How we mapped Artie's market and showed which accounts email could never reach

Introduction

Artie is a Y Combinator-backed data infrastructure company in San Francisco. Its platform replicates production databases like Postgres, MySQL and MongoDB into warehouses such as Snowflake in seconds, using change data capture, so data teams stop maintaining batch pipelines. Customers include Substack, ClickUp and Alloy. Artie sells top-down to heads of data, VPs of engineering and CTOs, with deals around $100K. When the founders came to us in October 2025 they had just raised a Series A, had a team of six, and were about to hire a sales team.

The Challenge

Artie’s pipeline ran on one channel, and that channel had a ceiling.

Email had died → cold email worked in 2023, then slid into spam folders through 2024. By late 2025 the founders called it “completely useless to us right now” and had stopped using it.
LinkedIn was capped → founder profiles got 10 to 15% reply rates, but 25 connection requests a day cannot feed a growing sales team.
A niche, hard-to-see buyer → the people who own a data pipeline rarely carry the title. Messaging had stayed broad because nobody could tell who ran Postgres and Snowflake.
No map of the market → Artie did not know how many accounts existed, how many had a data team, or how many could be emailed at all.

What We Built

We rebuilt the list from the company down, then matched every account to a channel.

1. Evergreen signals. Four campaign types that refill themselves: data people new in role, companies hiring data engineers, detected tech stack, and a fallback. Later, an AI step read each open data-engineering job post and wrote the first line from what the company was actually hiring for.

2. Company before persona. The qualification was split in two. First, does this company have a data team at all, using Sumble’s tech-stack and team data and Crunchbase. Only then find the people, scoring keywords across their whole LinkedIn profile rather than their title, tiering them, and cutting individual contributors entirely. Disqualified companies flowed straight into the Smartlead blocklist. The result was 62,808 double-qualified leads.

3. One visitor, two hundred prospects. Artie’s website-visitor flow emailed at most two people a day. We rebuilt it in Clay: qualify the visiting company, find 20 to 50 lookalikes with Ocean.io, find the data people at each with Surfe, Icypeas, Prospeo and Leadmagic, validate twice, and run them in their own evergreen campaign. One visitor became 80 to 200 reachable people, with 5,000 leads queued.

4. A deliverability map of the market. We checked the mail provider behind every one of 2,160 target accounts: 63% on clean Google Workspace domains, 18% Outlook-only, 19% behind Barracuda, Mimecast or Proofpoint. The last group never sees cold email. Each segment got its own treatment: full sequences, under-40-word emails for Outlook, and LinkedIn or calls for the gateway-protected accounts.

5. LinkedIn from the founder’s profile. HeyReach sequences from the CTO’s account carried the same first line and bullets as the email, aimed at the accounts email could not reach, with a live dashboard showing accounts and contacts touched.

The Outcome

Email never became a volume channel for Artie. The map of the market did its job.

• 62,808 double-qualified leads → a data-pipeline buyer list Artie did not have before, built company-first and stripped of individual contributors.

• 37% of accounts flagged before budget was spent on them → Outlook-only and gateway-protected companies were identified and routed to LinkedIn and calls instead of burning sends.

• 80 to 200 prospects per website visitor → the visitor flow went from two emails a day to an evergreen lookalike engine with 5,000 leads ready.

• An outbound stack Artie kept → when the team grew from 6 to 15 and took outbound in-house in March 2026, we handed over the sending infrastructure, lead lists and suppression data they now run on.

Why It Worked

1. Company before persona. Asking “does this company have a data team?” before “who is the CTO?” cut the noise that broad title searches had produced.

2. Knowing what email cannot reach. Mapping mail providers across the market turned a deliverability problem into a routing rule, account by account.

3. Assets, not just sends. Lists, lookalike workflows and sending infrastructure were built so Artie could carry them in-house, which is exactly what happened.

Ready To Own Your Pipeline?

If your buyer is hard to see, your email lands in spam, and your only working channel has a daily cap, this is the exact gap we close. We map the market from the company down, find the people who own the problem, and route every account to the channel that can actually reach it.

Book an outbound strategy call and we’ll map what an owned outbound channel would look like for you.