By The most common support message I get reads like this: “The campaign was doing great for two weeks, now nobody accepts. Should I change the opener?”
Almost every time, the opener is fine. What changed is the list. The campaign burned through the people it was built for, and something started filling the gap with people it was not built for. I know this because, for a while, that something was my own software.
Campaigns have a half-life
A LinkedIn campaign is a finite list of people with a message attached. Founders treat it like a faucet. You set it up, it runs forever, and if replies drop you fiddle with the copy.
But at a safe daily pace, a search that returns a few hundred good matches is a few weeks of sending. The first batch is the best audience you will ever have for that message, because you wrote the message with exactly those people in mind. Week one looks great. Week three looks worse, and the temptation is to A/B test openers on an audience that no longer exists.
Our own published benchmarks put the average acceptance rate at 25.2% across 8,400 connection requests, with a 14.6% reply rate across 2,511 messages. Those are averages over whole campaigns. The early weeks of a tight campaign run above them and the tail runs below. Blend the two and you get a number that hides what actually happened.
Where the bad leads come from
Here is the part I had to learn the hard way. When a campaign runs out of matching people, an automation tool has two options: stop, or relax the filters and keep going. Stopping looks like a bug to the user. So most tools, mine included at the time, kept going.
Linkdify's lead search used a relaxation ladder. If the exact search came back empty, it loosened the criteria one notch and tried again. That sounds reasonable until you watch what it does. A campaign aimed at wellness founders drifted into pharma companies and clinics, because the industry filter matched “health, wellness and fitness” and that bucket is enormous. A campaign with one specific keyword that zeroed out the search slid, notch by notch, down to location only. The user had asked for a particular kind of person in Austin and ended up messaging random people in Austin.
From the dashboard, that looked like a campaign with falling acceptance. From the lead list, it looked like a different campaign wearing the first one's message.
I changed how it works. Keywords relax first, because free-text keywords are the filter most likely to zero a search without meaning much. Location never relaxes. A guard now checks each lead's primary industry after enrichment, so a secondary tag can no longer pull a clinic into a founder campaign. And bring-your-own-list campaigns don't get topped up at all. If you uploaded a CSV, you meant that list, not that list plus whatever the system could scrape together.
Want LinkedIn outreach that runs itself?
Linkdify finds your leads, writes personalized messages, and sends them at safe, human-paced limits.
How to tell which problem you have
You don't need my data to diagnose this. Open your campaign and compare the people contacted in the first week with the people contacted last week.
- Same titles and the same kinds of companies? Then you might genuinely have a copy problem. It's rarer than people think, but it happens.
- Titles drifting and industries getting vaguer? The audience ran out. Your message is now being judged by people it was never written for.
- Acceptance fell before replies did? That's the strongest exhaustion signal. Strangers who don't recognize themselves in your profile never accept, so they never read the message you keep rewriting.
What I do instead
Size the audience before you launch. If the search returns 240 people, that is a 240-person campaign. Plan the message and the timeline around that number instead of pretending it's infinite.
Run several narrow campaigns rather than one wide one. A precise persona with a message written for it will beat a broad campaign with a message that tries to speak to everyone. Narrow campaigns exhaust faster, and that is fine. Exhausting is the point. It means you reached everyone you meant to reach.
When the list ends, end the campaign. Don't widen it. A new audience deserves a new message, so make it a new campaign. The only thing a broadened campaign inherits is a message that no longer fits.
Treat uploaded lists as sacred. If you hand-picked 150 names, no tool should be adding number 151 without asking you.
Know what healthy looks like. Roughly one in four accepts. Roughly one in seven messaged leads replies, and the median reply lands in about 14 hours. If your first week is near those numbers and your third week is far below them, go look at who you contacted in week three before you touch a word of the copy.
Want LinkedIn outreach that runs itself?
Linkdify finds your leads, writes personalized messages, and sends them at safe, human-paced limits.
The takeaway
A campaign that stops working usually didn't stop working. It finished. The dip you are staring at is the system continuing past the end of the list with worse and worse approximations of the person you wanted. No sentence fixes that. A fresh list does, along with a tool honest enough to tell you when the old one is done.