.png)
Early go-to-market efforts come with full stacks, active campaigns, and a steady flow of activity. Clarity around targeting, a sharp understanding of the problem, and consistent use of market signs in decision-making should receive the same level of attention, as they shape outcomes from the start.
In this conversation, Eoin McGuinness, Head of GTM at Instantly.ai, draws from experience across startups, sales leadership, and thousands of GTM setups to examine where teams drift off course. The focus is on how ICPs are defined and refined, how feedback is interpreted, and how those inputs translate into decisions that can sustain growth.
He is a 3x exited founder, with his first exit at 22, which gave him early exposure to building and selling in resource-constrained environments. He then moved into sales leadership, where he led a 4x increase in ARR over 14 months while expanding into 12 international markets, navigating the shift from early traction to structured scaling.
A significant part of his perspective comes from his time at HubSpot, where over eight years he worked on GTM strategies with more than 10,000 startups.
Tools are commoditized. The differentiator now is ICP precision and speed of iteration. Anyone can send 10,000 emails this afternoon. Very few teams can articulate why a specific list of 200 accounts should care about their offer this quarter. That's the work that's still hard, still manual, and still almost never done well.
The second piece is iteration cadence. Teams that ship a new sequence every week beat teams that perfect one over a month. The market gives you the answer, but most teams aren't shipping fast enough to hear it.
Founders confuse activity with motion. They have HubSpot, they have Calendly, someone's running campaigns, so they "have a GTM." But there's no documented ICP. No offer hypothesis being tested. No log of what got replies and what didn't. The market is telling them something every day through reply rates, no-shows, and the questions prospects ask on calls. They're not reading any of it. The disconnect is almost always this: founders think they're targeting a market. They're actually targeting a job title.
The tell is when reply rates are flat or falling, but volume keeps going up to compensate. Founders read "we sent 50% more this month" as progress. It's actually the early sign of brand erosion, domain reputation cratering, ICP fatigue setting in, and copy that worked in Q1 burning out by Q3. If your meeting count is holding steady only because you tripled send volume, outbound has stopped being a lever. You've started laundering bad efficiency through bigger numbers. Cut volume, fix the offer, restart smaller.
Months 1–3: prove offer-market resonance with manual outbound. 50 accounts, hand-picked, founder doing the sending. If you can't book conversations there, automation makes nothing better, just faster failure.
Months 4–6: tighten ICP based on who actually said yes, not who you thought would.
Months 7–9: build the infrastructure, domains, sequences, light enrichment, and follow-up logic.
Months 10–12: scale, but only the things that worked.
Most founders flip this. They invest in tooling and SDR hires in month two, before there's anything proven worth scaling. Then they spend months six through twelve trying to figure out why the engine isn't producing.
Replies are vanity. Booked calls are vanity if half of them are no-shows.
The signals that actually matter: prospects describing their pain back to you in your own language, which means the messaging is landing. Inbound trickling in from people in your ICP who heard about you secondhand, that means the market is talking. Pull-through from meetings to opportunities above 25%, which means the targeting is right. Existing customers asking for the next feature before you've shipped it, that means you've found a real problem.
If none of those are present and you're still celebrating reply rates, you don't have traction. You have noise.
Messaging gets 80% of the time. Targeting gets 20%. It should be the other way round. The best-written email to the wrong person fails. A mediocre email to a precisely-defined account converts. Teams polish copy because it feels productive and creative. Defining ICP feels boring and conceptual, so it gets skipped or done in a single afternoon and never revisited. The other big one: process. Founders build elaborate workflows before they've validated whether the workflow does anything. Process is what you formalise after something works, not before.
Pricing and proof. EU and UK founders often show up with €/£ pricing that signals "small" to a US buyer. The case studies are European brands that don't move the needle in a Texas pipeline review. The AE on the call is pattern-matched as junior due to accent or pacing. The send times are wrong; 9 am London is 4 am Eastern. The motion that worked at home doesn't translate. The fix isn't to send more, it's to rebuild the proof stack for the US buyer specifically before you scale spend.
Genuine improvement: research at scale, knowing which 200 accounts to prioritise this week based on signals like funding, hiring, or product launches. Summarising calls so the follow-up actually reflects what was said. Generating sequence variants to test in parallel.
Bad practice, accelerated: surface-level personalisation. "I see you went to X University, hope you're enjoying Boston." Every prospect clocks it inside two seconds now. The volume of low-effort AI outbound has roughly 10x'd in the last 18 months and reply rates across the board are getting hammered as a result.
AI is a force multiplier. If your fundamentals are weak, it multiplies the weakness.
Three places, every time.
One: stop building before you've sold. If you can't sell the wireframe, you can't sell the product. Get five paying customers committed to a problem before you commit a developer to a solution.
Two: ICP is not a job title. It's a behaviour, a stage in a company's life, a specific trigger that makes your offer urgent. Most founders' ICP doc reads like a LinkedIn search filter. That's not an ICP, that's a contact list.
Three: founder-led sales are non-negotiable for the first 25 customers. You cannot outsource the learning. Every founder who hires a salesperson at customer five regrets it by customer fifteen, they've paid someone else to develop the muscle memory the business needed them to build.
