The Standard ICP Workflow Has a Structural Flaw
Most outbound teams build target lists the same way: pick a persona, pick a title, pull the list, work it. The logic feels sound. In practice, it quietly kills coverage.
The problem isn't the persona — it's the sequence. When you pre-filter by assumed job title, you're betting that your internal language maps cleanly onto how target companies actually structure and name their roles. In verticals like manufacturing, logistics, or healthcare, that bet loses constantly. The person who owns the budget for a RevOps or operational AI decision might be called Director of Continuous Improvement, VP of Supply Chain Excellence, or Head of Clinical Operations. None of those show up on a "VP of Operations" pull.
Jordan Crawford quantified the gap: teams that pre-filter by title miss 40–60% of viable buyers at their own target accounts. That's not a list-quality problem. That's a methodology problem.
Invert the Query Order
Crawford's fix is straightforward and executable this sprint. Instead of starting with an assumed title and filtering down, you start by scraping the full job-title universe at your target accounts — every title that exists — and then apply your filters against real data rather than assumptions.
The workflow is AI-native. You can run it in Claude Code or a comparable pipeline tool without a new data vendor. The change is in query order, not tooling. Pull first, filter second.
For a non-tech operator running outbound against a fragmented buying committee, this resequencing does two things immediately:
- Expands addressable pipeline without inflating your ICP — you're finding real buyers, not adding noise.
- Surfaces title conventions you didn't know existed at accounts you've already been targeting, which means your next sequence can speak their language rather than yours.
The Actionable Move This Week
Pick one high-priority target account where your current list feels thin or your open rates are underperforming. Run a full title scrape against that account — Claude Code can do this with a prompt, no engineering ticket required. Then filter the output against your buying criteria.
Compare what you get against what your current list shows. If there's a gap, you've just found the exact problem to fix across your full account list. If there's no gap, you've confirmed your methodology and can move on.
This is a one-sprint diagnostic with direct revenue implications. The teams that do it now will enter Q4 planning with a materially more accurate picture of who's actually in their market.