on a journey to PMF: the playbook changed while we were running it

we went from “let me explain what Journey does” to “I already built yours — what do you think?”

Hey everyone,

Dan Layfield wrote something on HackerNoon this week called “AI Isn’t Ending SaaS — It’s Rewriting the Playbook.”

His argument is that the fundamentals of building a business haven’t changed. You still must find an arbitrage, get profitable distribution, build defensibility, but the way you execute against those fundamentals has shifted faster than anything since the internet.

He uses this analogy about the forward pass in football: when it was introduced, it changed how the game was played, but not what it takes to win.

You still needed to score more points.

What we’re doing right now at Journey isn’t changing what we sell, but instead, it’s changing how we find the people who need it, how we prove we understand their problem, and how fast we can do all of that before they default back to the status quo.

The product-led growth motion feels quite different.

fun stuff (before we get into things…)

We have a lot of family coming over this weekend, so should have some fun updates.
Kids were sick and family was busy, so we don’t have a ton of things to share here today.

Sorry!

where we were four weeks ago

When I wrote the last newsletter about competing against status quo, I was still mostly thinking about CRE funds as our primary segment. That was our first real experiment for the PQS framework: SEC EDGAR filings, BuiltWith scans, identifying funds actively raising capital with no investor communication system.

Good segment.

We have a customer paying ~$500/month there.

But the problem was the data signals were slow. SEC filings have lag. The total addressable market of small CRE funds actively raising at any given time is narrow. And the outreach motion required a lot of manual research per prospect.

So we decided to move on and take a look at the manufacturing space that also currently occupies a good chunk of MRR for Journey.

why manufacturing

Here’s the insight that kicked this off. Steve said it during one of our retros:

“Some things, you don’t need information to buy. Candy, for example. With these guys, the information is essential to the purchase.”

Manufacturing companies sell physical products that are too large, too complex, or too dangerous to demo on a Zoom call. A 50,000 lb food waste digester. Industrial conveyor systems. Custom-fabricated steel components. The buyer can’t see the product until it shows up on a truck. Which means the content — the spec sheets, the renderings, the dimension drawings, the case studies — that content is the product demo.

And right now, the way most of these companies share that content is: email attachments. PDFs. One at a time. Into a thread that will have 50–60 exchanges over a 6+ month sales cycle. By month four, neither the rep nor the prospect can find anything.

This isn’t a theoretical pain point I’m guessing at. Our customer Power Knot sells food waste equipment — literally 50,000 lb machines. Their VP of Marketing wants her sales team to have access to which prospect is actually looking at the specs for the LFC-200 versus who said “thanks, we’ll review” and went dark.

That’s the kind of pain where Journey finally isn’t a vitamin.

three segments, one ICP

We didn’t just pick “manufacturing” as a segment and blast emails. We broke it into three Pain-Qualified Segments, each targeting a different pain and a different buyer persona.

PQS #1: “Complex Product, Can’t Demo on Zoom”

This targets VP of Marketing and Sales Directors at companies selling physical products with extensive technical documentation. The core insight is that when the product can’t be demoed, the content IS the selling. Journey becomes the product demo that a Zoom call can’t deliver.

PQS #2: “Sales Reps Drowning in Email Threads During Long Cycles”

This targets Sales Directors and VPs of Sales at companies with 3–12 month enterprise cycles. Each deal generates 30–60 email threads. Reps can’t find what they sent, prospects reference outdated versions, and leadership has zero visibility into which deals have engaged buyers. Journey’s “single link that grows over time” value prop was built for this exact pain.

PQS #3: “Marketing Team Can’t Enable Sales with Product Content”

This enters through the marketing persona. Marketing creates beautiful product content — renderings, videos, case studies — but the moment it leaves marketing’s hands, it becomes email attachments with ugly file names. This is our path to multi-seat expansion: marketing builds the Journeys, sales distributes them, both get value. And at a fraction of what Highspot or Seismic costs.

the data recipe: how we actually find these companies

This is where the playbook has changed, so I’m learning as I go with people like Steve.

And this is where Layfield’s point about AI reshuffling execution really shows up in practice.

We built a hybrid enrichment pipeline in Clay that combines native integrations with Claygents.

Here’s the actual architecture:

BuiltWith (native Clay integration) scans a company’s website and tells us their tech stack. We’re specifically looking for two things: does this company run a CRM like HubSpot or Salesforce? And do they have a sales enablement tool like Highspot, Seismic, or Showpad? If the answer is “yes CRM, no enablement tool” — that’s our gap. That’s the whitespace Journey fills.

Then we run four Claygents in parallel. One analyzes product pages for complexity — is this a physical product? Is it large or complex? How many product lines? How many downloadable PDFs? Another scans job postings for sales cycle indicators — mentions of “long sales cycles,” “complex sales,” “technical sales.” A third maps the sales team size and hiring activity. A fourth assesses marketing content quality and team structure.

All of this runs automatically.

A company enters the top of the Clay table as a domain name. By the time it reaches the bottom, we know: what they sell, how complex it is, what CRM they use, whether they have a content platform, how big their sales team is, whether they’re hiring, and how much marketing content they’ve produced.

From there, a routing formula assigns each company to a PQS (or disqualifies them entirely). And each PQS routes to a specific message template with variable fields pulled directly from the enrichment data.

A year ago, this pipeline would have required a full-time SDR spending 30–45 minutes per prospect doing manual research.

Now it runs in the background.

the outreach: showing up with the value already delivered

In the last newsletter, I talked about competing against status quo and how the old approach was basically asking prospects to imagine a better future? And how that doesn’t work because status quo bias means they’ll default to “we’re fine”?

For this experiment, the sequence runs through HeyReach on LinkedIn. Day 0 is a blank connection request — no note, because acceptance rates are higher without one. Day 1, after they accept, is the real message. And this message is built entirely from the enrichment data:

We tell them we went through their product pages. We reference the specific number of product lines we found. We note that they’re running HubSpot (or whatever CRM) but don’t have a sales content platform. Then we name the gap: their reps are probably downloading PDFs from the website and emailing them to prospects as attachments.

And then the close isn’t “want to hop on a call?” It’s: “I built a sample of what your top product line would look like as a Journey using your public content. Want me to send it?”

That’s the shift.

what’s actually different about this versus how i used to do it

I want to name the specific things that have changed because it reminds me of the work I still need to do:

The research happens before the message, not during it. In the old world, a rep would google a company, skim their website, and write a semi-personalized email. Now the research is automated and systematic. Every company gets the same depth of analysis.

The segmentation is pain-based, not demographic. We’re not targeting “manufacturing companies with 10–200 employees.” We’re targeting “companies selling physical products too complex to demo, with a CRM but no content platform, and 3+ sales reps.”

The first touch delivers value, not a request. The standard outbound motion is: interrupt someone, describe a problem they might have, ask for 15 minutes. Our motion is: interrupt someone, show them we already understand their specific situation, and offer to give them something useful for free.

Disqualification is built into the system. If a company already uses Highspot or Seismic, they’re automatically disqualified — we don’t compete where there’s an entrenched solution. If they’re not actually a manufacturer (software/services), they’re out. If they’re over 500 employees, they’re outside our pricing sweet spot. This means every message that goes out has already passed through multiple qualification filters. We’re not spraying.

so what?

You’re right. So what… if it doesn’t work.

I’m running this test now, so I’ll share if I find it to net better results than the last segment.

But at the end of the day, right now, these are all just words because I don’t know.

What I do know is that this motion is structurally better than anything we’ve run before. The research is deeper, the targeting is tighter, the message is more specific, and the value delivery is upfront rather than gated behind a demo request.

The thing Layfield wrote that keeps rattling around in my head is his point about arbitrages: every great business opportunity eventually gets competed away. The window for using AI-powered enrichment and personalized outreach at this level of specificity is open right now. It won’t stay open forever.

For me, the question is whether we can use that change to solve our own status quo problem: the one where we know our product works, our best customers love it, but we haven’t been able to find enough people in the exact right situation to hear about it.

More soon.

Let’s keep it going.

Cheers, Danny