Picture a fourth-quarter budget meeting. Thirty slides, one that matters. Organic traffic — down eighteen percent year over year. The room does the math fast: fewer visits, fewer leads, fewer dollars. Someone suggests moving the search line over to paid, where at least you can see what you bought.
Two rows down, on the same slide, direct traffic is up double digits. Nobody asks about it. Direct is the column that means “we don’t know.” And you don’t build a plan on a column that means we don’t know.
So the decision gets made on the number that’s shrinking, and the number that’s growing goes back in the drawer.
I made that meeting up. But it’s the default shape of 2027 planning right now. And it’s backwards.
Hi, I’m Jeff Payne. You’re listening to The Jeff Payne Show, Episode #89: The Column Marked Unknown.
It’s planning season. And most companies plan organic the same way they have for fifteen years: set a traffic target, run it through a conversion rate, call the output revenue. Visits times conversion times order value. That waterfall is on the first slide of every deck I’ve ever seen.
It made sense when the visit was the first thing a buyer did. In a lot of categories, it’s now the last thing they do — if they do it at all.
In the first four months of this year, about sixty-eight percent of US Google searches ended without a click to anything. Two out of three. Not to you, not to a competitor, not to anybody. The answer was on the page, or in the AI summary, and the person moved on.
That’s the demand side. Here’s the other half: when buyers do finally show up, the systems that are supposed to tell you how they got there can’t see most of the trip anymore.
Walk a modern purchase. Someone reads a thread on a forum. Searches. Lands on a product page. Asks an AI assistant to compare two model numbers. Asks a second assistant. Goes to a competitor to check sizes. Comes back two days later, types the address in directly, and buys.
Eleven stops. Two of them were AI conversations your analytics will never register. The one that gets credit is the last one — a person typing your name into a browser bar. And your dashboard labels that one “direct.”
Here’s what I need every owner to hear: direct is not a channel. Direct is a confession. It’s your measurement system telling you that something caused that person to know your name and come looking — and it has no idea what.
Increasingly, what it has no idea about is AI. A buyer asks an assistant for a recommendation, gets your brand named, and never clicks — because there wasn’t a link, or they didn’t need one. Later they show up as direct. Or as branded search. Or as a phone call. In every case the influence happened before the visit, and the visit got the credit.
Now stack that on top of the planning waterfall. If your 2027 plan is built on what attributes cleanly, you will systematically defund the things working invisibly and over-fund the things that merely report well. Not because anyone’s careless. Because the instrument is pointed at the one moment you can still see.
And it gets harder from here. Presence in AI answers isn’t one thing you earn once. Research this summer across nearly four million AI citations found that ninety-one percent of them appeared on only one of the three major engines. One. Showing up in one assistant tells you almost nothing about the other two. So “are we visible in AI?” isn’t a question with a single answer. It has to be asked surface by surface — the way you’d never collapse every social platform into one number.
This is Decision Blur at the planning level. The buyer’s path got messier, the measurement got blinder, and the planning model kept its old shape.
So what does a defensible 2027 plan look like? Three shifts.
First: stop setting the target as visits. Set it as presence. Where is your audience researching this category — which assistants, which communities, which publications — and are you named there when the question gets asked? That’s the goal. Traffic becomes a lagging indicator of it, not the objective.
Second: stop asking any single metric to prove influence. None can. Report three layers together, on one page, every month. Exposure — are you being named and cited where buyers ask. Behavior — did people notice and go looking: branded search, engagement, discussion. Outcome — did attention become revenue, usually on a lag. Alone, each one is noise. Together, they tell you where the problem lives.
Exposure up but branded search flat? Mentions aren’t landing — that’s distribution. Branded search up but pipeline flat? That’s conversion — your pages, your offer, your handoff. Branded search and direct both up while AI citations stay flat? Something off-site is working and the machines haven’t caught up. Find it and fund more of it.
Third: fund it as a portfolio, not a line. Proven work at full capacity. Learning work capped, with one owner and a kill date. And a small set of bets funded on judgment — community, events, the brand work attribution would tell you to cancel. Lock the annual total. Re-split it every quarter against those three layers.
That’s the plan you can defend in a room full of people who want the number. You’re not asking them to trust you. You’re handing them a better instrument.
Three questions before your planning meeting.
One. If direct traffic is up, does anyone in your company know why? If the answer is no, that’s the first thing to find out — not the last.
Two. When your team reports AI visibility, is it from one assistant or across several? One is a sample, not a result.
Three. Does your 2027 plan have a single line funded on judgment rather than attribution? If every dollar has to prove itself on a dashboard, you’ve already decided what to cut.
Proof Over Proximity. Be the name they type in — then build the instrument that can see it.
If this was useful, subscribe on Apple Podcasts, Spotify, or TrueFans, and leave a five-star rating and review. It’s the fastest way to help other owners find the show. I’m Jeff Payne. See you tomorrow.
Why 2027 planning built on a traffic target will defund what’s working
Picture a fourth-quarter budget meeting. Thirty slides, one that matters: organic traffic, down eighteen percent year over year. The room does the math fast — fewer visits, fewer leads, fewer dollars — and someone proposes moving the search line over to paid, where at least you can see what you bought.
Two rows down on the same slide, direct traffic is up double digits. Nobody asks about it. Direct is the column that means “we don’t know,” and you don’t build a plan on a column that means we don’t know. So the decision gets made on the number that’s shrinking, and the number that’s growing goes back in the drawer.
That meeting is hypothetical. The shape of it is not. It is the default form of 2027 planning right now, and it is backward.
The waterfall stopped describing reality
Most companies plan organically the way they have for fifteen years: set a traffic target, run it through a conversion rate, call the output revenue. Visits “X” conversion “X” order value. That waterfall sits on the first slide of nearly every planning deck.
It made sense when the visit was the first thing a buyer did. In many categories it is now the last thing they do — if they do it at all. In the first four months of 2026, 68% of US Google searches ended without a click, according to SparkToro’s analysis of Similarweb clickstream data. Two in three searches. Not to you, not to a competitor, not to anyone. The answer was on the page or in the AI summary, and the person moved on.
That is the demand side. The other half is that when buyers do finally arrive, the systems meant to explain how they got there can no longer see most of the trip.
The visit was the first thing a buyer did. Now it’s the last — if it happens at all.
Direct is not a channel. Direct is a confession
Walk a modern purchase as tracked in real clickstream research: a buyer reads a forum thread, searches, lands on a product page, asks an AI assistant to compare two model numbers, asks a second assistant, visits a competitor to check sizes, and returns two days later by typing the address in directly to buy.
Eleven stops. Two were AI conversations that no analytics platform will register. The stop that gets credit is the last one — a person typing a brand name into a browser bar — and the dashboard labels it “direct.”
Direct is your measurement system admitting that something caused this person to know your name and come looking, and it has no idea what. Increasingly, the something is AI. A buyer asks an assistant for a recommendation, hears your brand named, and never clicks, because there was no link or they didn’t need one.
Later they show up as direct, as branded search, or as a phone call. In every case, the influence happened before the visit, and the visit got the credit.
This is one of the best written, most informative email newsletters I’ve ever read!
Thank you for sharing your expertise, and hope you’re doing well.
Beth Zimmerman
Founder, Executive Director
A plan built on clean attribution defunds what works
Stack that blindness on top of the planning waterfall, and the consequence is mechanical. If your 2027 plan is built on what attributes cleanly, you will systematically defund the things working invisibly and overfund the things that merely report well. Not because anyone is careless — because the instrument is pointed at the one moment you can still see.
It gets harder from here. Presence in AI wasn’t something you earn once. Growth Memo’s H1 2026 analysis of roughly 3.7 million AI citations found that 91% appeared on only one of the three major engines — ChatGPT, Perplexity, or Google AI Overviews. Visibility in one assistant tells you almost nothing about the other two. “Are we visible in AI?” has no single answer; it has to be asked surface by surface, the way no one would collapse every social platform into one number.
This is Decision Blur at the planning level: the buyer’s path got messier, the measurement got blinder, and the planning model kept its old shape.
Three shifts for a defensible 2027 plan
First, stop setting the target as visits and set it as presence. Where is your audience researching this category — which assistants, which communities, which publications — and are you named there when the question is asked? Traffic becomes a lagging indicator of presence, not the objective.
Second, stop asking any single metric to prove influence; none can. Report three layers together, on one page, every month. Exposure: are you named and cited where buyers ask? Behavior: did people notice and go looking — branded search, engagement, discussion. Outcome: did attention translate into revenue, usually with a lag. Alone, each layer is noise. Together they locate the problem. Exposure up but branded search flat is a distribution problem. Branded search up but pipeline flat is a conversion problem — pages, offer, handoff. Branded search and direct are up while AI citations stay flat, which means something off-site is working and the machines haven’t caught up; find it and fund more of it.
Third, fund the work as a portfolio rather than a line. Proven work at full capacity. Learning work capped, with one owner and a kill date. And a small set of bets funded on judgment — community, events, the brand work that attribution alone would tell you to cancel. Lock the annual total, then re-split it every quarter against the three layers.
That is a plan you can defend in a room full of people who want the number. You are not asking them to trust you. You are handing them a better instrument.
You’re not asking leadership to trust you. You’re handing them a better instrument.
Three Questions
Sources: Kevin Indig & Amanda Johnson, Growth Memo — “How to Plan 2027 When Traffic Is Down, and AI Won’t Attribute” — https://www.growth-memo.com/
Kevin Indig, Growth Memo — H1 2026 AI Halftime Report (91% single-engine citation finding) — https://www.growth-memo.com/p/ai-halftime-report-h1-2026
Rand Fishkin, SparkToro with Similarweb — “In 2026, Less Than One Third of Google Searches Still Send a Click” — https://sparktoro.com/blog/in-2026-less-than-one-third-of-google-searches-still-send-a-click/
Amanda Natividad, SparkToro — three-layer measurement framing (Exposure / Behavior / Outcome), via Growth Memo — https://www.growth-memo.com/
Buyer journey composite drawn from Semrush clickstream journey data as published in Growth Memo — https://www.growth-memo.com/
COMPLETE THE FORM TO
BOOK A STRATEGY CALL
"*" indicates required fields
COMPLETE THE FORM TO
BOOK A STRATEGY CALL
"*" indicates required fields
Subscribe and Share – WE APPRECIATE YOUR SUPPORT
