Brendan Hughes booked this one expecting a fight. Two founders who sell competing answers to the same problem, one episode title asking whether intent data is dead, and a subtitle promising that two signal companies would argue it out. What he got was Jack Porter of WhiteWhale and MarketSizer's Niall O'Gorman agreeing on almost everything, which Niall eventually had to call out on air: "again, we're meant to be fighting here."
The agreement turned out to be the interesting part. When two people building in different directions land on the same enemy, the enemy is probably real. Here is what came out of Episode 3 of Intent, Decoded.
The Supply of Keyword Intent Is Drying Up
Jack's framing of the category is the cleanest version of it I have heard. There are three generations of trying to predict which account is worth a rep's morning.
The first is what most people mean when they say intent data: this company searched this keyword. It is the model behind the familiar names, and it works by watching someone from a company IP address land on a page in a content network. The problem is structural rather than commercial. Buyers increasingly do that research inside a model instead of a search engine, and a model reading a blog post on your behalf does not leave the trace the whole method depends on. The supply is thinning out. That is a different failure from "the data was never any good", and it is not one a vendor can fix with better matching.
The second generation is the response to that: the pivot from intent to signals. Funding rounds, hiring, job changes, champion tracking. Useful events, and genuinely better than a keyword surge. But they are also the easiest signals to build, which is why they are now in every tool on the market.
Everyone Has the Same Signals, So Everyone Calls the Same Day
This is where Jack landed the line of the episode. "Everyone with a Clay table, everyone with an Apollo account, everyone with a ZoomInfo account emails the exact same prospect the exact same day at the exact same time. And there is no worse time to reach out than when everyone else is calling."
It is worth sitting with, because it inverts how most teams think about signal quality. A commoditised signal is not merely unhelpful. It is actively expensive. It concentrates the entire market's outbound onto the same accounts in the same 48 hours, which means a signal every tool ships is a reliable predictor of a crowded inbox rather than of a receptive buyer. The funding announcement that told you to call is the same announcement that told forty other vendors to call.
Niall's version of the same complaint comes at it from the data side. Aggregated datasets get refreshed on a vendor's schedule, not the buyer's: weekly for the large sites, every six months for the small ones. "That is not helpful, because the data is going to be stale before it is even received." Then the current fashion is to point a language model at that stale data, which adds fabrication on top of being out of date. His phrase for the result: the wild west.
So the shared enemy is not intent. It is static data that everyone owns and nobody can check. We have made the longer version of that argument in intent data didn't break, the wrong signals were sold as intent.
Work Backwards From the Buying Moment
Jack's answer to the commoditisation problem is to stop starting from the signal. Start from the moment.
His worked example: you sell a customer experience or customer support product. When does a company actually buy one of those? Not continuously. There is a tipping point where the pain of doing support manually becomes bigger than the hassle of buying software. Maybe it is crossing a thousand customers. Maybe it is hiring the first customer success person. Maybe it is landing an enterprise deal large enough to eat the team's week.
Once you can describe that moment, you can work backwards to the observable events that indicate an account has arrived at it. A first CX hire. An expansion into a new market. A deal announcement that changes the support load. None of those are intent in the classic sense. All of them are evidence that the pain is rising.
The exercise is free, and it is the part most teams skip. Write down the single sentence that describes when a company becomes a buyer of your product. If you cannot write it, no signal feed will supply it.
What the Account Does, Not What Is Said About It
This is where the two companies genuinely diverge, and it is a difference in the question being asked rather than a disagreement about the answer.
WhiteWhale watches what is said about an account: news, hiring, filings, earnings calls, company updates, read back with the quote and the source attached. MarketSizer watches what the account does with software. Niall's side of it is subscription movement: when a technology was first seen on an account, when it was last seen, and what the gap between those two dates implies. An active trial looks different from a churned trial. A trial that flipped to a subscription looks different again. From there you can place an account on a lifecycle - new, developing, established, loyal - and the placement changes the play.
His example of why that matters is the most practical thirty seconds in the episode. "If somebody has just renewed with a competitor, you put them into nurturing until you are 120 days out from the next renewal. If your own customer starts implementing a named competitor, you pick up the phone as fast as humanly possible." Two accounts, both showing competitor activity, opposite responses. A list that treats both as "high intent" has destroyed the only information that mattered.
Both of those are still evidence rather than inference, which is the thing they share. The detection side of it is covered in how to detect competitor trials, and the renewal-window play in the B2B winback deep dive.
Timing Is a Spectrum, Not a Binary
Brendan pushed on this and got the reframe the episode needed. Most tools present timing as a one or a zero: this account is hot, go. Jack's objection is that nobody buys that way. "Practically, it is not a binary. That is not how you and I buy things. We think about things, we sleep on it, we wait six months, we forget about it, and then we need to go buy it."
The honest version runs from an account with no competing technology in the stack at all, through an account evaluating, to an account fully committed to a competitor where any deal is a rip and replace. Each position calls for a different move. An account already locked into a competitor gets an air-cover warm-up: education on the differences, run patiently against a renewal date you can approximate. An account building out the function right now, having just put real salary behind it, is go time.
Same target list, two completely different motions. A tool that only tells you "hot" cannot tell you which one to run. That is the argument we made from the buyer's side in the winnable account is warm, not hot.
The Signal Is Half the Job. Routing Is the Other Half.
Niall's contribution here is the one most likely to be ignored and the cheapest to fix. Every signal has a correct recipient, and the default of "send it to the SDR" is wrong most of the time.
A named competitor being trialled inside an account where you are the incumbent is not an SDR signal. It goes to the CSM, the account manager, or whoever owns retention, and it goes there today. An account doing early-stage category research is not a sales signal at all; it is a marketing signal, and it is not even a qualified lead yet. Sending it anywhere near a rep guarantees a premature pitch to someone who has not decided they have a problem.
His framing of the work order is deliberately unglamorous: most of the value is extracted before any signals arrive. What are you trying to achieve, who are the personas, how do you identify them, in what format does the signal need to surface, which system does it need to land in, and who acts on it. That is a whiteboard exercise, not a purchase. The implementation detail for CRM-side routing is in how to build a signal-led GTM motion.
No Source, No Belief
Both founders had the same answer to the reliability question, which tells you something about where the category is heading.
Jack: "A signals business does not exist unless the data is reliable. As soon as that data becomes unreliable, people stop trusting it and they stop listening to it." He traces the collapse of trust in first-generation intent to a specific experience every rep has had: you call the account the platform flagged, and the person on the phone has no idea what you are talking about. Technically someone at that IP address did search the term. That fact is useless to the person dialling. So WhiteWhale attaches the source to every signal, whether it is a job post, a press release or a news article, and the rep can read it before they act.
Niall's version is the timeline. Nobody believes a subscription record until they can click into it and see when the technology was first detected, when it was last detected, when the competitor went live. The Chrome extension adds a second layer: a live detection on the site you are looking at, checked against the historical record, so the aggregate claim and the current reality are verified against each other.
The same idea shows up in how both companies price. Jack's is month to month with an explicit "if the data is bad, cancel", which he contrasts with a multi-year contract paid up front. Niall went one further: when MarketSizer tries to enrich something and gets it wrong, the credit is returned automatically with an audit trail recording the miss, before the customer has to ask. Admitting a failure in public is an odd growth strategy. It is also the fastest route to being believed by someone who has been sold unverifiable data before.
Why Not Just Build This Yourself?
Brendan asked the question every founder in this category gets asked weekly: I have a model, a few connectors and an afternoon. Why do I need either of you?
Jack's answer is operational rather than technical. You can ask a model about a thousand accounts this morning. Will you do it tomorrow morning? And the morning after? How do you know what changed between runs? The manual version does not survive contact with a Tuesday, and the automated version burns through more in credits than either product costs. In practice, he said, most of their customers do both: take the structured data in, and use the model for the writing, the call prep and the research on top.
Niall, who is candid about using these tools daily, made the guardrails argument. However the data is produced, it has to come back in a predictable shape. If an SDR has to work out how to read the answer every single time it arrives, in a new format, with different charts, the efficiency gain has gone. You can build that consistency yourself if you are willing to maintain what you build. Most organisations decide that the maintenance is not their core business.
His closing image for it was a strip mall with two barbers. One sign reads six dollar haircuts. The other reads we fix six dollar haircuts.
Signals Multiply. They Do Not Rescue.
The closing advice is the part to take to a planning meeting, and both founders arrived at it independently.
Jack: signals and intent data are a multiplier. Point one at a working outbound motion and it gets substantially better - his example was a team going from five meetings a week to fifteen, because the targeting finally matched the message. Point one at a motion that is not working and nothing happens. "If you multiply anything by zero, it is still zero. Intent data is not going to save a bad offer." The three things that have to be solved first are the offer, the infrastructure and the pitch. A team arriving at a signal product as a life raft, having decided outbound is broken, is close to the worst-case customer, and he tells them so.
Niall's version starts even further upstream: rubbish in, rubbish out, and master data is the piece nobody wants to own. His practical instruction for anyone overwhelmed by the tooling is to turn the computer off, take a sheet of paper, and draw the customer journey. Left is where they start, right is where they finish, and the messy middle is the actual work. Do that first, and the tools stop being a menu of options and become a set of answers to questions you have already asked. Go straight to a model instead and you get an impressive plan you had no part in creating, do not understand, and cannot repeat.
Which lands them both in the same place: it comes back to the ICP. What do we sell, who do we sell it to, and what is the value proposition. Nothing layered on top works without that, whether it is signals, intent, or a coaching platform.
Brendan's summary was more concise. Get the basics right, then add the magic beans.
Intent, Decoded is MarketSizer's podcast series on the mechanics of purchase intent and what it actually takes to use it well. Watch Episode 3 on YouTube, listen on Spotify, or catch the full LinkedIn Live recording here. Follow the show on Spotify or YouTube to catch future episodes, or visit the podcast hub for everything in one place.
Recommended reading
- Episode 3: Intent Data Is Dead. Or Is It? - the full conversation, with chapters, quotes and the video.
- Intent Data Didn't Break. The Wrong Signals Were Sold as Intent. - the long-form version of the shared-enemy argument in this episode.
- What Replaces Intent Data in 2026? - the compact answer to the question the episode title asks.
- How to Detect Competitor Trials - the detection methods behind the first seen, last seen timeline.
- B2B Winback: Recover Churned Customers on Competitor Renewal Timing - the deep dive on the 120-days-out renewal play.
- Purchase Intent vs Intent Data - the definitional pillar behind the split between what is said and what is done.
- The Winnable Account Is Warm, Not Hot - why the hottest account is rarely the most winnable one.
- From Signal to Trust - the Episode 2 recap with Leslie Venetz, on earning the right to a buyer's time.
- Most GTM Teams Use Intent Data at the Wrong Time - the Episode 1 recap this conversation builds on.
- Intent, Decoded - the podcast hub - every episode in one place.