Cold email versus LinkedIn outreach

Every outbound vendor sells the same promise: load a list, write a sequence, and the meetings arrive on autopilot. The pitch is not wrong about the mechanics — both channels really can run at volume with modest effort — but it is silent about what volume costs, and the cost is not the subscription. It is a reputation, and the two channels spend a different one.

Cold email spends the reputation of a domain. Every mailbox provider watches how recipients treat mail from a sending domain — opens, replies, deletions without opening, spam complaints — and adjusts where that domain's mail lands accordingly. Push too hard, too fast, with too little relevance, and the domain's mail starts arriving in spam for everyone, not just the people you annoyed. That damage is not really reversible on the domain that caused it. The industry's answer is to never run cold volume through the domain your company actually uses for email. A dedicated sending domain, warmed slowly and separate from the one on your invoices and your team's inboxes, is not an optimisation — it is the minimum condition for running the channel at all, because it means a bad sequence burns a domain you can retire rather than the one your customers already trust.

LinkedIn spends a different reputation: a named person's account, and by extension their professional network. Message enough strangers with enough repetition and the platform throttles or restricts the account, and unlike a sending domain, that account has history attached to it — years of connections, a name people recognise, sometimes a personal brand the person built deliberately. A restricted or banned profile is not something you spin up a replacement for by lunchtime. This is the frame worth holding through the rest of the comparison: cold email risks an asset the company owns and can regenerate, LinkedIn risks an asset a person owns and often cannot.

The ceiling on LinkedIn is structural, not a setting

Cold email's constraint is mostly about pacing and infrastructure — how many domains you are willing to warm, how much of the list is verified, how much personalisation each message gets. Given enough sending domains and enough discipline, the volume can grow roughly in proportion to the money and time put into it.

LinkedIn does not offer that trade. The platform caps how many connection requests and messages an account can send, and those caps apply per account regardless of budget — you cannot buy your way past them, you can only add more accounts, which means more people whose names and networks are now on the line. This is the same structural point the pillar piece on which channels actually scale makes about outbound in general: it has a hard, countable ceiling, and on LinkedIn specifically that ceiling is set per seat, not per company. A five-person sales team can run five people's worth of LinkedIn outreach and not a company's worth. Cold email has no equivalent per-seat wall — the limiting factor is deliverability infrastructure, which is a problem you can solve with process, not one you hit a wall on by definition.

The practical consequence is who each channel suits. A single founder or a two-person team with real personal credibility in a niche can get disproportionate results from LinkedIn, because the ceiling per account is exactly the resource they have — their own time and their own network. A company trying to run outbound at real volume, with several people or a dedicated function behind it, runs into LinkedIn's per-seat ceiling almost immediately and has to lean on email, where the constraint is money and process rather than headcount.

The arithmetic that gets left out of both pitches

Every outbound vendor quotes a reply rate as if a reply were the unit of value. It is not. A reply that says "not interested, please remove me" and a reply that books a call are both replies, and the sequences that generate the most total replies are frequently the ones generating the most of the first kind.

The honest version of the metric has three lines, not one: positive replies that indicate real interest, negative replies that at minimum cost you the researching and sending time and at worst damage the relationship with a company you might want later, and unsubscribes or spam complaints, which on email actively feed the deliverability problem described above — a rising complaint rate is one of the signals that gets a domain throttled. A sequence that produces ten meetings and forty complaints is not obviously better than one that produces six meetings and four complaints, because the first one is also quietly spending down the asset that lets you send the next campaign at all. On LinkedIn the equivalent cost is a connection request ignored or reported, which does not carry the same deliverability mechanic but does draw down the same finite budget of requests the account gets before the platform notices the pattern.

This is also where list quality does more work than either channel's marketing admits. A narrow, well-researched list run at a slower pace consistently beats a large purchased or scraped list run fast, on both channels, because the negative-reply and complaint lines scale with irrelevance far faster than the positive-reply line scales with volume. The company that believes it needs a bigger list usually needs a better-targeted smaller one instead.

The legal ground shifts under email, and does not move much under LinkedIn

Messaging someone on LinkedIn sits inside a platform whose terms already govern the relationship — you agreed to them, so did they, and the legal exposure is mostly about violating those terms rather than a separate body of external law. Cold email is different, because it is regulated as email, by rules that vary by jurisdiction and do not converge the way marketers sometimes assume.

Some European jurisdictions allow unsolicited B2B email to a professional address under a legitimate-interest basis when the offer is genuinely relevant to that person's role. Others require something closer to prior consent even in a business context, and a sequence written and cleared for the first kind of jurisdiction can be a compliance problem the moment the list includes contacts in the second. This is not a detail to leave to the sending tool's default settings — a template built once and pointed at a global list is being pointed, unknowingly, at several different legal regimes at the same time. Anyone running cold email outside a single home jurisdiction needs to know which rule applies to which part of the list before the first send, not after a complaint.

The recommendation, and where neither channel is defensible

Start with cold email, run from a domain built for the purpose and never the one your company actually depends on. It is the channel whose main constraint — deliverability infrastructure — is a problem you can solve with process and money rather than one that caps out with the size of your team, and it is the channel where the worst-case damage lands on an asset you can replace. Bring LinkedIn in as the second channel, and put it in the hands of whoever's personal credibility is genuinely relevant to the buyer — a senior person messaging as themselves, not a rotation of SDR accounts running a script, because the entire value of the channel is a real name attached to a real network, and a scripted account throws that away for nothing extra.

Neither channel is defensible for a company whose buyer cannot be found on a list or a platform at all — a genuinely local service business, a product sold through distributors who never see the end customer, anything where the addressable population any outbound sequence could reach is a rounding error next to the market. For those companies the honest answer is that outbound is not the channel, and the money is better spent where the buyer already looks, which is closer to the territory covered in what makes paid social work for B2B companies or in strengthening whatever is already working informally, the subject of the referral channel nobody measures. Running either cold email or LinkedIn outreach at a market that structurally cannot be reached by list or platform does not fail quietly — it produces the same low-signal noise the wider channel discipline warns against, just with a domain or an account paying for it.