Are you filtering prospects *before* sending, or are you letting ai personalization fix bad targeting?

I’ve been thinking about this a lot, and I realized I might be approaching this backwards. I’m currently pulling a large list of prospects by broad criteria—basically anyone in my ICP—then relying on LiSeller’s personalization to make the outreach land. But the more I think about it, the more I wonder if I should be getting strict about filtering upfront.

The argument for filtering first: I spend time now identifying truly high-intent prospects, verify they’re actually a fit, then send hyper-personalized stuff to a smaller, better list.

The argument for casting wider: More volume, and the personalization might convert anyway. Plus, I’m not sure how to efficiently filter 500 prospects down to real high-intent ones without spending 4 hours manually reviewing.

I’m averaging around 4% reply rate on my bigger lists, and I’m wondering if that would jump significantly if I was more selective. Or am I overthinking this and just need better personalization? Curious how others are approaching this—do you filter hard, or do you rely on volume + good AI?

You’re actually asking the wrong question. It’s not one or the other—it’s both, but in the right sequence. You need to filter for intent signals, not just ICP fit. An ICP check just tells you they’re a potential customer. Intent tells you they’re ready to listen.

Intentity signals: 1) Recent job changes, 2) Company news/funding announcements, 3) They’ve engaged with content related to your solution, 4) Specific technologies or keywords in their profile. Start there. That drops your 500 to maybe 150-200. Then personalize heavily for those 150.

4% on a big list usually means you’re hitting a lot of people who have no reason to care. 10%+ on a smaller, intent-filtered list is way more achievable. Volume only matters if the quality is there.

I use a combo of LiSeller’s smart lead filtering plus some custom logic I’ve built in Zapier. Here’s my stack: Pull prospects based on basic ICP, then use LiSeller to flag high-intent signals. Then I pipe those through a webhook to a Google Sheet where I manually validate top 30-40 per week. Pure math game—if I send to 500 nobody, I waste time. If I send to 50 perfect fits with great personalization, I get 8-12% response rate.

Take the time to set up your filtering criteria now. It pays dividends later. What specific intent signals are you currently tracking?

For recruiting, I always filter first. High-level talent is allergic to spray-and-pray outreach. They can feel when they’re one of 500. I spend 30 minutes upfront identifying people who actually match the role—not just the title, but the specific tech stack, background, what they’re probably looking for next.

Then when I send a message, it’s so specific that the personalization matters because I’m not saying generic stuff. “I noticed you built the infrastructure team at [Company], and we’re growing a similar function here.” That works because I actually verified it matters.

Smaller list, better results. Always.

Safety angle: sending to 500 randoms looks terrible for account health. LinkedIn’s algorithm notices if you’re connecting to people with no real profile interactions, no mutual connections, misaligned backgrounds. Your account gets flagged as suspicious.

Filtering first actually protects your account. You’re sending to people who look like a real, intentional network. That keeps your account warm and your deliverability high. Plus, your reply rates are better, so you’re not burning through your outreach faster for worse results. Win-win.

Both are important, but the sequencing matters. LiSeller’s smart lead filtering was specifically built to help with this—it identifies buying signals and engagement patterns that predict higher response rates.

Here’s how I’d suggest approaching it: Use LiSeller’s built-in filtering to segment your 500 into tiers based on intent signals, then personalize heavily for Tier 1 (your highest-intent prospects). You get volume and quality. Setting this up takes maybe an hour, then it recalculates automatically.

What filtering criteria are you currently using in LiSeller? Let me know and I can help you refine it.

The data is clear: 4% on 500 = 20 responses. 8% on 150 = 12 responses with a fraction of the effort and account risk.

But here’s the strategic insight: Intent matters more than volume at scale. You want your list small enough that you can truly personalize. Not just AI-fill-in-the-blanks personalization—real personalization that shows you actually know why they matter.

Filter for intent first. Recent job changes, active profile engagement, company news. Then personalize. That’s the winning formula.

What’s your current filtering taking into account?