Is cold calling back? Outbound when every inbox is full of AI

For about ten years, the advice to SaaS founders was simple: nobody picks up the phone anymore, so send email. Then AI made a well-written, "personalized" cold email cost nothing to produce. Everyone started sending them. Inboxes filled up, mailbox providers tightened the rules, and buyers learned to ignore anything that looks like outreach.

Meanwhile, the phone didn't get cheaper in the same way. A real person who has done their homework and calls at the right moment is rare again. So yes, I think cold calling is back. But not the version from 2010, with a rep dialing a list of 300 names a day.

What changed

Three things happened at the same time:

1. Writing got free. An AI tool can research a prospect and write a plausible first line in seconds. That's great for the sender and terrible for the channel, because when everyone's email is "personalized", none of it feels personal.
2. Mailbox providers pushed back. In 2024 Gmail and Yahoo started requiring bulk senders to authenticate their email, offer one-click unsubscribe, and keep spam complaint rates low. Google's sender guidelines ask senders to keep reported spam rates below 0.3% (the announcement is from October 2023). Volume-based cold email got a lot harder to do well.
3. Buyers got better at filtering. Most of us now scan an inbox with a trained eye for templates. The bar for getting a reply is higher than it's ever been.

The result is that a thoughtful phone call has become a strong signal. It says: a person decided this conversation was worth their time.

The modern playbook

The version that works now looks more like careful account work than a call center.

Call on signals, not lists. A new funding round, a job posting for the role you help, a switch in their tech stack, a pricing page that just changed. For product-led companies, the best signal is your own product: a trial user who invited their team, hit a limit or connected their billing system. Trial insights show you where those trials are and how they convert. Those people are already interested. A call from a human who can help often converts better than any email sequence.

Let AI do the research, not the conversation. This is where AI is genuinely useful: summarizing a company, finding the right person, pulling the three facts that make your call relevant. Five minutes of research used to take thirty. Use the time you save to make fewer, better calls.

Open with a reason. "I saw you're hiring your first RevOps person, and we help teams set up revenue reporting before that hire starts" beats any clever opener. If you can't write the reason in one sentence, you shouldn't be making the call.

Keep the list small. A tight ideal customer profile and fifty well-chosen accounts will beat five hundred loosely matched ones. Your reputation is part of the channel too.

Follow up briefly. After the call, send a short, plain email that references the conversation. This is the one email that gets opened, because the person now knows who you are.

Respect the rules. Calling rules differ by country. In parts of Europe even B2B calls can require prior consent, and many countries have do-not-call registers. Check before you dial.

When it doesn't work

Outbound has a cost, and the math has to work. A rough, illustrative example: if a salesperson costs $8,000 a month fully loaded and closes four new customers a month at $300 MRR each, that's $1,200 of new MRR a month. At an 80% gross margin, it takes about eight months of that rep's new MRR to pay back one month of their cost. That's workable. At $30 MRR per customer, it isn't.

So as a rule of thumb:

• Low-priced self-serve products shouldn't cold call new logos. Use the phone for trial conversion and expansion instead, where the signals are strongest.
• Mid-market and up can make outbound work, if calls are signal-driven and the list is tight.
• Founders should make the first hundred calls themselves. Not because it scales, but because nothing teaches you faster what your market actually cares about.

Measure pipeline, not dials

The trap with outbound is measuring activity. Calls made and emails sent tell you how busy the team is, not whether the channel works. What matters is what comes out at the end:

• Meetings and opportunities from each channel.
• New MRR by channel. Tag new customers with how they were acquired (for example with a metadata field in Stripe) so you can see MRR from outbound next to MRR from inbound and product-led signups.
• CAC payback by channel. How many months of gross margin it takes to earn back what you spent.
• Retention by channel. Customers you talked into buying sometimes churn faster than customers who came to you. Compare their cohorts before you scale the channel.

GrowPanel pulls custom fields like acquisition channel from your billing data and lets you break MRR, churn and cohorts down by them, so you can see which channel brings customers who stay. RevOps teams use it for exactly this (here's how).

Cold calling isn't a silver bullet, and it never was. But in a world where every inbox is full of AI, a well-prepared human voice stands out again. Use that while it lasts.

If you want to see your revenue by channel, explore the live demo or start for free.