Two to three contacts per company is the right number for cold email: chosen by role, staggered across a week or more, each carrying a message written for that specific person. That is the whole answer to how many contacts per company you should work. One contact leaves the deal hanging on a single inbox. Emailing the full org chart turns a campaign into a domain blast that the receiving company’s mail server learns to treat as one spammer.
That range is not theory for us. Our parent agency, Referral Program Pros, has run more than 4,000 outbound campaigns and booked over 7,000 meetings, and GTM Bud productizes that playbook under a written guarantee: 5 percent positive replies on LinkedIn, 1.5 percent on email, or a full refund. Across those campaigns, the accounts worth multithreading always got a small set of differentiated contacts entering the sequence days apart, never a synchronized blast, because the campaigns that hit every address at a domain at once are the ones that stopped delivering.
How many contacts per company should you cold email?
The working consensus across sales engagement vendors and community threads is 2 to 3 well-chosen contacts per company, never entered on the same day, with roughly 3 touches landing at any single company in a given week as the practical ceiling. The data that exists points the same direction. Hunter’s State of Cold Email report, built on 31 million emails sent through its platform, found that campaigns contacting 1 to 2 people per company earned a 5.1 percent reply rate against 3.5 percent for campaigns emailing 3 or more, a 46 percent difference in favor of restraint. That is vendor data from one platform’s user base, not a controlled study, but it matches what practitioner threads on the question keep reporting: the second and third contact add real coverage, and every contact past the third adds mostly risk.
Multithreading is the practice of engaging two or three relevant stakeholders at the same account with coordinated, role-specific messages. It is not the same thing as contacting everyone whose address your enrichment tool returned, and the distinction is the entire game. The case for the second and third thread is insurance against the oldest outbound failure there is: Modern Inbound’s sequencing guide argues that the bottleneck in larger deals is almost never touch count but single-threading into one person who cannot make the call alone. Your champion goes on leave, changes jobs, or simply never forwards the email, and the account dies with the thread. Two or three threads mean the account survives any one inbox going dark.
Why does emailing everyone at the same domain backfire?
Because the receiving company sees your campaign as one sender, not as ten separate emails. Corporate mail servers and Google Workspace tenants aggregate signals per sending domain, so every additional identical message to the same company compounds the previous ones. Emelia, a cold email platform, reports in its guide to campaign mistakes that when senders target hundreds of contacts at a single company, the campaign lands in spam within about ten messages in 99 percent of the cases it has observed, often with no bounce and no warning, just opens flatlining. And the human failure mode is worse than the technical one. A cautionary tale that circulates among outbound practitioners on LinkedIn describes a sender who emailed the same pitch to 12 people at one company; several forwarded it to the same decision maker, who answered with a complaint and a blacklist entry. Identical messages do not multiply your chances. They multiply your exposure.
The complaint math makes the ceiling concrete. Google’s email sender guidelines require bulk senders to keep spam complaint rates under 0.3 percent, and one annoyed office can produce several complaints from a single tenant in one afternoon, which is a complaint spike no copy improvement can undo. Domain-level restraint also has to live inside your overall sending discipline: the per-inbox and per-domain ceilings are their own topic, covered in our guide to how many cold emails per day are safe. The contacts-per-company cap is the account-level version of the same principle. Volume that nobody asked for is the thing that gets punished, and nowhere is volume more visible than ten identical emails arriving at one company in one morning.
Who should the 2 to 3 contacts be?
Pick roles, not names: the person who lives with the problem daily, the person who owns the budget for solving it, and the person who sponsors the outcome at the executive level. Those three read the same pitch three different ways, which is exactly what you want, because any of them can start the internal conversation and each forwards upward or downward along a different path. Finding which humans hold those roles at a given account is its own research step, and our guide to finding decision makers in a company covers the mechanics. Whether the account deserves three threads at all is an upstream question your ICP definition should already answer: a 20-person company does not have three distinct stakeholders for most offers, and pretending it does just burns a contact slot.
Company size sets the cap. These are the bands we run at Referral Program Pros:
| Company size | Contacts per company | How they enter |
|---|---|---|
| 1 to 50 employees | 1 to 2 | 3 to 4 days apart |
| 51 to 500 employees | 2 to 3 | 3 to 4 days apart, junior before senior |
| 500+ employees | 3, within one division | Staggered across 2+ weeks, one department at a time |
The enterprise row carries the rule that saves the account for later: stay inside one division. Three contacts in one business unit is multithreading. Three contacts scattered across three unrelated departments is a blast wearing a strategy costume, and it spends your credibility with every team at once.
The staggered multithreading schedule for small senders
Stagger the entries so the account hears from you as a sequence of individuals, not as a campaign. The pattern community threads describe, and the one we run at Referral Program Pros, starts low and climbs: the hands-on owner of the problem first, the budget holder a few days later, the executive sponsor last. By the time the senior message lands, there is a nonzero chance your name has already been mentioned internally, which converts the cold exec touch into something closer to a warm one.
| Day | Who enters | What the message does |
|---|---|---|
| 1 | The IC or manager closest to the pain | Specific, tactical, easy to say yes to |
| 4 | The director who owns the budget | Business case framing; may reference the team-level conversation |
| 8 | The VP or executive sponsor | Shortest note of the three, strategic, one clear question |
| 11+ | Follow-ups interleave | Each thread follows up on its own clock, total kept near 3 per week |
Two rules keep the schedule honest. First, the weekly cap governs the whole account: with three threads running, each contact gets a short sequence rather than a long one, because three contacts on aggressive five-step cadences is fifteen emails at one domain in two weeks. Keep individual cadences lean and let the craft of spacing and re-angling live where it belongs, in your follow-up sequence design. Second, never let two entries share a day. The same upload that looks staggered in your plan becomes a synchronized blast the moment a sequencer schedules every row at once, which is a tooling problem before it is a strategy problem, and it is covered below.
How do you differentiate the three messages?
Write each message so it survives a forward. The test is simple: if the director forwards your note to the IC you emailed three days earlier, the two messages should read as one person doing real research from two angles, not as a mail merge with the first name swapped. That means a different opening observation, a different slice of the value proposition, and a different ask for each role. The IC gets the workflow pain and a low-commitment question. The director gets the cost or revenue framing and a short business case. The VP gets one sentence of context and one strategic question. Role-based message variation is segmentation applied inside a single account, and the same discipline you use to split a list by persona applies here; our guide to cold email list segmentation covers how to build those role buckets so the variation is systematic instead of improvised.
Differentiation is also the deliverability play, not just the courtesy play. Receiving filters flag near-identical bodies arriving at one tenant from one sender, so three genuinely different messages look like correspondence while three templated clones look like a campaign. The cautionary tale above was not punished for emailing 12 people. It was punished for emailing 12 people the same thing.
How should same-domain dedupe work in your sequencer?
Same-domain dedupe is the set of rules a sequencer applies so one company is never treated as a list of unrelated rows. At minimum it should do four things. Group contacts by company domain before anything is scheduled, so every cap applies to the account instead of the row. Enforce a contact cap per domain, with 2 to 3 as the sane default and anything above it requiring a deliberate override. Space same-domain sends automatically, so an upload of three contacts produces staggered entries instead of a synchronized morning blast. And stop on reply at the account level: the moment one contact answers, every other sequence at that domain pauses until a human decides whether the remaining threads continue, convert into referenced names, or end. A sequencer missing any of these four will eventually blast a domain for you, politely, on schedule, while you sleep.
This is the part of the playbook we refused to leave manual. GTM Bud’s cold email automation tool applies domain grouping, per-company caps, staggered entry, and account-level stop-on-reply by default, and writes a different angle for each role rather than cloning one template across the account. The strategy in this article only works if the software enforces it on the days nobody is watching.
Frequently asked questions about contacts per company in cold email
Should you mention that you emailed someone else at the company?
Yes, as long as it is true, and especially after a reply or a forward. A line like “I also reached out to your colleague on the data side” reads as organized rather than sneaky, because multithreaded prospects compare notes whether you acknowledge it or not. Hiding the other threads is the only losing move, since a single forward exposes it instantly and turns coordination into something that looks like a blast.
What should happen when one contact replies and the others are mid-sequence?
Every other sequence at that domain should pause the moment any reply lands, whatever its sentiment. A positive reply turns the remaining contacts into people you reference in the live thread, not people you keep cold emailing. A hard no from a senior contact usually closes the account for the cycle, while a redirect to a colleague means you restart one thread carrying the referral.
Do follow-ups count toward the per-company cap?
Yes. The practical ceiling of about three touches per company per week counts every message that lands at the domain, follow-ups included. Three contacts each receiving a three-step sequence is nine emails at one company, which is exactly why the entries are staggered days apart and the individual sequences are kept short.
How long should you wait before re-approaching a company that never replied?
Our agency works on 60 to 90 days of quiet before a company re-enters a campaign, and the second pass needs a different angle and at least one different contact. Re-sending the same pitch to the same three people after a month reads as the same campaign and inherits its silence. The fresh entry point usually comes from re-running your automated lead generation pass on the account rather than recycling the same three names.
Does the contacts-per-company rule apply to LinkedIn outreach too?
It applies more tightly. LinkedIn is a visible channel where profile views, connection requests, and mutual connections surface to the whole team, so three simultaneous connection requests at one small company look coordinated in a way three emails never do. We generally run one LinkedIn thread per account, at most two, and let email carry the rest of the coverage.
Cover the account without burning the domain
The contacts per company question has a stable answer because every input agrees on it: 2 to 3 people chosen by role, entering days apart, each carrying a message the other two could forward without embarrassing you. The reply-rate data rewards restraint, the blacklist stories punish blasting, and the multithreading upside only exists when the threads are genuinely different. Pick the three roles, stagger the week, keep the account under three touches, and let your sequencer enforce the caps you decided on while calm. If you would rather the differentiation and the dedupe both happen automatically, GTM Bud’s AI cold email writer drafts a distinct angle for every contact at the account and the sequencer holds the caps, the same discipline behind the 7,000+ meetings our agency booked to earn these rules.