How to Improve Cold Email Deliverability (Data-Grade List Hygiene)

To improve cold email deliverability, fix the list before you touch the sequencer. In a 37,040-contact US outbound sample processed by LeadSonar, 99.98% of addresses carried a valid MX record and only 0.55% produced a hard bounce when the campaign was actually mailed, with the bounces dominated by Google 550-5.1.1 NoSuchUser and 550-5.2.1 DisabledUser responses. Those are addresses a verification pass removes before a single message leaves your outbox. Based on Q3 2026 LeadSonar production data.

Cold email deliverability is decided across two layers that fail independently. The sending layer covers authentication, domain warmup, and volume pacing, and it lives in your own outbound tool. The data layer covers whether each address is real, reachable, and deliverable, and it is the layer where most preventable bounces are born. This guide is about the data layer: how to build a verified, MX-valid, low-bounce list so your sequencer spends its limited daily volume on addresses that actually accept mail. Sending itself stays in your sequencer, where authentication and warmup belong.

Deliverability Is Decided Before You Hit Send

A cold email bounces for one of two reasons: the receiving server never accepted the address, or your sending reputation was already too low for the inbox to trust the message. The first reason is a data problem, and it is almost fully solvable before send. The second is an infrastructure problem, and it compounds every time the first goes unaddressed.

The mechanism is direct. Every hard bounce, spam-trap hit, and "user unknown" rejection is a negative signal that Gmail, Microsoft, and other receivers record against your sending domain. A list with 10% invalid addresses mailed from a perfectly authenticated inbox will still post a 10% bounce rate on the first send, and a bounce rate above the 2% threshold that Gmail and Microsoft treat as a reputation risk starts pushing your later, valid messages into spam. The dirty list poisons the deliverability of the clean part of the same list.

LeadSonar operates entirely at this data layer. The platform finds, verifies, and enriches B2B contacts through a waterfall cascade of 18-20 data providers and gates every email through EmailShield verification before it is stored, sustaining 23.5 enriched profiles per second and resolving roughly 85% of contacts across the cascade while charging only for verified results. The actual sending, warmup, and cadence run in your own sequencer. The division of labor is deliberate: LeadSonar hands your outbound tool a list that has already been stripped of the addresses that generate bounces.

The Two Layers of Cold Email Deliverability

"How to improve cold email deliverability" collapses two different engineering problems into one search query. Separating them is the first practical step, because the fixes live in different systems and different teams often own them.

| Layer | What it controls | Where it lives | Primary failure mode | |---|---|---|---| | Sending / infrastructure | Whether a valid message reaches the inbox | Your sequencer and DNS | Missing SPF/DKIM/DMARC, cold domain, over-sending | | Data / list | Whether the address is real and deliverable | Your data platform | Invalid addresses, catch-all, role-based, stale records |

The sending layer is well documented across the deliverability-vendor ecosystem, and the rules are stable. Authenticate the sending domain with SPF, DKIM, and DMARC. Warm the inbox gradually. Keep daily volume per inbox modest and spread it across multiple inboxes and domains. Monitor your reputation with tooling from the receivers themselves. Those are the correct sending-layer moves, and they are handled inside your outbound tool.

The data layer gets far less attention, and it is where cold email programs quietly bleed deliverability. Every address that should have been removed before send burns a fraction of your daily volume and adds a negative reputation signal. The rest of this guide is the data-layer workflow, followed by the sending-layer references you will hand to whoever owns your DNS and warmup.

Where Bounces Are Actually Born

Bounces come from identifiable, removable classes of addresses. Naming them is what turns "improve deliverability" from a vague goal into a checklist.

Invalid addresses. The mailbox does not exist or has been disabled. These return SMTP 550-5.1.1 or 550-5.2.1 and are the single largest source of hard bounces. In the 37,040-contact sample, the 0.55% that hard-bounced were dominated by exactly these two response codes.

Catch-all domains. The domain accepts every address at the SMTP layer, so no verifier can confirm a specific mailbox by probe alone. Some catch-all addresses are real and some are dead, and sending to a scraped catch-all list from an untrusted source is one of the fastest ways to run up a bounce rate.

Role-based addresses. info@, sales@, support@, and similar shared inboxes are frequently filtered, forwarded, or monitored by spam systems, and they convert poorly. A corporate outbound list should carry named individuals on custom company domains.

Disposable and free-consumer domains. Disposable domains exist to be thrown away, and free consumer domains such as gmail.com and yahoo.com signal a non-corporate contact on a B2B list. In a 103,000-contact B2B database processed by LeadSonar, 100% of contacts carried a deliverable email on a custom company domain and 0% sat on a free consumer provider, which is the shape a clean corporate list should have.

Stale records. B2B contacts change jobs constantly, and an address verified six months ago may now bounce. Freshness is a deliverability variable, which is why verification close to send time matters as much as the initial clean.

Each class has a removal or segmentation rule, and applying all of them in sequence is the data-hygiene workflow.

The Data-Hygiene Workflow: Six Steps to a Low-Bounce List

This is the workflow LeadSonar runs to produce the numbers cited above. It applies whether you are building a list from scratch through Lead Finder or cleaning an existing CSV through bulk enrichment. Each step maps to a removable bounce class from the section above. For a deeper treatment of the cleaning pass on an existing file, see the companion guide on how to clean a B2B email list.

Step 1: Validate mail-server reachability

Resolve the MX record for every domain and drop any address whose domain has no reachable mail server. This is cheap, deterministic, and catches typo'd and defunct domains before any expensive SMTP work. A well-sourced list clears this stage almost entirely: in the 37,040-contact outbound sample, 99.98% of addresses had a valid MX record and only 6 of 37,040 rows lacked one.

Step 2: SMTP-verify every address

Run each surviving address through the full verification pipeline: RFC 5322 syntax check, MX confirmation, and an SMTP handshake that classifies the mailbox as Valid, Invalid, Risky, Catch-all, Disposable, or Role-based. LeadSonar routes this through EmailShield, its sister verification engine, and an address is treated as found only after it clears verification. When a provider in the waterfall returns an address that fails the SMTP check, the cascade discards it and moves to the next source, so an invalid hit never reaches your export. The verification engine runs at 99.9% accuracy on the classification decision, and the measured downstream result is 94.6% average deliverability on verified emails over a trailing 30 days.

Step 3: Strip role-based, disposable, and free-consumer addresses

Filter out info@, sales@, and support@ role addresses, disposable domains, and free consumer domains on a corporate outbound list. These carry the Role-based and Disposable flags from Step 2, so the removal is a filter on data you already have. The target is a list of named individuals on custom company domains, the same profile as the 103,000-contact sample where every contact sat on a corporate domain.

Step 4: Segment catch-all by source confidence

Catch-all domains return SMTP 250 to every probe, so a per-address verdict is impossible at the protocol level. The right move is segmentation by source confidence. A catch-all address from a form submission, an existing customer, or a named partner is generally safe to mail. A catch-all address from a scraped or purchased list is high-risk and belongs in an excluded or warm-up bucket. LeadSonar flags catch-all explicitly and reduces the confidence score on catch-all domains, so you can sort your export and decide policy per segment. For the accuracy mechanics behind these flags, see B2B contact data accuracy.

Step 5: Re-verify close to send time

A list cleaned in April and mailed in July has decayed. People changed roles, mailboxes were disabled, and domains lapsed. Verify addresses close to when you actually send them. LeadSonar verifies at lookup in real time, with sub-1s response on cached or single-source checks and roughly 2s for a full waterfall pass, so the deliverable share reflects the state of the mailbox today.

Step 6: Export the clean list to your sequencer

Export the verified, MX-valid, catch-all-segmented list to CSV or XLSX, or push it directly to your outbound tool. Sending, warmup, and cadence run in your own sequencer, which now receives a list engineered for a hard-bounce rate under 1%. LeadSonar's role ends at the clean, deliverable list. What you do with authentication and pacing is the sending-layer work covered next. For the verification step in isolation, the outbound list build workflow covers the API and bulk paths.

The Sending Layer: Authentication, Warmup, and Volume

The data-hygiene workflow gives your sending infrastructure a list that will not sabotage it. The sending layer then determines whether those verified messages reach the inbox. This work happens in your own sequencer and DNS, and the standards are public.

Authenticate the domain. Configure all three records on the sending domain. SPF (defined in RFC 7208) publishes the servers allowed to send for your domain. DKIM (defined in RFC 6376) attaches a cryptographic signature to every message. DMARC (defined in RFC 7489) tells receivers how to handle mail that fails the first two. Gmail and Yahoo now require authentication for bulk senders, and the receiver-side expectations are documented in Google's sender guidelines.

Warm the domain and inboxes. A brand-new sending domain has no reputation, so ramp volume gradually over two to four weeks before running full campaigns. Warmup is a function of your sequencer or a dedicated warmup tool.

Pace volume. Keep per-inbox daily volume modest, commonly 20 to 50 cold sends, and spread campaigns across multiple inboxes and domains. The SMTP transport itself is defined in RFC 5321, and the messaging-industry best practices for bulk senders are maintained by M3AAWG. Monitor reputation with Google Postmaster Tools and the equivalent Microsoft dashboards.

The reason the sending layer is a short section here is that it is thoroughly covered by the deliverability-tooling ecosystem and it is orthogonal to data quality. A flawless authentication setup mailing a 10%-invalid list still bounces 10%. The data layer is the multiplier that decides how much of your sending-layer effort survives contact with the list.

Two Production Examples

A revenue-operations SaaS SDR team, 240,000-contact list

A six-person SDR team selling revenue-operations software built a 240,000-contact target list through Lead Finder and waterfall enrichment, filtering on job title, headcount buckets, and industry. Across the enrichment run, roughly 50% of contacts resolved on the first, cheapest provider tier, another 35% were recovered by deeper tiers, and about 15% were genuine misses the team was never charged for under stop-on-hit billing.

Before adopting a verify-at-source workflow, the team had mailed an earlier list assembled from mixed exports and posted an 11% bounce rate that dragged their primary sending domain into spam folders for two weeks. On the verified 240,000-contact list, MX validity sat at essentially 100%, role-based and free-consumer addresses were stripped in export, and the mailed hard-bounce rate came in at 0.6%. The team consolidated three separate tools, a finder, a verifier, and an enrichment vendor, into one LeadSonar workspace billed on the Pro plan at $199 per month with 800,000 lead credits, and ran the overflow through credit rollover across two months.

An outbound agency running 12 client ICPs

An outbound agency running cold campaigns for 12 client ICPs enriches roughly 50,000 leads per month across SaaS, fintech, and professional-services accounts. Each client list previously went out with whatever addresses the source export contained, and bounce rates hovered around 11% on the messier verticals, forcing constant domain rotation and reputation firefighting.

The agency moved every list through EmailShield verification inside LeadSonar, excluded catch-all rows from unverified sources, and segmented high-source-confidence catch-all contacts into a slower cadence. Cold-email bounce dropped from 11% to 1.4% within two weeks, and the reduction in burned sending domains cut the agency's warmup overhead materially. Billing ran on the dual credit model: one lead credit per contact revealed and one enrichment credit per AI action, with CSV enrichment charged per filled cell and unfilled cells refunded, so the agency paid for the verified data it kept rather than the rows it discarded.

What "Good" Looks Like: The Deliverability Numbers

Concrete targets make the workflow measurable. The table below pairs each metric with the LeadSonar first-party observation behind it.

| Metric | Healthy target | LeadSonar observation | Sample | |---|---|---|---| | Valid MX rate | Above 99% | 99.98% valid MX | 37,040-contact outbound sample | | Hard-bounce rate | Under 2%, ideally under 1% | 0.55% hard-bounce when mailed | 37,040-contact outbound sample | | Measured deliverability | 90%+ | 94.6% average deliverability, trailing 30 days | Verified-email population | | Corporate-domain share | High on B2B lists | 100% on custom company domains, 0% free-consumer | 103,000-contact B2B sample | | Verification-engine accuracy | Above 99% | 99.9% classification accuracy | EmailShield engine |

Two cautions on reading these numbers. The 99.9% figure is the accuracy of the verification engine's classification decision on an address, and it is a separate thing from inbox placement, which the 94.6% trailing-30-day deliverability figure measures. And the 0.55% hard-bounce number is what a properly verified list produces when mailed, so it describes the output of the workflow rather than a raw scraped list, which routinely bounces 8% or more.

The economics of running this workflow are pay-per-result. LeadSonar starts at $29 per month on the Starter plan with 50,000 lead credits, and the category it competes in starts higher, from $49 to $199 per month at comparable tools. Because the waterfall stops on the first verified hit and misses are never billed, you pay an industry-leading blended cost per verified email rather than a flat fee per lookup attempt. Cancel one seat of a legacy sales-intelligence tool and the LeadSonar subscription is covered for a year.

FAQ

What is a good deliverability rate for cold email?

A healthy cold email program keeps hard-bounce under 2% and lands the large majority of sent mail in the inbox. LeadSonar measures 94.6% average deliverability on verified emails over a trailing 30-day window, and in a 37,040-contact outbound sample only 0.55% of addresses hard-bounced when actually mailed. Bounce rates above 2% start to damage sender reputation at Gmail and Microsoft, so list quality is the first lever to pull before touching sending infrastructure.

How do SPF, DKIM, and DMARC affect cold email deliverability?

SPF, DKIM, and DMARC are authentication records that let receiving servers confirm your mail is genuinely from your domain. SPF (RFC 7208) lists the servers allowed to send for your domain, DKIM (RFC 6376) cryptographically signs each message, and DMARC (RFC 7489) tells receivers what to do when the first two fail. Configuring all three on the sending domain is required for consistent inbox placement, and it happens in your own sending infrastructure rather than in a data platform.

Why are my cold emails going to spam?

Cold emails land in spam for two independent reasons: a dirty list that generates bounces and spam-trap hits, and a sending setup that lacks authentication or warmup. The data-layer causes are invalid addresses, catch-all domains, role-based inboxes, and stale records that bounce. In a 37,040-contact sample, hard bounces were dominated by Google 550-5.1.1 NoSuchUser and 550-5.2.1 DisabledUser responses, both of which come from addresses that verification removes before send.

How many cold emails should I send per day per inbox?

Most senders ramp a warmed inbox to roughly 20 to 50 cold emails per day and spread volume across multiple inboxes and domains rather than blasting from one address. Volume pacing and warmup are sending-infrastructure decisions handled in your sequencer. The data platform's job is to make sure every one of those sends goes to a verified, deliverable address so the limited daily volume is not wasted on bounces.

Does list quality or sending setup matter more for deliverability?

Both matter, and they fail in different places. Authentication and warmup determine whether a verified message reaches the inbox, and list quality determines how many of your sends bounce or hit spam traps in the first place. A perfectly authenticated inbox mailing a list with 10% invalid addresses will still post a bounce rate that triggers reputation penalties, which is why data-grade list hygiene is the foundation the sending layer sits on.

How do I clean a cold email list to improve deliverability?

Validate the MX record, run each address through an SMTP handshake to classify it as valid, invalid, risky, catch-all, disposable, or role-based, strip the undeliverable classes, segment catch-all by source confidence, and re-verify close to send time. LeadSonar gates every email through EmailShield verification and charges only for verified results, so misses are absorbed rather than billed. The output is a list built for a hard-bounce rate under 1%.

Methodology

First-party statistics in this article come from LeadSonar production data and real anonymized datasets processed by the platform. The 99.98% valid-MX rate, 0.55% hard-bounce rate, and the 550-5.1.1 / 550-5.2.1 bounce-code breakdown are computed from a 37,040-contact US outbound sample that was streamed and actually mailed. The 100% corporate-domain and 0% free-consumer-domain figures are computed from a separate 103,000-contact B2B sample. The 94.6% average deliverability figure is measured across the verified-email population over a trailing 30-day window, and the 99.9% figure is the classification accuracy of the EmailShield verification engine, which is a distinct measurement from inbox placement. Throughput (23.5 profiles per second), latency (sub-1s cached, roughly 2s full waterfall), and the roughly 85% cascade hit rate are measured on production runs.

Client examples are anonymized real engagement shapes; exact figures are representative of the platform's aggregate performance and are consistent with the samples above. Pricing reflects the plans live as of Q3 2026 (Starter $29, Growth $79, Pro $199 per month) and the dual credit-pool model, where revealing a contact costs one lead credit, each AI action costs one enrichment credit, and CSV enrichment is billed per filled cell with unfilled cells refunded.

Sending-layer standards are drawn from the published RFCs for SMTP, SPF, DKIM, and DMARC, from Google's sender and Postmaster documentation, and from M3AAWG best-practice publications, all linked inline. Limitations: bounce and deliverability figures describe verified lists produced by this workflow and do not represent raw, unverified, or scraped lists, which bounce at materially higher rates. Sending, warmup, and cadence are performed in the user's own outbound tool and are outside LeadSonar's measured scope.

Last updated: July 2026.

Written by Nikita Stoletov, CTO at LeadSonar, where he leads lead-intelligence and waterfall-enrichment engineering. Nikita architected the multi-provider cascade and the EmailShield-gated verification pipeline that keeps LeadSonar's verified emails at a measured 94.6% deliverability. Read more at leadsonar.io/authors/nikita-stoletov or connect on LinkedIn.