The GTM Launchpad is where we open the hood on B2B revenue - fixing structure, sharpening outbound and building pipeline that actually holds up

A few days ago I pulled apart one slice of an outbound motion I’ve been working inside.

From a distance, the activity looked healthy enough

1,995 dials and 111 connects

Then we went one layer deeper.

Only four of those connects had turned into what I would call a legitimate sales conversation and none became meetings.

That is a pretty violent gap and it would be really easy to look at those numbers and start coaching the wrong problem. Tell the rep to improve his opener, work harder on objections, ask for the meeting more confidently maybe even make more calls.

Some of that might eventually matter but before I coach any of it I need to know whether the seller was actually being given enough real opportunities to sell.

A phone system calling something a “connect” does not automatically mean a buyer showed up on the other side. Sometimes it is a wrong person, a personal number, somebody who left the company, or a six second interaction that was dead before the rep finished saying his name.

Technically, those calls connected. Operationally they are nowhere close to the same thing as a sales conversation.

That distinction ended up shaping a lot of what I built this week.

I’ve watched some version of this problem show up inside outbound teams for years.

A rep gets told to prospect, which sounds simple until you watch what prospecting actually requires.

They have to find an account, figure out whether the account is worth working, find the right people, determine which persona matters, check whether the contact data is usable, research the company, look for a reason to reach out, make sure nobody else already owns the account, decide which campaign applies and choose whether to call, email or do something else.

Then, finally, they get to sell.

That is a ridiculous amount of routing logic to put on the person whose highest value job is talking to buyers.

While I was rebuilding part of this motion inside Apollo this week I started simplifying the architecture around four questions:

Who? Which account? Why now? What next?

Who? means the actual people inside the company connected to the problem. Economic buyer, champion, technical stakeholder, gatekeeper. Not 47 titles stuffed into one giant persona because somebody once downloaded an ebook with “Director” in their title.

Which account? is a different decision. Just because I found a CTO does not mean I found a good company to prospect. The organization still needs to fit based on size, market, geography, business model, use case or whatever else makes that company more or less likely to have the problem we solve.

Then comes why now? This is the layer I think gets more important every year. Intent, hiring, technology changes, website activity, a new initiative, an existing tool, a funding event, or some other signal that helps explain why one otherwise good account belongs above another one today.

Finally: what next? Which workflow should this person enter? Which channel makes sense? What needs enrichment? What should be excluded? Is there already an open opportunity or customer relationship? Does the account belong in an automated sequence, a call block or a higher-touch human workflow?

Put those four decisions together and suddenly the seller’s day looks very different.

The rep should not have to rebuild that logic from scratch every morning.

This is getting more important because prospecting tech has become absurdly powerful.

There is no shortage of data anymore.

There are hundreds of millions of contacts available across the platforms we use. We can enrich records through multiple providers, scrape signals, monitor intent and automatically push people into sequences before a rep has finished breakfast.

Technically, I can manufacture a giant amount of activity very quickly. That is exactly why filtering matters more.

I had a separate conversation this week with a founder building his first outbound email motion. We had new domains, new mailboxes, multiple data sources and plenty of possible markets we could attack. The tooling would happily let us start stuffing every possible industry into campaigns.

So we did almost the opposite.

Let the infrastructure warm up. Get clear on the ICP. Decide which few campaigns actually deserve the limited sending capacity then launch.

We only had a handful of mailboxes to work with and that constraint was useful because it forced us to make decisions.

Do we spread them across four random experiments?
Do we run two intentional motions?
Do we put all of them behind one market we actually believe in?

The software makes volume easy.

It does not mean volume should make the decision.

The same principle showed up on the call coaching side.

I rebuilt part of the Trellus environment this week because I didn’t want every logged connect treated like a coachable sales conversation otherwise we create a nasty measurement problem.

Imagine a rep gets ten connects

Six are wrong people or bad records
Two conversations end almost instantly
Two become legitimate buyer interactions

If an AI scorecard grades all ten for discovery, listening, problem fluency and next-step control, the score may look scientific but it is garbage. The seller never had an opportunity to demonstrate those skills on most of the calls.

So the first job of the coaching system became classification:

did an actual sales conversation exist?

If not, capture the data problem. Do not pretend it is automatically a rep skill problem.

If yes, now we have something worth coaching

and when we did have legitimate conversations this week, the coaching became much simpler.

The behavior I kept coming back to with the rep was: ask one more question before you explain

Someone says the problem is not a priority? Do not immediately launch into why it should be. Understand the condition.

Someone says another team owns it? Find out who

Someone already has a solution? Understand what they are doing today, why they chose it and what is working or not working

Diagnose before explaining

The tooling can surface the moment
The scorecard can help us find the pattern

But the actual behavior still has to be simple enough for a seller to remember when another human picks up the phone

This is probably the biggest shift in how I’m thinking about outbound infrastructure right now.

I do not want the rep spending half the morning operating a miniature research department.

I want the machine doing the work around the selling moment: building the relevant account universe, prioritizing it, mapping personas, surfacing signals, cleaning the data, removing duplicates, excluding customers and active opportunities and routing the right people into the right workflows.

Then hand the human a much better starting point

Not “Here are 4,000 contacts. Good luck”

More like “Here are the twenty five people most worth your attention today, here is why they are here, here is what we already know, here is the next best place to start”

That is a very different sales experience and I think it becomes even more important as AI increases the amount of information available to sellers.

The rep does not need another ten tabs. They need fewer decisions that a system could have made safely before they ever showed up.

We still frame a lot of the AI-in-sales conversation around one question

How much of the seller can we automate?

I’m becoming more interested in a different one

How much friction can we remove before the seller reaches the part only a seller can do well?

Research, routing, enrichment, prioritization and preparation are exactly where I want the system to be relentless.

Once another human answers, the job changes.

Listen, adapt, understand the problem and ask the next question because of what the buyer said not because line seven of the script told you to.

That is the part I want sellers spending more of their time doing.

If we build the operating environment correctly we can create more of those moments without forcing the rep to become the operating system themselves.

Open whatever your SDR uses first thing Monday morning and pick one prospect sitting in front of them.

Can the rep answer these four questions in about ten seconds?

Who is this person?
Why does this account fit?
Why might this matter now?
What should I do next?

If answering those requires opening six tools, searching LinkedIn, asking RevOps, rebuilding the account history and guessing which campaign applies, the seller is doing routing work.

That does not automatically mean you need another AI tool. It does mean your operating system still has decisions living in the wrong place.

The goal is not to automate the seller.

It is to make sure the seller arrives at the human moment with everything they should already know.

System finds the work
System prepares the work
Seller does the selling

That is the outbound machine I want

By the time this lands in your inbox today I’ll probably be somewhere between Arizona and Dallas.

I’m heading out for the TitanX event and somehow Tuesday’s agenda includes playing flag football on the Dallas Cowboys field 😂

I’ve never been to Dallas and I’ve actually never been to Texas at all so I’m looking forward to getting in the room, meeting more operators and inevitably coming home with another notebook full of things to tear apart in next Monday’s issue.

If you’re around the event come say what’s up! And if I pull a hamstring trying to relive my athletic prime on an NFL field we simply will not discuss it here next week 😂

Have a great Monday 🤙

Thanks for reading. See you again next week at 8 am MST!

- Tom Slocum ✌️

Open the hood with me →

Know a founder, sales leader or operator who’d get value from this field note?