We use cookies. Read Privacy Policy
Ok

Cold Email Lessons From Lead Gen Jay: Live Q&A Recap

Webinar Recaps
Blog > Webinar Recaps > Cold Email Lessons From Lead Gen Jay: Live Q&A Recap
Vlad Mae | Published: | min read | reads
On November 17, Jay Feldman, better known as Lead Gen Jay, joined us for an hour of live, unscripted questions. No slides, no pitch. People brought real problems from their campaigns, and Jay answered with what's working for him and his clients right now.

We pulled the most useful parts into one article. If you run cold email for your own company or for clients, here's what you need to know.
Recap of our live Q&A with Jay Feldman, hosted by Nikita Bykadarov, CEO of Maildoso.

Nikita runs Maildoso, a cold email infrastructure platform, and talks with agency owners and outbound teams every day. In this session he put the questions our community cares about most to one of the best-known people in cold email.

Key Takeaways

  • Spread mailboxes across domains to isolate risk. Scale by adding mailboxes, not by sending more from each one.
  • Inbox placement tests can lie for Microsoft. Aged domains and short copy are Jay's fix.
  • Clean bad-fit leads out of your list before you send. In a typical export, around 4 in 10 aren't a real fit.
  • The highest reply rates come from lead data nobody else is using and from emails that don't sell anything.
  • Agency clients stay when you make them money. If their offer is weak, fix the offer first.
  • AI reply agents work best when they audit their own work and get checked by a second model.

Who is Jay Feldman?

Jay Feldman is the founder of Lead Gen Jay, a cold email agency and education brand, and co-founder of Otter PR. He went to medical school before switching to business, and he's been building cold email systems for about 10 years.

Otter PR is where he first got good at cold email. Today it has around 60 staff and two headquarters, and it made the Inc. 5000 list. Lead Gen Jay has about 20 people and runs cold email for clients. Jay also runs Consulti AI, a lead data platform, and teaches cold email to 91K+ subscribers on his YouTube channel.

Cold email infrastructure and deliverability

Most questions on the call were about infrastructure, so we'll start there.

How many mailboxes per domain?

Jay's starting rule for Google mailboxes is 3 to 5 per domain. The number that matters most is how many emails each domain sends per day, not how many mailboxes sit on it.

The reason is risk. If you're new, some people will report you as spam, and that hurts the whole domain. When a burned domain has 3 mailboxes, you lose 3. When it has 10, you lose 10.

One attendee had already set up 20 mailboxes on 2 domains. Jay's advice was to keep them and lower the volume per mailbox, from 30 to 50 emails a day down to 15 to 20.

Nikita added how Maildoso users usually split it. With around 30 mailboxes, stay at 4 to 6 per domain so you always have spare domains. With 1,000+ mailboxes, 50 to 100 per domain is workable. Either way, keep your warm-up volume higher than your cold volume. Deliverability usually drops when cold emails start to outnumber warm-up emails.

Both agreed on the core idea: scale horizontally. You send more by adding mailboxes and domains, not by pushing more emails through each one.

Run daily inbox placement tests

For teams managing hundreds of mailboxes, Jay uses two automated rules in his sequencer:
  1. Run an inbox placement test on every mailbox every day.
  2. If a mailbox scores below 90%, pull it out of sending and warm it up for 2 weeks. Once it's back above 90%, put it back in rotation.
He runs this setup for Lead Gen Jay, Otter PR and all his clients, and says it's the simplest way to keep deliverability high at scale.

Why Microsoft inboxes are harder

This was one of the most useful points of the session. According to Jay, inbox placement tests often lie for Microsoft. The email lands in the inbox, the test reports success, and a few seconds later Microsoft quietly moves it to junk. You see a clean test and almost no replies.

His fix has three parts:
  • Aged domains. Jay buys domains with 20 to 30 years of history at auctions. Domain age matters a lot for Microsoft reputation.
  • Solid mailboxes. Google or Microsoft both work. Neither is clearly better.
  • Short copy with lots of variation. Use heavy spintax or fully AI-generated emails so no two messages look the same.
On aged domains, he'll go up to about 5 Google mailboxes per domain.

Security gateways and enterprise leads

Nikita added a point for anyone targeting large companies. Many enterprises route mail through security gateways before it reaches Microsoft or Google filters. Your warm-up network doesn't include those gateways, so to them you look like a sender with no warm-up at all.

Some agencies handle this by using separate domains for gateway-heavy lists and replacing them as soon as replies stop. Jay's approach is simpler: he mostly targets companies under 100 employees, so gateways aren't a big enough problem for him to plan around.

Don't cut corners on tools

Jay was blunt about cheap alternatives. Cheap verifiers, lead databases, sequencers and mailboxes are tempting, but any one of them can wreck a campaign. His advice is to use tools that people with real results actually use.

That doesn't mean overpaying, though. One attendee was paying about $9 to $10 per mailbox through Google Workspace directly. Jay pointed out that infrastructure providers like Maildoso offer the same kind of Google mailboxes for cold email at a fraction of that price. You keep ownership of your domain either way, since the provider only needs access to its DNS records.

Lead lists and data

Remove bad-fit leads before you send

Jay's favorite quick win is a step most people skip. Export a list from a standard database, go through it one by one, and you'll often find that about 4 in 10 leads aren't a real fit for your offer. Most senders email them anyway, and those are the people who mark you as spam.

Add a step between the export and the campaign. Use AI, Clay or an automation tool to remove anyone who's a bad fit and enrich the ones who are a good fit. One attendee sells to authors and found that mentioning each author's book title made a big difference. Jay suggested using AI research to find missing titles instead of dropping those leads.

Use data nobody else is using

Most B2B databases sell the same people, usually pulled from LinkedIn. Decision makers at companies under 50 employees get hit by everyone's campaigns, so they rarely reply.

Jay's advice is to step outside the usual databases. Scrape local businesses from Google Maps, creators from YouTube or Instagram, event lists, review sites like Clutch or G2. As he put it, the harder you have to work to find someone's email, the better the chance they'll reply.

Can you email the same list again?

Yes, with two conditions. Wait about a month after the last sequence, since most people won't remember you by then. And make the new sequence clearly different, because people who ignored the first one will likely ignore a similar second one.

Jay's take on lead data tools

Jay tests API-only lead databases head to head regularly. In his latest test, Blitz came out as the most polished and most expensive, GetLeads was his pick for most people, and LeadOcean was the cheapest but about half as good on data quality.

For teams with budget, he likes pairing Clay with GetLeads or Blitz. Clay gives you a lot of signal data on its mid-tier plan, and you plug in your own API key to enrich those leads. Together it runs around $1,000 a month, which he called a huge unlock for agencies sending at scale.
We analyzed 10,000 outbound campaigns to create the ultimate guide on the most common mistakes – it solves up to 80% of all deliverability problems (according to our clients).
Facing deliverability issues?
GUIDES

Cold email copy and offers

How many emails in a sequence?

Three is Jay's default. It gives you enough follow-ups to catch people who missed the first email without overloading your list.

Pick an offer that isn't oversaturated

Asked how he decides whether an offer will work over cold email, Jay said AI can give you a decent first answer, but experience fills in the nuance. He puts himself in the recipient's shoes and asks what would actually make them interested.

He also avoids crowded audiences. Marketing agency owners, coaches and creators get bombarded with cold email, Jay included. He doesn't reply to any of it. In fact, he has an automation that spots cold emails in his inbox and reverse-pitches the sender, since anyone sending cold email is a good lead for cold email tools.

How to get 10 to 20% reply rates

Those dashboard screenshots with 10 to 20% reply rates usually come from stacking two things:
  1. A lead source nobody else is using.
  2. An email that doesn't sell. Jay gets 10 to 20% consistently on partnership outreach, like inviting creators to collaborate or teach in his community. Generous giveaways work too. One of his clients sends $500 worth of free product, and the results are, in his words, ridiculous.

From positive reply to booked call

Solve one bottleneck at a time

One attendee had good positive reply rates, but leads went quiet after he responded. Jay framed it as the next bottleneck. Getting interested replies is usually the hardest step, and once that works, the job becomes turning a raised hand into a booked call.

The levers he suggested:
  • Reach the lead on a second channel, like LinkedIn or Instagram.
  • Leave a voicemail if you have their number.
  • Retarget them with ads after they click something.
  • Look at your follow-up wording and timing, not only the first email.
After that come the next bottlenecks in order: no-shows, then close rate. Fix them one by one.

How to raise your show-up rate

Another attendee was sending 900 emails a day and booking 2 meetings a day, but too many people didn't show. Jay's answer: make the meeting feel personal.

His team sends iMessage reminders that look like they come from the salesperson, often with a short voice note like "Getting on in 5 minutes, see you there." A real person waiting on the other end makes people much less likely to skip. A quick confirmation call right after booking and connecting on social media help too.

Use email for volume, other channels for follow-through

Jay sees LinkedIn, Instagram and cold calling as good supplements to cold email, not replacements. Social channels are hard to scale, since you'd need dozens of accounts. So he doesn't send cold DMs. He waits for a positive reply to an email, then uses it as a signal to reach that person on other channels.

Running a cold email agency

How to reduce client churn

An agency owner with about 30 past clients asked how to keep clients longer. Jay's answer was simple: clients stay when you make them money.

Reporting every week, regular calls and always having a next test planned all help. But if a client's offer is unproven or overpriced, you can book plenty of calls and they'll still leave in 3 to 6 months. So when a new client comes in with a weak offer, Jay coaches them on it first, for example by adding a lead magnet or reshaping the offer so it actually converts.

How long should a good client stay? With established clients and proven offers, he expects at least 12 months. With newer businesses and untested offers, 6 months is more realistic. And look at why a client leaves, not only when.

Pay-per-call, retainer, or build and hand off?

Jay walked through three models:
  • Pay-per-call. The client pays for each booked call. It's easier to sell and clients stay longer, but only do it in a niche you know well, or you'll burn money.
  • Retainer. A fixed monthly fee for a set amount of sending. The most common model.
  • Build and hand off. A lump sum to build the system until it hits a target, then an optional monthly fee to maintain it. This is what Lead Gen Jay does. Most of the work is upfront, and the client keeps everything if they leave.
For pay-per-call, he added a trick: if you specialize in one niche, your mailboxes can stay generic to that niche. When one client leaves, you point the same warmed-up senders at the next one.

Under his own model, getting a system to its target usually takes 60 to 90 days. Maintenance after that costs $1,500 a month.

Closing deals faster

Jay said some people now buy his agency services without a sales call at all, which is rare for anything over $1,000. He credits years of YouTube content and community building for that trust.

For everyone else, his advice is to nurture interested leads. Show them results from past clients, retarget them with case studies, and put them in a drip sequence with testimonials. Nikita added that agencies he talks to see strong results from case studies in retargeting ads.

How Jay uses AI in cold email

AI reply agents

An agency owner with a flood of inbound replies asked which AI setup handles them best. Jay's view: if you're comfortable with AI, build your own agent instead of relying on a sequencer's built-in one. Built-in agents are limited to what they were trained on, while your own can adapt as you test new offers and CTAs.

Jay now runs his whole cold email flow as a scheduled job on a Mac Studio. Every 30 minutes it adds leads to campaigns and handles replies, and he keeps refining it in Claude Code.

Two tips for any AI system you build:
  • Make it self-improving. Add a step that reviews yesterday's replies and suggests what could have been better.
  • Use two models. Jay has Claude and Codex check each other before acting. When they have to agree, mistakes drop.

More automation, more to maintain

Jay used to spend his first hours each day fixing broken Zapier and n8n automations. Now AI maintains most of that for him. The catch is that AI also lets you build far more, so the total work hasn't shrunk. As a founder who never learned to code, though, he says it finally lets him build the tools he always wanted for his clients.

The hardest part of being a founder

We closed with a personal question. Jay said it's the pressure. With around 80 people across two companies, a bad month feels like it's all on you.

He pointed to October 2025, when Google shut down the legacy reseller panels many providers used for mailboxes. Jay said 1,000 to 2,000 of his mailbox clients went down at once, with domains locked and mailboxes disabled. Fixing it took months and couldn't be handed off. In his view, that's what separates real founders from people who just like the idea: they push through setbacks instead of quitting.

FAQ

How many mailboxes should I put on one domain?

For Google mailboxes, Jay starts at 3 to 5 per domain, up to about 5 on aged domains. With large setups of 1,000+ mailboxes, 50 to 100 per domain can work. What matters most is total daily volume per domain and keeping warm-up volume above cold volume.

How many emails should one mailbox send per day?

If you're cutting back to protect your domains, Jay suggests 15 to 20 emails per mailbox per day instead of 30 to 50. To send more, add mailboxes rather than raising volume per mailbox.

Why do I pass inbox placement tests but get no replies from Microsoft?

According to Jay, Microsoft can move an email from the inbox to junk a few seconds after delivery, after the test has already reported success. Aged domains and short, highly varied copy help.

How long should I wait before emailing the same list again?

About a month, with a clearly different sequence.

How many emails should a cold email sequence have?

Jay's default is 3.

Want to ask the next expert yourself?

We run live Q&As like this one with people who send cold email at scale. Join the Maildoso Slack community to hear about the next one, get your copy roasted and ask deliverability questions any time.

To follow Jay, check out his YouTube channel and links.

A note on deliverability

Your copy and targeting only get you so far; the mailboxes you send from decide whether your emails actually land. On cold, misconfigured domains, even the best outreach ends up in spam. Maildoso supplies dedicated, warmed-up domains and mailboxes that get your messages in front of people, ready in about 3 minutes and starting at just $0.49 each.

And with Maildoso's API and MCP server, you can automate domain and mailbox provisioning and plug it straight into your existing stack and AI workflows.
Most cold emails fail simply because they land in spam. People never even see your offer. Our private mailboxes are built specifically for outbound; this means your emails will finally be seen, and you’ll start getting more positive replies.
Boost your outbound with our infrastructure!