Think about the week after a Series B announcement. The new VP of Sales opens their inbox and finds the same email twenty times. Different logos, same opening line: "Congrats on the raise."
I've sent some of those emails. Most of us have.
Funding rounds, new hires, job changes, tech installs, office openings. A few years ago, spotting these early was an edge. Today it's a subscription. If my competitor pays for the same data providers I do, we both see the same trigger on the same morning, and we both write to the same person before lunch.
Kyle Poyar described this better than anyone last month in Growth Unhinged. His point: third-party intent data has been commoditised, and the plays that last pair it with first-party data that is genuinely yours. When I read it, I recognised our own motion almost step for step. I've designed our business development and sales motion around these plays.
So this isn't a summary of his piece. It's what I've learned running these plays, and what I think comes after them.
When, what, whether.
Every signal-led play has to answer three questions. Most of the tooling is built for the first two.
Three questions every play answers
-
WHEN
Rented data
Do I reach out?
Third-party signals: a job change, a funding round, a new tech install.
-
WHAT
Owned memory
Do I say?
First-party data: what this champion achieved with us, why this deal was lost.
-
WHETHER
Judgment
Should I reach out at all?
Rules a human writes down: is this the right moment for them, or just for my pipeline?
Rented data is the commodity. Owned memory isn't. Nobody else has my call recordings, my closed-lost notes or my product usage data. I've started calling this the memory advantage.
When everyone can buy the same timing, the edge goes to whoever remembers the buyer best.
The third question, whether, is the one no tool asks for you. I'll come back to it, because I think it's about to matter more than the other two.
The stack I've been testing these plays on.
So you can map this to your own setup:
- SalesforceSystem of record.
- ClayEnrichment, audiences and orchestration, including its Account Agent for reading deal context.
- HG Insights + DemandbaseTechnographic and company signals.
- Factors.aiIntent, and identifying which accounts are on our website.
- Influ2Reaching the buying group at target accounts with person-level ads.
- OutreachSequences.
Four plays, and what running them taught me.
-
The champion who moved
OwnedWho championed or signed for us, and what they achievedRentedA job changeThe instinct is to reach out the day the announcement goes up. I don't. I wait about a month.
In week one, a new leader is buried in onboarding, introductions and congratulations. Your message is one of fifty. By week five, they're writing their first plan and looking for things they know already work. That's when a message from someone who helped them before actually lands.
The message never says "congrats on the new role." It reminds them what they achieved with us, in their own numbers, and asks whether the same problem exists in the new place. I don't stop at the main champion either. Power users get promoted. People who sat next to the product and watched it work remember it too.
Whether checkMemory cuts both ways. If a champion's last project with us ended badly, the job change isn't an opening. Before this play fires, I want to know how the relationship ended, not just that it existed.
-
The deal that wasn't dead
OwnedWhy the deal was lost: calls, emails, Salesforce notesRentedA sign the blocker has movedThis is the play I'd start with if I had to pick one. Poyar points out that roughly one in four qualified opportunities becomes a customer. That means most of a team's warmest prospects are sitting in closed-lost, and most teams reactivate them with a calendar reminder and a "just checking in."
Deals rarely die because the buyer didn't need you. They stall on timing, budget or priority, and all three change. So I let Clay's Account Agent read the loss context and match it to the signal that would undo it:
- Lost on budget: wait for funding or a new CFO.
- Lost on priority: wait for job posts that show the initiative is now staffed.
- Lost to a competitor: watch for a tech change.
If there's no matching signal, there's no outreach.
Whether checkNot every signal means "now." A deal lost to a competitor six weeks ago, plus a tech change, probably means they're implementing the competitor. Some signals mean "not yet," and the play has to know the difference.
-
The visit that means a buying group
OwnedWhich pages a company visited, and how oftenRentedA target account with a known buying groupRetargeting everyone who visits your website is mostly paying to follow students, job seekers and competitors around the internet. So the play works the other way round. Factors.ai identifies the accounts behind high-value page visits. I check them against our tiered target list, and only qualified accounts move forward. For those, Influ2 puts our message in front of the actual buying group, not the anonymous visitor. And the account owner gets an alert with what was viewed and when.
The learning that matters here is which behaviour actually predicts a deal. It's rarely the obvious one. Poyar's example: three visits to documentation can say more than one visit to the pricing page. Your patterns will be your own. The only way to find them is to compare what accounts did before they became real opportunities.
Whether check"I noticed three people from your team on our pricing page" is accurate, and it's unsettling. I use visit data to decide who to reach and when. I don't quote it back to the buyer.
-
The users sales never saw
OwnedProduct usage, including personal-email signupsRentedCompany fit, hiring, buying groupIn AI products, a big share of signups use a personal email address. Poyar estimates 75–90% in his experience, and sales usually filters them out as noise. That's a mistake. Some of those people are evaluating you for their team without telling anyone yet.
So we match them to their companies and roll usage up to the account. Clay qualifies the account, the rep gets the full usage context, and the account goes into an Outreach sequence. These are the triggers I care about:
- several active users at the same company who don't know about each other
- users hitting product limits
- a director-level user
- a usage spike in the same week as a pricing-page visit
Whether checkSomeone who signed up with a personal email may have chosen not to tell their employer yet. Connecting them to their work identity is a decision with consequences, which brings me to the part of this I think we're underestimating.
What happens when everyone runs these plays.
Bought signals became a commodity because everyone could buy them. I think the plays built on top of them are next.
When every vendor in a category runs champion tracking, a good champion doesn't get one thoughtful note in week five. They get ten. When everyone reactivates closed-lost deals on a leadership change, the new leader's first month is full of the same trigger from different senders. The play still technically works. It just stops standing out.
The plays can be copied. Two things can't:
- 01
How deep your memory goes
"They were a customer" is not memory. "They cut their team's onboarding time, and their CFO signed off because the payback was under a year" is. That detail only exists if reps capture it and the system keeps it.
- 02
How fast you learn
Every play produces evidence about which signal, timing and message worked for which kind of buyer. A team that feeds that back every week will keep pulling ahead of a team running the same play from a template.
Three ways memory becomes a liability.
- Dirty memoryAn agent working from a stale Salesforce record personalises confidently and wrongly. It's the mirror effect from Signal 02: an agent scales whatever data it inherits.
- Creepy memoryPrecision can feel like surveillance. Knowing something doesn't mean you should say it. Use the data to change who you contact and when, not to recite it.
- Unlawful memoryIdentifying visitors and matching personal emails to work identities means processing personal data. Build consent in now, not after the first complaint.
Where this goes next.
Today, people design the plays and decide which data matters. Poyar ends his piece with the obvious next step: agents that read every signal, spot the patterns and propose new plays on their own.
I think there's a step after that. Sangeet Paul Choudary argues that most agents today work for one side of a deal. A buyer's agent picks from offers sellers have already fixed. Markets change when agents on both sides can express what they need and what they'll trade, and put together deals that didn't exist before.
If that happens, owned memory stops being just an input for outreach. It becomes what your agent brings to the table when the buyer's agent asks: what did customers like this achieve? What did they need? What would make this work?
Teams that have been capturing that memory carefully will have real answers. Everyone else will have a list of triggers.
How I measure it.
I treat each play like a channel and give it one quarter to prove itself against a baseline. If it can't beat the baseline in a quarter, it isn't a motion yet. It's an experiment.
- PipelineSourced or influenced by the play.
- Win rateSignal-triggered opportunities compared with cold outbound.
- Cost per opportunityIncluding data, enrichment and tooling.
- Negative replies per 100 touchesHow I measure "whether." If the relevance is real, this should fall as volume falls.
- Memory freshnessThe share of target accounts whose champion and loss-reason fields were updated in the last 90 days.
The last two aren't in most playbooks. I think they're the ones that will separate teams a year from now.
Sources (5)
- Kyle Poyar, "The new battleground in GTM". Growth Unhinged, 20 Sep 2026.
- Sangeet Paul Choudary, "How agents unlock new markets". 11 Oct 2026.
- Sangeet Paul Choudary, "Agents, and their better half". 4 Oct 2026.
- Sangeet Paul Choudary, "Who Owns the Customer in the Age of AI Agents?". Harvard Business Review, 9 Oct 2026.
- Hogan Lovells, "India's Digital Personal Data Protection Act 2023 brought into force". DPDP Rules notified 13 Nov 2025; notice and consent obligations from 13 May 2027.
Navneet Kumar
Building Neolism in the open.