 reviews behind it first. The commitment you can ask for is a function of the trust you've earned.

So split it. Run one route with the lowest possible barrier... one product, clear price, no strings. Run one built around the best-outcome bundle. Run one built around continuity, and make that one coherent instead of coupon-y (Graymatter does this beautifully... their subscribe-and-save isn't a percentage off, it's a month-by-month unlock of gifts and cash back). Hold the customer angle steady, test them one at a time, and let the commitment be the thing that changes.

Do it stacked, and a winning test can't tell you whether price, gift, duration, or subscription did the work. Do it split, and suddenly the result means something. And measure past the first order... a subscription push that wins the checkout and bleeds cancellations before the second shipment didn't win anything.

The route doesn't end at the buy button

A page can make a clean promise and still lose the customer when the cart changes the price, the checkout buries the commitment, or the confirmation email explains nothing. Whatever decision the customer thinks they made, the rest of the route has to keep it.

So test the actual route, with real orders and a real credit card. Not previews. One team I know is doing exactly this right now... placing separate one-time, 30-day, and 90-day orders, abandoning carts on purpose, and reading every confirmation and post-purchase email the way a customer would. Unglamorous? Extremely. It's also closer to a real test than another hero swap.

Want to know what a broken route looks like downstream? I've seen brands fielding customers trying to cancel subscriptions they never knowingly purchased. Support didn’t cause that. The route never communicated the decision. If the customer can't explain what they just bought, the route didn't convert cleanly... and every "winning" test that created those customers was lying to you.

Judge the route by the customer it creates

Conversion rate is an intermediate signal. A route can produce a cheaper first order while attracting confused subscribers and refund machines. The route that wins is the one that creates the best customer for your economics, and you only find that out if you look past the purchase.

The simplest way in is the post-purchase survey, and most brands stop one question short. If someone says they found you on Facebook, keep going... was it a static ad, a creator video, press? Feed those answers back into the map of which route created which customer. Acquisition is a focus group at scale. Every route you run should teach you something about why people buy, and honestly, even a losing route earns its budget if the hypothesis was clear.

If I were running this test tomorrow

Here's the Monday version. Pick the control first... wherever your paid traffic and money are actually going today, PDP included. Write down the whole route as it exists: the ad, the page, the offer, the commitment, the emails.

Then write one sentence: "We believe [this customer] will buy [this product, this way] when the route leads with [this argument], because [this insight]." If you can't fill in the blanks, you're not ready to build. And change one thing that matters... the argument, the source, the offer, the commitment. Not the button.

Build only what the idea requires. Sometimes that's a new page. Sometimes it's a pre-sell article or just a cleaner checkout. Decide before launch what winning looks like and what you'll do about it, and when it's over, write down what it taught you. Not "LP B won." Which customer responded, which argument earned attention, what they were willing to commit to, and what the next route should test.

Quick word on AI, because it's the reason any of this is finally practical. The production cost of a genuinely different angle has collapsed... research synthesis, copy variants, modular page assembly, all of it. The first template still takes real hours. After that, testing a new customer argument stops being a six-week project. But be careful with the speed. AI can produce weak variants faster than any team in history, which means it raises the cost of sloppy thinking. It can increase your testing velocity. It cannot tell you whether the tests add up to learning.

One more thing this fixes that nobody talks about: the org chart. Right now your creative team, your media buyer, your CRO person, and your lifecycle team each optimize their own surface, and nobody owns the complete argument the customer experiences. A route test forces everyone onto one hypothesis. Creative knows the promise it's introducing. Web knows the argument it has to continue. Lifecycle knows the expectation the customer was given. The customer never experiences your departments... they experience one path to a decision.

That's all for this week!

20 or 30 landing pages is a lot of work. It is not automatically 20 or 30 ideas. Test the path, not the page. Do that consistently and remember that even the losing tests move you forward.

Before I go... the Q4 Summit is coming to New York on September 24th. One day, one room of serious operators, everything you need to win the biggest quarter of the year. Apply here for a free ticket and get it on your calendar now!

It's Sunday night, so I hope you plan to get a full 9 hours of sleep going into the new week (8 last night for me, and 9 tonight). I'll see you here next Sunday. Same time, same place.