Field Notes
Sharper Stories.
Better Pipelines.
Practical writing on B2B messaging, positioning, AI search visibility, and why most homepages don't convert. Written from inside the work, not the conference circuit.
Greg Rosner
Founder of PitchKitchen · Author of StoryCraft for Disruptors
Find your problem
All ten, with the short answers →
Start Here
Anchor · Magnetic Messaging Framework
How do I know if my B2B messaging is broken, not just underperforming?
Underperforming B2B messaging needs tuning. Broken B2B messaging needs rebuilding. The difference between them decides whether the next $50,000 you spend lifts pipeline or just amplifies the silence. Three tests, 30 minutes, no agency, tell you which one you're actually looking at.
9 min read
Anchor · Magnetic Messaging Framework
What changes about B2B positioning when AI is doing the buyer research?
The buyer changed and so did the first-pass evaluator. ChatGPT, Claude, Gemini, and Perplexity now summarize, rank, and filter B2B vendors before the buyer ever clicks a homepage. Three diagnostic prompts tell you in 20 minutes whether AI buyer research has already filtered you out.
10 min read
Anchor · Magnetic Messaging Framework
What does it mean when my marketing spend is going up and my pipeline is going down?
When spend climbs and pipeline falls, the problem is almost never the tactics. It's bottlenecked marketing ... stuck behind a positioning problem upstream. Five diagnostic questions you can run this week, and the four moves that fix it, in order.
10 min read
Anchor · AEO Strategy
Why doesn't AI cite my B2B company when buyers ask for recommendations?
AI engines don't cite B2B companies that sound like every other B2B company. If you're missing from ChatGPT, Claude, Gemini, and Perplexity answers, the model didn't fail to find you. It found you and decided your content didn't earn the citation slot. Five tests, the Princeton GEO numbers behind the pattern, and the order in which to fix it.
11 min read
Anchor · Sales-Marketing Alignment
Is my B2B sales cycle slow because of sales execution or because of my message?
Most B2B founders blame sales execution when cycles stretch from 80 days to 120. The cause usually sits one layer up. If your message isn't doing the work, your reps have to do it inside every call, and the deal slows under the weight of the explaining. Five tests to separate message-caused friction from execution-caused friction, plus the order to fix it in.
11 min read
All Insights
How do I use win/loss data to improve our positioning?
Read win/loss data one positioning decision at a time: lost deals show your real alternative, fast wins show your best-fit customer, buyer words show your value. Here's the method, plus why 278 of 302 scored homepages never name the alternative.
What should our company be known for?
Your company should be known for one problem, one buyer and one belief, said in seven words anyone can repeat. Here's how to find it, test it, and why only 12 of 302 scored homepages get there.
How do I turn positioning into messaging?
Positioning is the decision. Messaging is that decision said in your buyer's words, format by format. Here's the translation, piece by piece, and the six-step process to close the gap.
Why does our sales team struggle to explain our value?
Your reps can describe the product. What they can't describe is the buyer's before: what the problem costs today and what changes after you. Here's how to write it down so they can say it.
How do we answer "why should we choose you over competitors?"
Stop starting with features. A five-part answer that moves the comparison onto the buyer's problem, names the old way, proves one difference and says who you're wrong for.
How do we figure out what makes our company different?
Your difference is already hiding in deal stories and your best customers' words. A five-step way to dig it out, the questions that work, and why AI engines need it in your own language.
How do we know if we have a positioning problem?
You have a positioning problem when buyers can't say who you're for, what you replace, or why they'd pick you. Seven symptoms, a 30-minute leadership test, and how to tell positioning from messaging.
Who should we hire for B2B product positioning?
Positioning help covers three different jobs: making the decision, documenting it, and carrying it. A straight hire-X-if-Y answer, an honest credit to April Dunford, and why the product marketer comes second.
How do we position an AI product?
Lead with the problem you end and the buyer you built it for, then describe the AI as a mechanism a buyer can check. Five steps, the AI-in-the-category question, and data from 105 AI-category homepages.
How do we create a differentiated sales pitch?
A differentiated sales pitch names the buyer's old way, what it costs, and a point of view a competitor wouldn't co-sign, before the product shows up. Five steps, the competitor-rep test, and data from 302 scored homepages.
What category should our B2B product be in?
Your B2B product belongs in the category your buyer already types into ChatGPT and Google, with the problem you solve named inside it. Four tests to pick it, with data from 302 scored homepages.
How do we turn product features into customer value?
You turn features into customer value by running the chain backward: start with the problem the buyer already feels, in their words, then show the feature that ends it. Here's the five-question method, with data from 302 scored homepages.
How do we position a product in a crowded market?
You position a product in a crowded market by changing what buyers compare you on: pick the slice you win, name the old way, and stand against it. Here's the method, with Dunford's three styles credited and data from 302 scored homepages.
Why does our messaging sound like everyone else's?
Your messaging sounds like everyone else's because it was built from the same inputs: the product's features, the category's vocabulary and a committee's edits. Here's what that looks like in 302 scored homepages, and how to fix the inputs before the copy.
How do we identify our best-fit customers?
Your best-fit customers are already on your customer list. Here's how to sort them by behavior, what to ask the champions, and how to write a profile your team and your AI tools can both use.
How do we create a compelling B2B value proposition?
A B2B value proposition compels when the buyer recognizes their own problem in it, before they ever see your product. Here's the four-line method, where the raw material comes from, and three tests to run this week.
How do we create differentiated B2B messaging?
Differentiated B2B messaging starts with what you're different from, not with better adjectives. Here's how to find the difference, write it down once, and keep AI from averaging it away.
Why do our AI marketing agents sound like everyone else?
AI marketing agents write generic work because they start every job with nothing of yours to read. Marketing engineering, the role Greg Isenberg is putting a name on, starts with the brain. In B2B, where you sell to a few hundred accounts, the brain matters even more. Here's what goes in it and which agents to run first.
Shared space or MCP: how do we keep every AI on our story?
Your reps draft in ChatGPT, marketing writes in Claude, and every tool tells its own version of your story. There are two ways to give them one source of truth. Here's when a shared space is enough, and when it's time for Brand Twin over MCP.
Are buyers using AI to fact-check our sales claims?
Sixty-two percent of B2B buyers now use AI to check vendor claims against public sources, and 56% have dropped a vendor after it found a gap. The problem usually isn't lying. It's five assets telling five slightly different stories. Three tests show you what your buyer's AI sees.
How do we know if ChatGPT actually sent us a customer?
AI engines strip the referrer, so your AI-sourced buyer shows up as direct traffic weeks later. Here's the three-signal join that proves whether AI search sent you a customer, and what one of our own 24-day bookings showed.
There is no best B2B messaging company until you name the job you're hiring for
A CEO sent me the list ChatGPT gave him for 'best B2B messaging company.' Every list like it, ours included, is written by a firm that sells messaging. Here's how to read them, and the three jobs hiding under that one search.
Does our company need a new name, or just a clearer story?
Pure Storage renamed itself Everpure in 2026. Almost no B2B company should copy that. Here's the one condition that makes a rename the right call, what it costs now that AI engines file your reputation under your old name, and the cheaper decision most founders actually need.
Is calling our product AI-powered still helping us, or is the word costing us trust?
Every competitor can generate the same AI sentence for free, and as of August LinkedIn is demoting polish that lacks substance. Here’s the deletion test that tells you whether your AI claim is load-bearing, and what to do with the word when it isn’t.
How do we run April Dunford's first positioning step ourselves?
Competitive alternatives are what your buyer would do if you vanished tomorrow, and the list is usually a spreadsheet, an analyst, or nothing at all. Here's how to build it yourself in a week, with your own deal data.
What can't a go-to-market consultant fix for us?
A GTM consultant can extract, document, and pressure test your position. They can't make the decision, supply the conviction, force adoption, or fix a product gap. Here's the line, with data from 302 scored homepages.
We scored 10 real B2B positioning statements. One passed.
We graded 302 real healthtech homepages on the four parts of a positioning statement. Exactly one scored full marks. Here are ten of those cases with their real language, their real scores, and the one part nearly everybody drops.
One customer is half our revenue. Do we chase another like them, or sell to everyone else?
One customer at 40 or 50 percent of revenue is tracked as a financial risk and never as a messaging risk. Meanwhile the homepage, the case studies and the roadmap all describe that account's problem, so every other buyer reads material written for somebody else. Three tests to find out, and what to do about it.
Nobody copies the true part of your message
Last Tuesday a construction SaaS CEO sent me two homepages and asked which one was his. A competitor had lifted his headline, his three value props in order, even his customer outcome sentence. Then I asked him to point at one thing in it only his company could say, and he couldn't find it. That's the copyable claim, and almost every B2B homepage is built out of them. Here's the villain, three tests you can run in twenty minutes, and what to fix first.
Does StoryBrand work for B2B? Where it fits, where it breaks, and what a complex sale needs instead
StoryBrand works exactly as far as it promises: a clear page for one buyer who already believes in the problem. Roofers and HVAC companies prove it. A company selling AI-enabled RCM to a hospital committee fails all four fit conditions. Where it breaks, what the critics say, and what a complex sale needs instead.
What should actually be in our go-to-market plan?
Six decisions, each with one owner and one date. That's the whole plan. Here's what belongs in each row, what to cut, and the ten-minute test that tells you whether yours exists on paper or only in somebody's head.
Your AI needs two knowledge bases, and most companies have built neither
AI writes generic work for your company because it has nothing of yours to read. Real work needs a knowledge base with two halves: documented SOPs on the inside, an AI Brand Twin on the outside. Most companies buying AI seats have built neither, so the model fills the gap with the average of the internet.
Do we need a GTM consultant, a GTM advisor, or a fractional CMO?
Three roles, three prices, three completely different things in your hands when they leave. A consultant hands you a decision, a fractional CMO hands you execution, an advisor hands you a second opinion. Here's what each costs and how to tell which one you're actually short of.
What is an AI Brand Twin, and how does it shape what AI engines say about us?
Three narrators describe your company to buyers every week: your team, your AI tools, and the buyer's AI. You supervise one. An AI Brand Twin is the written source that makes all three agree, which is how it changes what the engines say about you.
Is a strategic narrative in our sales deck enough?
A strategic narrative in a sales deck only works in a room where someone presents it. Two of the three places a B2B deal now gets decided have nobody from your company in them. Here's what the deck can't reach, a ten-minute test, and what to build instead.
Your reputation doesn't cross the border
Last Tuesday a cybersecurity CEO showed me six months of flat German pipeline and told me the problem was the rep he'd hired in Munich. I asked him one question: who is the buyer in Munich paying right now to do this job? Nobody on the team had an answer. That's the Home Field Message, and almost every company carrying one finds out at the border. Here's the villain, three tests to run before your next market entry, and what to fix first.
Explaining is the tax you pay for a story you never published
A rep I listened in on last Thursday spent eleven minutes explaining what his company does. The buyer already had the answer open in another window. She'd asked an AI three days earlier, read what came back, and booked the call to check it against a human. He never got to that conversation, because he was still doing the introduction. That's the Live Rebuild, and most B2B companies run it on every first call. Here's the villain, three tests you can run on calls you've already recorded, and what to fix first.
The week a competitor raises is the week your category gets decided
A funding round doesn't hand your competitor the market. It hands the category a few weeks of attention it won't get again all year, and whoever answers the buyer's question most clearly during that stretch becomes the reference point everyone else gets compared to. Most CEOs respond by going quiet until the noise passes. Here's why that's the expensive choice, three tests to run this week, and the one thing that lets you move in days instead of quarters.
Why isn't our go-to-market working, and how do we tell which layer broke?
A go-to-market has five layers, and they fail downward. The break almost always sits above the layer you're staring at, because the layers with dashboards are the ones at the bottom. Here's the day-long diagnostic that finds the real one.
Which positioning framework should we use, and where does each one break?
Seven positioning frameworks compared by their breaking condition rather than their reputation, with the input each one quietly assumes you're bringing, and a 302-homepage finding that explains why most rollouts stall.
How we use AI before and after a sales call
A practical routine for preparing with evidence, asking better questions and catching what you missed before you follow up. From Greg Rosner’s conversation with Paul Kirch.
Your chief of staff is the only copy of your company story
Every artifact that carries your company story has a deadline and an owner. The story underneath them has neither, which is why it gets rebuilt from scratch every quarter by whoever has the board deck due, and why the company's narrative memory quietly becomes one person's working files. Here's the Rebuild Tax, three tests to measure what it's costing you, and the structural fix.
Does our AI company need brand strategy, or just a better model?
AI founders keep choosing between the next model release and a real positioning rebuild. The model is the one part of your company a buyer never sees, and the language around it is the only part an AI engine can read. Here's how to tell which one is actually blocking your pipeline.
Your forecast knows the message is broken before marketing does
Deals keep moving their close date and every one of them arrives with a reason about a calendar. Budget cycle, a new VP, a reorg. Those reasons are true and they are almost never the cause. The CFO is sitting on two years of evidence about why buyers don't move, filed under scheduling, and nobody has ever asked to read it that way.
We finished a brand transformation sprint. Why hasn't anything actually changed?
A brand transformation sprint changes what you've decided and how clearly you say it. It doesn't change what your team does in week nine when a big wrong-fit deal shows up. Here's the honest accounting of what a strategy sprint delivers, what it leaves to you, and how to tell six months later whether it took.
Which of our products should have their own brand, and which should just be features?
Most products shouldn't have their own brand. One earns a name when it has its own buyer, its own budget line, and its own competitor. Here's the test, what a stray name costs you in sales and in AI visibility, and which architecture actually fits a $5M-$75M company.
Our champion just left. Do we work their replacement or re-open the problem?
A founder called me Tuesday about the best deal in his quarter, eighty percent in the forecast for two quarters running. Then his champion took a job two states away and the replacement went politely, permanently quiet. He didn't lose a relationship. He lost the only person who thought the problem was worth solving this year. Meet the Borrowed Deal, why chasing the successor is the one play that guarantees a slow no, and three tests to run on your forecast this afternoon.
Nobody put the founder's story in the handover
A new CEO takes over from a founder and everything transfers cleanly: contracts, roadmap, comp plans, brand guide. One asset never got written down, because it lived in the founder's head, so nobody put it on the list. What he wrote keeps working for about two quarters, which is exactly long enough for everyone to agree the handover went fine.
Want this thinking built into your story?
The 90-Day Magnetic Messaging Sprint rebuilds it for you. Fixed price.
Strategy plus execution. Your story, homepage, sales deck, and an AI Brand Twin trained on your narrative ... rebuilt in 90 days. You own everything at the end.