By I sell LinkedIn automation, and the advice I give most often to new founders is to hold off on using it for a week.
Automation is a multiplier. It takes whatever your outbound already is and does more of it. If what you have is a working message to the right people, you get more conversations. If what you have is a guess, you get the same guess delivered to five hundred people, and you burn a list you can't get back.
So before you turn anything on, send 100 connection requests yourself. At 20 a day that's one working week. It's tedious on purpose. You're not doing it for the pipeline. You're doing it to answer five questions that no tool can answer for you.
1. Can you write your buyer down as a search filter?
Most founders describe their customer in a sentence that sounds fine in a pitch deck and falls apart in a search box. B2B companies that care about growth is not a filter. Head of Operations at logistics companies with 50 to 200 employees in the UK is.
Picking 100 people by hand forces the issue. You'll notice yourself skipping certain profiles and lingering on others, and those instincts are the real definition of your buyer. Write down why you skipped someone. By the end of the week you should have titles, a company size range, a geography, and a short list of exclusions. If you can't produce that, software can't either. It will happily fill your campaign with people who match a vague description.
2. Does your acceptance rate clear the baseline?
You need a number to compare against, so here's ours. Across 8,400 real connection requests sent through Linkdify, the average acceptance rate was 25.2%, about one in four. The full breakdown is in our LinkedIn outbound benchmarks.
One detail from that data matters for your test. Requests sent with a note were accepted 8.6% of the time. Requests sent without one were accepted 48.6% of the time. So run your 100 without a note, or you're testing your note-writing instead of your targeting.
If you land far below one in four with blank requests, the problem is usually one of two things: your profile doesn't look like someone worth knowing, or your list is wrong. Both are cheap to fix at 100 requests. Both are expensive to discover at 2,000.
Want LinkedIn outreach that runs itself?
Linkdify finds your leads, writes personalized messages, and sends them at safe, human-paced limits.
3. Do you have one message that got a real reply?
Not a message you like. A message a stranger answered.
Our benchmark here is a 14.6% reply rate across 2,511 messages. With roughly 25 accepted connections from your 100 requests, that works out to three or four replies if you're average. That's a tiny sample, and I wouldn't draw statistical conclusions from it. What you're after is the text of those replies.
Read them closely. People will tell you which word in your message they reacted to, what they assumed you sell, and what they're actually worried about. I'd trade a dashboard for four honest replies at this stage. Rewrite the message using their language, send it to the next batch of accepts, and see if it holds. Once one version keeps earning answers, you have something worth multiplying.
If you get zero replies from 25 conversations, that's information too. Change the message or change the audience before you scale either one.
4. Do you know what happens after someone says yes?
This is the question founders skip, and it's where most of the waste is.
In our data the median reply arrived about 14 hours after the message went out, and most replies landed within the first day. That means interest shows up fast and in a narrow window. If your plan for a positive reply is to think about it for two days and then send a calendar link, you've spent the moment.
Doing the first 100 by hand makes you build the second half of the conversation. What do you ask next? When do you offer a call, and when do you just answer the question? What do you say to not now? By the end of the week you should have a handful of replies you've written more than once. Those become your playbook. Automating the top of the funnel without them just produces more conversations you handle badly.
5. How many conversations can you actually hold?
A solo founder who is also building the product and doing support has a real ceiling on how many live threads they can keep warm. For some people it's five a week. For others it's twenty. You find out by doing it.
This number should set your sending volume, and almost nobody uses it that way. The usual instinct is to send as much as the account can safely handle and sort out the replies later. I think that's backwards. The job of automation is to keep your conversation capacity full every week without you thinking about it. Past that point, extra volume turns into slow responses and dropped threads, and a prospect who was ignored once is harder to win than one you never contacted.
Want LinkedIn outreach that runs itself?
Linkdify finds your leads, writes personalized messages, and sends them at safe, human-paced limits.
What changes after the week
If you come out with a filter, an acceptance rate near or above one in four, a message that earned replies, a rough playbook for what comes next, and a sense of your own capacity, you're ready. Hand the repetitive part to software. This is the point where a tool like Linkdify earns its cost, because finding the next 500 people who match your filter and pacing the requests is work a machine does better than a tired founder at 11pm.
If you come out with gaps, you saved yourself money and a lot of wasted leads. Fix the gap and run another 100.
The founders I see get the most from automation are rarely the ones who started it earliest. They're the ones who could tell me, before they turned it on, exactly who they were writing to and which sentence got the reply.