The agency that pitched our client before us sent 11,000 LinkedIn connection requests in 90 days. They got 47 replies. Three were interested. One became a meeting.
We ran 74 accounts over the same period. We got 14 replies. Nine were interested. Four became meetings.
The difference was not the copy. It was not the timing. It was who was on the list and why they were on it.
Why Volume Lists Fail
The volume approach is based on a false assumption: that outbound is a numbers game where more always means more. It is not. It is a targeting game where relevance compounds.
When you send 10,000 messages to a loosely filtered list, three things happen. First, your reply rate drops below 1%, which signals to LinkedIn and email providers that your messages are unwanted. Second, your deliverability degrades. Third, you spend three hours a week managing replies from people who were never a fit.
The cost of a bad list is not just low conversion. It is infrastructure damage.
A high-volume list does not just underperform. It actively harms future campaigns by flagging your accounts as spam sources. The recovery time for a damaged LinkedIn account or sending domain can be months.
The Signal-First Framework
Signal-first list building starts with a question most agencies skip: what is happening in this account right now that makes them more likely to need what we offer?
That question leads to signals. Not demographics. Not firmographics. Events.
The 5 Signals We Track
1. Hiring Signals
Job postings are the most reliable indicator of budget and priority. A company posting for a Head of Sales is about to build a pipeline. A company posting for a BDR is about to run outbound. We track this through Apollo's job change feed and Apify scrapers on LinkedIn.
The key is not just what they are hiring for. It is how recently and at what volume. Three sales hires in 30 days means something different than one hire posted six months ago.
2. LinkedIn Activity
When a founder or executive starts posting about a specific pain point, they have usually already decided to do something about it. We monitor LinkedIn for posts from decision-makers that mention problems we solve: "scaling outbound," "building a sales team," "too much manual work in our sales process."
Serper.dev runs daily searches against these keyword patterns and flags accounts for review.
3. Funding Events
New funding means new budget, new priorities, and a 90-day window where leaders are actively building the systems they need to hit their new targets. Series A and B rounds are the sweet spot. Seed is usually too early. Series C and beyond usually means they have the team already.
4. Tech Stack Changes
When a company adds a CRM, a sequencing tool, or a sales engagement platform, it signals that they are actively investing in their go-to-market motion. We track this through Builtwith and Clay's tech stack enrichment. A company that just adopted HubSpot is a different prospect than one that has been on Salesforce for five years.
5. Web Mentions and Community Activity
Reddit posts, G2 reviews, and Quora answers are underrated signals. When someone posts "what do you use for cold email outreach" in a founder community, they are at the beginning of a buying journey. We monitor this through Serper and manual community checks for our highest-value ICP segments.
How We Score Signals
Not all signals carry equal weight. A funding event on its own is weak. Combine it with three sales hires and a LinkedIn post about building pipeline, and that account jumps to the top of the list.
Accounts that score 12 or above get moved into the active outreach pool. See the full scoring breakdown for per-signal weights. Accounts between 8 and 11 go into a watch list and get checked again in two weeks. Below 8, they stay in the research queue.
Why 50 Accounts Instead of 500
The instinct is to maximize the list. More accounts, more chances. Here is why that logic breaks down in practice.
Personalization quality drops at scale. When you have 50 accounts, you can write a first line that references the specific role they just hired for, the funding round they just closed, or the LinkedIn post they wrote last week. When you have 500 accounts, you default to generic openers.
Generic openers get filtered. Not by spam detectors. By humans who receive ten cold messages a day and can tell in three words whether someone did any research.
What Gets Excluded
Signal-first also means filtering hard on the negative side. We exclude:
- Accounts where the decision-maker has been in the role less than 90 days (still finding their footing, not buying)
- Companies that were recently acquired (budget and priorities unclear)
- Accounts where we have already reached out in the last six months with no reply
- Any contact at a company where a teammate is already in conversation
The exclusion list is as important as the inclusion criteria. Every list also runs through a 7-point quality audit before launch. A tight list of genuinely qualified accounts is worth ten times a bloated list of maybes.
Build your exclusion criteria before you build your targeting criteria. Knowing who you are not reaching out to forces clarity about who you actually are.
Frequently Asked Questions
How do you find signals at scale without a data team? We use a combination of Clay for enrichment and waterfall data, Serper for Google-based signal detection, Apify for LinkedIn scraping, and Apollo for job change monitoring. Most of this runs automatically. The output is a scored list that we review weekly. Total manual time: about two hours per client per week.
What if there are not enough accounts hitting the signal threshold? That is useful information. It means either the ICP is too narrow, the signals are too restrictive, or this is genuinely a small market. Better to know that before you launch a campaign than six weeks in.
How often do you refresh the list? Weekly. Signals decay. An account that was a strong fit three months ago may have hired the person they needed, changed direction, or already bought from a competitor. Fresh signal data changes the priority order significantly.
Does signal-first work for all B2B products? It works best for products where buying is triggered by a specific event or circumstance. If your ICP is defined more by company size and industry than by what is happening in the account, you may need to rely more on demographic filters. But even then, layering in any one or two signals improves conversion meaningfully.
Tight lists of signal-triggered accounts outperform volume lists every time. The signal is the targeting.
Want to see how we would build this for your ICP? Book a call.