LinkedIn DM sequences: what to say after they accept

What is a LinkedIn DM sequence?

A LinkedIn DM sequence is a series of messages sent to a prospect after they accept a connection request, usually two to four messages with different angles, stopping the moment they reply. It is a distinct object from the connection request, it runs on a different set of rules, and it is where copy earns most of its value in this channel.

The distinction matters because the two get planned as one thing. The connection request is a permission gate where less is usually more. The DM sequence is a persuasion surface where specificity wins. Advice that is correct for one is often wrong for the other, which is a large part of why LinkedIn copy advice reads as contradictory.

What is the structural mistake that quietly kills sequences?

Never chain a DM directly after a connection request. The sequence must fork on an accepted condition, with messages on the accepted branch only, and the not accepted branch ending.

This sounds obvious and it is the single most common build error we correct on inherited campaigns. When the DM node sits straight after the request node, the sequence attempts to message people who never connected. Those messages either fail, or on a paid seat they consume InMail credits you did not intend to spend, and the reporting fills with sends that had no chance of a reply. Reply rate looks catastrophic, and the team starts rewriting messages that were never delivered to anybody.

The build rule we run on every client sequence: connection request node, then an accepted within days condition, then the DM branch on yes, with the no branch left to end. Get the tree right before enrolling anyone, because most sending platforms freeze structural edits once a sequence has live enrollments. Copy can usually be patched on a running sequence. Structure cannot.

What should the first DM say after someone accepts?

The first message should ask a short question that the recipient can answer in one line, and it should make it obvious why they were contacted.

The reason is mechanical. A LinkedIn DM appears in a messaging interface that people read on a phone, in the same list as messages from actual colleagues. A long message reads as a document and gets postponed. A question reads as a conversation and gets answered, and once there is a reply the interaction has changed shape entirely.

Three things belong in that first message. A reason you contacted this person specifically, which can be a segment level reason and does not need to be individual. A single sentence of credibility that establishes why they should spend a reply on you. One question with an easy answer. Nothing else, and in particular no calendar link, because a link in the first message asks for the largest possible commitment at the moment of least trust.

One distinction from our own build standard that is worth stating, because it is the kind of detail that only shows up when you run this at volume: the opener changes depending on how the connection happened. A person who accepted your request has given you a small signal of interest and the opener can acknowledge the connection. A person you are reaching by InMail or by another route has given you nothing, and using a warm opener on them reads as a false claim of familiarity. We maintain separate cold and warm openers for that reason and route them by branch.

Should you pitch in the first message or build rapport first?

Neither, in the form both are usually meant, and the framing is the problem.

Immediate pitching in the first DM works in high volume campaigns against mid market accounts where speed matters more than relationship depth, and it fails badly with senior buyers and considered purchases. Extended rapport building before mentioning why you are there fails almost everywhere in cold outreach, because the recipient knows there is a reason and watching you avoid it costs you the credibility you were trying to build.

What works across the campaigns we run is the middle option that is neither: be transparent about why you are there in the first message, but ask a question instead of making an offer. The recipient knows the shape of the conversation, which respects their time, and they are being asked for one line, not a meeting.

The one variable that genuinely shifts the answer is deal size. The larger the deal, the more the first message should qualify and the less it should sell. On smaller and faster deals, directness wins and the extra touch costs you more than it earns.

How long should the sequence be, and how far apart?

Two to four messages is the range we run, and the spacing question is genuinely not this page's subject.

Sequence length, touch spacing, when the first DM should fire relative to acceptance, and how LinkedIn touches interleave with email touches are all orchestration questions, not copy questions. They belong to our sibling property MultichannelPros. This page owns what the message says. That property owns when it fires.

The one thing that belongs here: every follow up needs its own angle. A follow up that restates the first message adds volume without adding information, and the recipient who ignored the first one has already made the decision it was asking for. New angle, or stop.

What does a 9% DM reply rate actually mean?

It means that on one white label programme running across a set of advisor workspaces, roughly nine in a hundred LinkedIn DMs sent received an answer of any kind, in the same window in which email from the same senders was replying at roughly 1.5%.

What it does not mean. It is not a positive reply rate. Not interested is a reply and it is counted here. It is not a booking rate, and the gap between a reply and a booked meeting is where most of the real work sits. It is not a promise about your programme, because it came from one audience in one vertical with a sender that audience recognised. And it does not mean LinkedIn produced more total replies than email on that programme, because the two channels did not send comparable volumes. We do not publish the absolute sends and replies per channel, because they are not yet stated with the precision the rates are.