Most LinkedIn outreach fails before the first message is sent because there was no strategy, just a tool and a hope. A strategy is the set of decisions that make every message land on the right person at the right time with the right ask.
Start with the ICP, not the message
The single biggest lever in outreach is who you contact. Define your ideal customer profile precisely: industry, company size, role, and a trigger that makes now the right time. A sharp list with an average message beats a brilliant message sent to the wrong people every time.
Make your profile do half the selling
Before anyone replies, they look at your profile. Your headline should say who you help and how, not just your job title. Your banner, about section, and recent activity should reinforce that you are a credible peer. A weak profile caps your response rate no matter how good your message is.
Design the sequence
- Connection request: short note or no note, depending on your audience.
- First message after they accept: value-first, no pitch, a single small ask.
- Follow-up one: a relevant resource or a specific reason to talk.
- Follow-up two: a soft, final nudge, then stop.
Space the steps over days, not minutes, and never send all of them if the prospect replies. The goal is a conversation, not a completed sequence.
Set your cadence and limits
Decide how many new prospects you contact per day and stick to a safe, ramped volume. Consistency beats bursts, both for results and for keeping your account healthy.
Warm the account before you scale
A strategy that starts at full volume on a dormant or brand-new profile is a strategy for a restriction. LinkedIn weighs how an account has behaved, so the order matters: complete the profile, accept inbound requests, react and comment like a person, and only then add outbound requests.
Warmerly builds this into its ceilings. The ramp day is the larger of the days LinkedIn warmup has run and the days since you connected the account, so an account used for campaigns alone still climbs with time. The ceilings below are what Warmerly lets an account do on each ramp day. They are Warmerly's settings, not LinkedIn's.
| Action | Days 1 to 3 | Days 4 to 7 | Days 8 to 14 | Days 15 to 21 | Day 22 on |
|---|---|---|---|---|---|
| Connection requests, LinkedIn Basic account | 5 | 6 | 8 | 10 | 10 |
| Connection requests, LinkedIn Premium account | 10 | 12 | 16 | 20 | 20 |
| Messages to connections | 10 | 17 | 27 | 37 | 50 |
| Profile views | 8 | 14 | 22 | 30 | 40 |
| Reactions | 4 | 7 | 11 | 15 | 20 |
| Comments | 1 | 1 | 2 | 3 | 5 |
| Posts | 0 | 0 | 1 | 1 | 2 |
Invitations also have a weekly ceiling: 25, 30, 40 and 50 on a Basic account across the same bands (50 from day 15), and 50, 60, 80 and 100 on a Premium account (100 from day 15). You can set a lower limit for any action, and the lower setting wins.
What LinkedIn publishes, and what it does not
LinkedIn does not publish a number for invitations. Its help page "Invitation limit reached" (read 24 October 2026) says LinkedIn sets invitation limits to protect member experience, that every member, Basic and Premium, is subject to them, and that reaching the limit usually brings a restriction that lasts about one week. During a restriction you cannot send invitations, withdrawing pending invitations does not lift it, and LinkedIn Support cannot say which type of restriction it is or why. Any weekly figure you read elsewhere is a community observation.
What LinkedIn does publish is the allowance for personalised notes. This matters for strategy, because on a Basic account "note or no note" is not a free choice. The same "Invitation limit reached" page says Basic members can add a personalised message to five connection requests a month. A second page, "Personalize invitations to connect" (read the same day), says up to three a month, with up to 200 characters each. The two pages disagree, so check what your own account shows. Both say Premium members can send unlimited personalised invitations. Plan the notes for your best-fit prospects and expect most Basic-account requests to go without one.
The ceilings in the table are Warmerly's. LinkedIn's own limits are not published and can differ by account. Treat both as ceilings, not targets, and stop at the first warning prompt.
How the sequence waits for an acceptance
A message can only go to a first-degree connection, so a sequence that starts with a connection request has to wait. In Warmerly the first message is held until the invitation is accepted. The connection is re-checked every six hours, no message budget is spent while it is pending, and a lead that has still not connected 21 days after the invitation is dropped from the sequence instead of being retried. Plan the calendar around that window: an invitation sent on day 0 can still produce a message on day 20.
| When | Step | Stop rule |
|---|---|---|
| Day 0 | Connection request, with a note only if you have one to spend | Waits for the acceptance |
| On acceptance | First message: one question, no pitch | Leaves on a reply |
| A few days later | Follow-up one: a relevant example or resource | Leaves on a reply |
| About a week after that | Follow-up two: a soft last nudge | Leaves the sequence, whatever happens |
| 21 days after the invitation | Still not connected | Dropped from the sequence |
How long it takes to learn anything
The weekly ceiling decides how fast you can find out whether a strategy works, because a rate is only worth reading on a decent sample, and our guide to measuring your response rate suggests at least 100 to 150 delivered invitations. On a Basic account at the weekly ceiling of 50, 150 invitations take three weeks, and longer in the first fortnight, when the ceiling is lower. On a Premium account at 100 a week it takes a week and a half. That is the real reason to start with one segment, one message and one change at a time: you only get a few readable tests in a quarter.
Work backwards from what you need. Meetings wanted a month, divided by those three rates multiplied together, gives the invitations to send. You do not yet have the rates, and this page does not supply them. Fill in the three blanks from your first sample and the volume you need follows.
| Number | How to work it out | Yours |
|---|---|---|
| Acceptance rate | Accepted requests divided by requests sent | |
| Positive reply rate | Positive replies divided by messages delivered | |
| Meetings per positive reply | Meetings booked divided by positive replies | |
| Invitations a month you need | Meetings wanted, divided by all three rates multiplied together |
If that number is above what the weekly ceiling allows, the answer is not to push past it. Add a second channel: email goes to the same people without spending LinkedIn budget, and Warmerly runs both from one sequence. See email and LinkedIn together.
Measure the right things
- Connection acceptance rate: is your targeting and request note working?
- Reply rate: is your first message earning a response?
- Positive reply rate: are the conversations qualified?
- Meetings booked: the only metric that pays the bills.
A worked example: one ICP, start to finish
Say you sell scheduling software to multi-location dental groups. A vague ICP would be "healthcare operations people." A usable one is: practice managers and operations directors at dental groups with several locations, in one country, who have recently opened a new site or hired for a front-desk role. Each clause removes people who would never buy, and the trigger (a new site or a hiring post) gives you a true, specific reason to write.
From that definition, the rest of the strategy falls out almost mechanically. The search filters come from the role and company clauses. The connection note references the trigger. The first message asks how they are handling the problem the trigger creates. The follow-ups bring a relevant example from a similar group. Nothing in the sequence is generic, because the ICP was not generic.
The decisions a strategy has to make
| Decision | The question to answer | Example answer |
|---|---|---|
| Who | Which role at which kind of company, and what makes now the right time? | Practice managers at multi-site dental groups that just opened a location |
| Why you | What does your profile say within five seconds of a visit? | Headline names the audience and the outcome, not just your job title |
| First touch | Note or no note, and what is the one true reason for it? | Short note that mentions the new location |
| The ask | What is the smallest next step that is still useful to you? | One question about how they handle scheduling gaps today |
| Volume | How many new people per day, and how fast do you ramp? | A small daily number, raised only while account health stays good |
| Stop rule | When does a prospect leave the sequence? | On any reply, on a decline, or after the final follow-up |
Your first two weeks, step by step
- Day 1: write the ICP in one paragraph, including the trigger. If you cannot name a trigger, pick the closest proxy, such as a hiring post or a role change.
- Day 1: rewrite your headline and about section for that audience. Ask a colleague who has never seen it what you do, and fix whatever they get wrong.
- Day 2: build a first list of a few hundred people and read through a sample by hand. If a quarter of them look wrong, tighten the filters before you send anything.
- Day 3: write the four-step sequence and read every message out loud. Cut any sentence you would not say to someone at a conference.
- Days 4 to 7: send a small number of requests per day and answer every reply yourself. You are learning what people actually say back.
- Days 8 to 14: compare acceptance and reply rates by segment. Keep the segment that performs, change one variable in the one that does not, and ramp volume only if the account stays healthy.
Common strategy mistakes and the fix for each
| Mistake | Why it hurts | Fix |
|---|---|---|
| Starting with the tool | You send volume before you know what works, and burn a good list on a bad message | Run a small manual test first, then automate what worked |
| Targeting a title instead of a situation | Everyone with the title gets the same message, so most of it misses | Add a trigger or a company-level filter to every list |
| A pitch in the first message | The recipient sees an ad, not a conversation | Ask one question and keep the pitch for the reply |
| Changing five things at once | You cannot tell which change moved the numbers | Change one variable per test and give it enough sends to read |
| Ignoring the profile | Interested people visit it, see nothing relevant, and leave | Treat the profile as the landing page for every message you send |
| Counting messages sent as progress | Activity metrics hide that no one is replying | Report on positive replies and meetings, not sends |
Reading your numbers
Your metrics tell you which part of the strategy to fix, if you read them in order. Start at the top of the funnel and only move down when it is healthy.
| What you see | Most likely cause | What to change first |
|---|---|---|
| Low acceptance rate | Wrong people, a salesy note, or a weak profile | Tighten the list, then test no note against a short note |
| Good acceptance, few replies | The first message reads as a pitch or asks for too much | Shorten it and replace the ask with one easy question |
| Replies, but few are positive | Targeting is close but not right | Review the objections you hear and exclude that segment |
| Positive replies, few meetings | The handoff is slow or the call-to-action is heavy | Reply within the same day and offer two specific times |
Edge cases worth planning for
- A very small market: if your whole ICP is a few hundred people, you cannot afford to burn them. Send fewer, more personal messages and use email as the second channel instead of repeating on LinkedIn.
- Several reps on one list: agree on one owner per prospect before anyone sends. Two people from the same company messaging the same buyer looks careless.
- Long sales cycles: a prospect who says not now is not a no. Add them to a list you revisit on a calendar date, and send something genuinely new when you do.
- Existing customers and warm contacts: keep them out of cold sequences entirely. A cold template sent to someone you already know does damage that one message cannot repair.
A good strategy works at a small scale by hand first. Automating it later saves time, but it does not remove the risk that LinkedIn restricts an account that behaves unnaturally. Treat every volume number as conservative practice, since LinkedIn does not publish exact limits.