Working stacks, costed to the cent
A directory tells you what exists. A stack tells you what to wire to what, in which order, and what it costs per contact when it runs. Four here, each one a real motion with the arithmetic shown rather than asserted.
Three questions, in the same order
Which providers, in which sequence, and what one unit of output costs when it runs.
Every stack below answers those three questions. The last one is what decides whether a motion survives its first quarter, and it is the one almost nobody works out before building.
The four are waterfall email enrichment at $0.0364 a contact, hiring signal to outbound at $0.13, inbound speed to lead at $0.059 a lead, and agent-driven account research at $0.09 an account.
Each one names the providers it uses, and every provider in them is in the directory with verified rates.
The prices are verified from vendor rate cards. Where a figure depends on a match rate, the page says so and shows the assumption, because no provider in this category publishes a match rate you can hold them to.
Start at the marginal cost column, not the total. A stack's average cost per contact always looks cheap. The marginal cost of its last step is what tells you whether that step should exist at all, and on most of these it is 20 to 30 times the first step.
Four stacks, filtered by what you are trying to fix
Filter by the stage of the motion, by how much build effort you can spend, or by a provider you already pay for.
Waterfall email enrichment
Four providers chained cheapest-first, each taking only what the last one missed, then verification routed per source.
Hiring signal to outbound
A company hiring three SDRs has already decided to spend. Turn that job posting into a sequenced contact inside a day.
Inbound speed to lead
Form submission to enriched, verified, routed and sequenced in under a minute, with no human in the path.
Agent-driven account research
Point an AI client at two MCP servers and have it write a researched account brief straight into the CRM.
Five rules that survive every stack
Learned the expensive way, on client pipelines.
Cheapest provider first, always
Sort by cost per successful lookup ascending and pass only the misses down the chain. In the waterfall stack, reversing the order takes the bill from $303 to roughly $417, buying 5% more contacts for 38% more money. It is a config change.
Qualify before you enrich
The cheapest lookup is the one you never make. Firmographic and signal filters cost a fraction of a cent, contact enrichment costs 20 to 50 times that. Cutting a list from 10,000 to 4,000 before enrichment saves 60% and usually converts better, because the rows you dropped were never going to be contacted.
Verify per source, not at the end
Findymail and Prospeo verify at find time. FullEnrich bundles verification free. Apollo does not. Bolting one verification step onto the end of a pipeline means paying twice for most of your rows. Route verification per source and it halves that line.
Cap the budget before anything runs unattended
Every overspend we have been called in to look at came from a loop rather than from volume: a retry that did not back off, a scheduled job that fired twice, an agent that worked the same list four times. Set a hard credit ceiling per workspace on day one.
Write back, or none of it counts
The step teams skip. An enrichment pipeline whose output lives in a spreadsheet gets used for about a quarter and then quietly abandoned. HubSpot gives you full CRM API access on its free tier, so there is no cost argument for skipping it.
These are the models. Frontal builds the pipelines.
Provider order, regional routing, per-source verification, credit ceilings that stop a retry loop emptying a balance overnight, and the write-back that puts a verified contact in front of a rep the same day it was found. That is the whole engagement, run by a senior operator.
Worked with 300+ companies. A senior operator runs your account.
Questions we get
What is a stack here?
A real motion with the providers named, the order fixed and the cost worked out per unit of output. There are 4: waterfall email enrichment at $0.0364, hiring signal to outbound at $0.13, inbound speed to lead at <60s, agent-driven account research at $0.09.
Are the costs real?
The rates are, read from each vendor's own pricing page and carrying the date they were read. The hit rates behind a waterfall are not: no provider in this category publishes a match rate you can hold them to, so those percentages are modeled assumptions and every page that uses one says so and shows the figure.
Can I build these without an engineer?
Three of the four, with a workbench like Clay plus a sequencer and a CRM. The arithmetic is the same either way, which is the point of showing it: the cost per unit does not depend on who wires it up.
Do the providers have to be the ones named?
No. Each step says what job it does, so a provider doing the same job at a different rate substitutes cleanly. The per-unit figures let you redo the sum yourself rather than take ours.