Strategy · 2026-09-07 · 12 min

LinkedIn Outbound for B2B SaaS and Dev Agencies: The 2026 Playbook

What still works in LinkedIn outbound for B2B SaaS companies and dev agencies in 2026: how buyers read messages now, how to build the account list, what the first message must contain, and the numbers to hold yourself to.

Two things happened to LinkedIn outbound over the last two years. Volume got cheaper, and attention got more expensive. Anyone can now send a thousand well spelled messages a week. Which is exactly why a thousand well spelled messages a week no longer produce anything.

Buyers adapted. A CTO who receives fifteen approaches a week has built a filter, and that filter runs in about three seconds. It is not reading your message. It is scanning for one thing: is there any evidence that a human looked at my company before writing this?

This is the playbook we run in 2026 for B2B SaaS companies and dev agencies. It is not a list of hacks. It is the sequence that survives the three second filter.

What Changed, and What Did Not

What changed is the baseline. Correct grammar, a first name, a company name and a tidy value proposition used to be enough to earn a reply. Now they are the signature of everything the buyer already ignores. Polished and generic reads worse than plain and specific.

What did not change is simpler. People still reply to messages about their own situation. They still open the first message after accepting a connection, at close to a 98 per cent rate, because it lands as a phone notification rather than a promotions tab. And they still buy from people who are already in their network when the timing finally turns.

So the channel is intact. The method has to carry more weight than it used to.

The Two Markets Behave Differently

We work with two audiences, and treating them the same is the fastest way to waste a quarter.

B2B SaaS

SaaS is a committee purchase. The person who feels the problem is rarely the person who signs, so the first conversation is an entry point rather than a close. Timing dominates. Software gets bought when a contract renews, a system breaks, a leader arrives with a mandate or a team doubles in size. Your job is to be inside the account and remembered before that moment, not to argue someone into urgency they do not have.

Practical consequence: multi threading is not optional. One accepted connection at a target account is a foothold. Three is a position.

Dev Agencies

Agencies sell capacity, and capacity has a clock on it. A bench that is idle in November is money that is already gone. Referrals are excellent and completely unschedulable, which is the whole problem: they arrive when they arrive, and never when the bench opens.

Practical consequence: the target list is built from live technical need. Companies hiring three backend engineers, migrating a stack, shipping a mobile product with no mobile team, or opening a market they have no engineers for. These are visible from the outside if someone bothers to look.

Step One: A Small List You Can Defend

Most outbound fails before a single message is written, at the list stage. A list of 4,000 companies is not a list, it is a spreadsheet of hope.

We work from a few hundred named accounts per profile, chosen because each one has a plausible reason to care. For a SaaS client that means firmographics plus a real trigger. For an agency it means an observable technology need. If we cannot write one sentence explaining why a company is on the list, it comes off the list.

A smaller list is not a limitation. It is what makes the research layer possible at all.

Step Two: The Research Layer

Before any message exists, we read. The company site, recent posts from the person and the company, job adverts, funding news, product releases, and where relevant the customers they publicly work with. This takes minutes per prospect and it is the entire competitive advantage.

The output of research is not a compliment. Nobody replies to "loved your post about culture". The output is a hypothesis about a problem the company plausibly has right now, in language they would recognise as their own.

The connection request itself goes out clean, without a note. The note field adds friction and gives the recipient something to dismiss before they have looked at the profile. The profile does the arguing.

Step Three: The First Message, in Three Sentences

This is the highest value message in the entire engagement, and it is written from scratch, per person. Its structure is fixed:

Their problem. One sentence naming the specific situation you believe they are in, based on what you read. Not a category. Their situation.
How we solve it. One sentence on what we do about exactly that, with no feature list and no adjectives.
A clear ask. A direct request for a short intro call.

That is the whole message. Three sentences, no attachment, no deck, no paragraph of positioning. If the first sentence is wrong, nothing after it matters, which is why the research comes before the writing rather than the other way round.

The reason for the direct ask is respect for the reader's time. Softening the ask into a vague question does not increase replies, it just makes the exchange longer and less honest. Senior people prefer to know what you want.

Step Four: Follow-Ups That Do Not Nag

Most replies do not come from the first message. They come from the second or third, sent days apart, and each one has to add something rather than repeat the request.

A relevant follow up looks like a specific observation, a short piece of proof from a comparable company, or a question that is easier to answer than the first one. Three to four touches over two or three weeks, then stop. Silence is an answer, and pushing past it costs you the account permanently.

We send these follow ups on a sensible cadence with tooling, because the sequencing is mechanical. The content is not. The distinction matters: what we refuse to do is send messages with no context behind them, regardless of how they are delivered.

Step Five: Qualify Before the Handoff

A booked call with the wrong company is worse than no call. Before anything reaches your calendar we check that the company is roughly the right size, the person is close enough to the decision, and the need is real rather than polite curiosity.

This is why we count qualified conversations rather than meetings. Meetings are easy to inflate. Conversations with people who can actually buy are not.

Numbers Worth Holding Yourself To

Honest ranges on a researched motion, per LinkedIn profile:

Metric	Realistic range
Invites sent	Up to 150 per week
Connection acceptance	35 to 55 per cent
Qualified conversations	Up to 15 per month
First replies	Typically from week 2 to 3
First booked calls	Typically from week 4 to 6

Note what is missing: closed deals in month one. In high ticket B2B, a first conversation in October is revenue in February. Anyone quoting you a faster figure is selling you something they cannot deliver. We wrote about the money side of this in Why You Need a Six Month Budget.

One profile is excellent for validating that your market responds to your message. If the goal is closed revenue rather than validation, volume matters, and that is an argument for running more than one profile rather than waiting longer.

When This Is the Wrong Channel

We would rather say this early than take the money.

Your average contract value is a few hundred euros a year. The maths does not work; the research cost per prospect exceeds what the customer is worth.
You sell to a role that barely uses LinkedIn.
You cannot commit six months. Three months of outbound produces a list of warm relationships and almost no closed revenue, which reads as failure to anyone who wanted a quick win.
Your sales team cannot take the calls. Generating conversations you cannot follow up on damages the network you just built.

The Part Nobody Talks About

Every accepted connection stays yours. When someone says "not now" in March and their situation changes in September, they are already connected to your founder, already exposed to whatever you post, and already familiar with your name. A significant share of the deals we have seen close came from exactly this: an approach that went nowhere the first time, revisited later.

That is the compounding asset in this channel, and it is the reason volume based approaches end up costing more than they save. Every burned prospect is an account you cannot go back to.

Where to Start

Pick twenty five companies you would genuinely like to work with. Write one sentence for each explaining why now. If you cannot write that sentence for at least fifteen of them, the problem is not your outreach, it is the target definition, and no message will fix it.

If you would rather have this run properly from the start, see how we work, the SaaS breakdown, the dev agency breakdown and the pricing.

Related Dispatches

Outbound glossary · How it works · Pricing