Case Study · MailGenius

Find out where youremail lands beforeyou press send.

MailGenius is the free email spam test of EmailMarketing.com, Troy Ericson's group, and the deliverability platform built around it. We built and run the product: the scoring engine that grades a message out of 100, the guidance attached to every finding, and the monitoring, extension and accounts behind the paid tiers.

Emails tested per year: 1M+Emails tested per year
Checks on every test: 12+Checks on every test
Domain monitoring: 24/7Domain monitoring

Engagement Overview

MailGenius is the free email spam test that most of the deliverability world has run at least once. Send a campaign to the address it gives you, press See Your Score, and it grades the message out of 100 and tells you what is costing you the points. Its users test over a million emails a year, and none of them pay to do it. It belongs to Troy Ericson’s group, alongside EmailDeliverability.com and EmailMarketing.com.

We are the engineering team behind it. Nortik is the group’s core technology and AI team, eight engineers across three products, and this is the third of them: the scoring engine that reads headers, content and sender reputation, the guidance attached to every finding, and the paid platform built around the free test, with domain monitoring, automatic testing, a Chrome extension, a deliverability chatbot, accounts and billing.

One structural fact shapes the rest of this page. The front of MailGenius is free and anonymous, with no account and no setup, and the business is the subscription behind it. Both are one product, and the free half is the half that has to survive a million tests a year without anybody paying for a single one of them.

1M+
Emails tested per year: 12+
Checks on every test: 24/7
Domain monitoring

Engagement: Team as a Service
Team: 2 AI Engineers, 1 Delivery Manager
Focus: Email deliverability testing platform
Working model: Owned delivery

Stack

Product & frontend

React
TypeScript
JavaScript
Chrome Extension APIs
HTML / CSS

Services

Node.js
SMTP

Data & storage

PostgreSQL
Redis

Email & DNS

SPF / DKIM / DMARC
BIMI
Reverse DNS
Blacklist feeds
HTML analysis

Commerce & AI

Stripe
Dante AI

Someone holding a phone above an open laptop, reading messages on the smaller screen

The Challenge

Email fails quietly. Nothing bounces, nobody complains, the campaign just underperforms, and the sender finds out weeks later from a flat open rate that tells them nothing about why. The answer is buried in authentication records, reverse DNS, blacklists, domain age, link hygiene and the HTML itself, which is specialist work. The people sending the mail are marketers, and asking them to read a DMARC record is asking the wrong question.

So the product had to be two things at once. A free test any stranger can run in under a minute with no account, no integration and no setup, because that is the front door and it carries over a million emails a year that nobody pays for. And a subscription platform behind it, with history, monitoring, alerts, entitlements and billing, because that is the business. One codebase, two audiences, and the free half is the one under load.

The Solution

It starts with a throwaway address. Copy it, send the campaign you were about to send anyway to it, press See Your Score: no account, no integration, no code, and no change to how the mail was built, which is the only reason a marketer runs it at all. It is also why the free tier is an engineering problem rather than a pricing one, because every one of that million tests a year is an anonymous inbound message that has to be accepted, parsed, scored and thrown away. What happens in between is the product.

A MailGenius score report reading 87 of 100 beside a severity table of failing, warning and passing tests, the sender address redacted

A score out of 100, and the shortlist behind it

We built the scoring engine: it reads the message headers, the content and the sender's reputation, runs the authentication, blacklist, domain, link and HTML checks against them, and returns one number with the arithmetic still visible. The report leads with the things to fix rather than the things that passed, each one carrying the points it cost, so a sender who reads three lines has already read the useful part.

Proprietary Scoring
Headers & Content
Sender Reputation
Ranked Findings

A MailGenius finding expanded, explaining why DMARC matters and how to move the record off a policy of none

Every finding comes with the reason and the fix

A score on its own changes nothing, so no finding is left as a verdict. Each row opens into why the check matters, what specifically failed, and the change to make: a DMARC policy sitting at none is explained as the hijacking risk it is and answered with p=reject or p=quarantine, in that order, with the warning that setting it up out of sequence blocks your own mail. That translation is the product. The checks are commodity, the guidance is not.

SPF / DKIM / DMARC
BIMI
Blacklists
Plain-English Fixes

A MailGenius DNS monitoring result for a domain, listing DKIM and DMARC as passing with the reasoning under each

And it keeps watching between tests

A pre-send test only covers the send you remembered to test. Domain monitoring closes that gap: it reads the DNS around the clock and mails the account the moment something breaks, with a weekly digest when nothing does. Test history keeps every result beside it, so a domain's deliverability has a record rather than a memory, and a fix can be shown to have worked.

Domain Monitoring
24/7 Alerts
DNS Watch
Test History

The MailGenius account screen showing an active Automatic Email Testing and Domain Monitoring plan, the domains monitored and the static audit address, personal details redacted

And there is a business behind the free tier

The account screen is where the free tool becomes a product. A plan, its renewal, the card behind it, the domains being monitored, and a static audit address the customer drops into their own bulk sends so campaigns are scored without anybody running a test at all. All of it is self-serve, which is what a tool with a million anonymous users a year needs: nobody at the client wants to be in the path between somebody trying MailGenius and somebody paying for it.

Plans & Renewals
Stripe Billing
Automatic Email Testing
Self-Serve

And what is around it

The same read, inside the tab
A Chrome extension puts the audit where the email already is, so a sender can check how likely a message is to be delivered without leaving what they were doing to go and test it.
A chatbot with exactly one subject
An AI model tailored to email deliverability answers the questions the report raises, in the report. It is scoped on purpose: a narrow assistant that knows SPF alignment is worth more here than a general one that will guess at it.
Free checkers as the wide end of the funnel
SPF, DKIM, DMARC, BIMI and blacklist checkers each stand on their own, which is what puts MailGenius in front of people who came looking for one answer rather than a platform. An API and a whitelabel tier sit behind the same engine.
Rebuilt on Node.js and React
The platform has moved off its original Ruby on Rails codebase onto Node.js with a React front end. One language across the services and the app, and a stack the rest of the group's engineering already runs.

A Nortik engineer photographed against a plain wall in the Novi Sad office

Business Impact

MailGenius is the widest thing the group owns. A million tests a year come through a page that asks for nothing, and every one of them ends with a specific, ranked, fixable answer from a tool that has just proved it knows what it is talking about. The monitoring, the extension, the checkers, the consulting practice and the sister products all sit behind that moment. A free tool only works as a front door if it is genuinely good, and building one that is has been the work.

And it does not stand on its own. The same engineering team builds EmailDeliverability.com, the software that fixes an email inside the marketer’s own editor, and InboxPro, the infrastructure their mail actually leaves on. Three products, one team, and a group that can answer a deliverability question from the test that finds it all the way down to the machine that sends the message.