A product that won by doing one thing perfectly spent seven years becoming several. Now its parent publishes growth figures with it removed, and the platforms where new apps get built don’t list it at all.
Prompted by a read of Doug Dennison’s “The Fall of Mailchimp.” Dennison founded MailNinja in 2005 and spent years ranked the top Mailchimp partner in the world out of 900+. Worth reading alongside this.
On its Q3 FY26 call on May 20, 2026, Intuit published its own growth rate twice: once with Mailchimp included, once with Mailchimp stripped out.
- Intuit Global Business Solutions grew 15% to $3.3B. Ex-Mailchimp, 17%.
- Online Ecosystem grew 19% to $2.5B. Ex-Mailchimp, 22%.
- On the same call, management confirmed Mailchimp revenue declined year over year, and announced a 17% workforce reduction, roughly 3,100 roles, that named “rightsizing in Mailchimp” directly. Restructuring charges of about $300M land mostly in the quarter ending July 31, 2026.
or context on what was bought: Intuit paid approximately $12B in 2021, partly funded with a $4.7B term loan, for a business doing $800M in revenue and growing about 20% a year. It got there. In Q4 FY22, the one quarter Intuit broke the number out, Mailchimp did $265M, or roughly $1.06B annualized. Intuit has not disclosed a Mailchimp revenue figure since.
Five years later the parent company is publishing its growth with that asset removed and cutting investment in it.
The decline started before AI, and the timing is what makes it fatal. Mailchimp had already spent years widening a product that won by being narrow. What the agent era added was a second problem on top of the first: the entry point for new email moved from a marketer signing up for a UI to a developer or a model provisioning it with an API call. Mailchimp is losing the customers it has, and it is not in the room where the next ones are being created.
This isn’t a story about a bad company or bad people, although it might be in part a story about what happens when founders move on, and a product is run by a large corporate behemoth, just one product of many. Mailchimp hasn’t logged out. It shipped a real ecommerce release in February 2026 and a native conversational analytics agent in May 2026, with integrations across Shopify, WooCommerce, Wix, Canva, ChatGPT and Claude. The people building inside it are doing serious work. What happened is structural, and a lot of B2B companies are one roadmap decision away from the same place.
The Trail Was Public for Four Straight Quarters #
Intuit’s own disclosures mapped the decline for over a year:
Segment definitions shifted between periods, so read the gap rather than the absolute levels. Half a point at Intuit’s September 2025 Investor Day. Four points by December. Three points and an outright revenue decline by May.
The guidance language decayed on the same curve. August 2025: CFO Sandeep Aujla called Mailchimp a near-term drag with initiatives underway, expected to be performing well by year end. February 2026: double-digit growth pushed out to sometime beyond fiscal 2026. May 2026: growth language gone from the conversation, replaced by profitability.
Eleven Million Users, Zero Growth #
EmailToolTester’s January 2026 update of self-reported ESP customer counts shows how the field moved while Mailchimp held still:
Self-reported counts mix free users with paying customers, so treat the absolute numbers as soft. The direction holds regardless. Everyone with a functioning growth engine added double-digit or better percentages over eighteen months. Mailchimp added none.
The 11 million number is also doing less work than it looks like. Mailchimp’s headline metric counts businesses on the platform, most of them free. When your biggest number is a denominator you don’t monetize, it stops telling you anything about the health of the business underneath.
What Best-in-Class Looks Like in the Same Market, Right Now #
Klaviyo, Q1 2026: revenue of $358M, up 28%. Over 196,000 customers. Customers generating more than $50,000 of ARR up 38% year over year to 4,175. NRR of 110%, up two points. Non-GAAP operating margin of 16%, the highest since IPO, and the first quarter of positive GAAP operating margin. Revenue per employee above $600,000, up more than 25%. Revenue outside the Americas up 39%, EMEA excluding the UK up 51%.
110% NRR in a customer base that skews small is the number that matters most there. Small businesses are supposed to be structural churners. Klaviyo is expanding inside them anyway, because it does one job (B2C customer data and messaging for commerce brands) deeply enough that customers grow their spend rather than shop the category.
Mailchimp sells to that same segment. Its own CFO told Reuters those small businesses, the bread and butter of the platform, found it “a bit harder to use,” and said that hurt retention and expansion.
Same market, same buyer, opposite roadmap philosophies, and a 28-point spread in growth rate.
Focus Was the Moat. The Suite Strategy Spent It. #
Mailchimp won 2001 through roughly 2019 by being the best way for a small business to send email. Then it became an all-in-one marketing platform: websites, landing pages, social ads, postcards, CRM features, appointment tools, SMS. The acquisition thesis stacked another layer on top, joining QuickBooks purchase data to Mailchimp reach to produce lifecycle marketing tied to revenue.
That thesis is coherent in a deck. It has never once been how an SMB buys software. A 12-person business does not buy a data-synergy story. It buys “send my customers an email that gets opened, in under twenty minutes, without learning anything new.” Every feature added to serve the platform narrative made that specific job slightly harder to do. The cost surfaced years later in retention and expansion, which is both where product complexity always lands and where it is hardest to trace back to a roadmap decision made in 2019.
The competitive set went the other direction and specialized. Klaviyo took commerce, Kit took creators, Brevo took price and volume-based pricing, beehiiv took newsletters. Each of them beat Mailchimp at a slice of what Mailchimp was trying to hold all at once.
The turn also started before Intuit owned it. Mailchimp added the bulk of its non-email functionality in 2019, two years ahead of the deal, and Doug Dennison, who built the top-ranked Mailchimp agency practice, watched his large-organization work dry up as Mailchimp’s marketing focus changed and the businesses on the platform got steadily smaller. So Intuit paid $12B for what looked like a compounding SMB platform and got a category-defining brand already drifting down-market with a product widening faster than it was deepening. It then spent five years pushing in both directions at once, chasing a mid-market that grows north of 30% inside Intuit with a customer base moving the other way.
Then the Incentives Flipped from Growth to Harvest. Freemium is Being Killed #
Mailchimp’s free plan was the growth engine. It popularized freemium in email and built the 11 million user base. Its history since the acquisition:
- Pre-2022: 2,000 contacts, 10,000 sends per month
- 2023: 500 contacts, 1,000 sends per month
- Effective February 17, 2026: 250 contacts, 500 sends per month, daily send cap cut from 500 to 250
Multi-step automation came off the free tier by mid-2025. Contacts are counted including unsubscribed and non-subscribed records, so lists hit the ceiling faster than owners expect. Exceed the contact limit and sending can entirely, test sends included. In April 2026, legacy accounts created before May 2019 that had never migrated took an 11% to 13% price increase.
Compressing the free tier raises this year’s ARPU by taxing next year’s paid cohort. It is a rational move if you have decided the cohort isn’t coming. Intuit has said close to that outright: the position at Q3 FY26 was that Mailchimp’s revised cash flow profile would generate more value for Intuit than a third party would likely pay in the current equity and debt environment for software, with Aujla adding that third-party terms simply are not there right now. Goodarzi’s memo language was “reducing investments in areas including Mailchimp.”
They explored a sale, didn’t get a number they liked, and are now running the asset for cash.
A Cut Too Far? Has Deliverability Taken a Hit? #
Intuit hasn’t published a function-by-function breakdown. Affected staff have. Within hours of the announcement, Mailchimp employees posted publicly, naming email and SMS delivery operations, software and infrastructure engineering, and engineering management among affected functions. Evan Burke, a known voice in the email industry, wrote that about half of his team was let go. Staff engineers with 12 years at the company posted similar notices.
Deliverability engineering is not overhead on an email platform. It is the function that determines whether mail arrives, how fast an incident gets resolved, and how quickly problems with Gmail or Microsoft get worked. Cutting it degrades the product’s only real promise.
For anyone running meaningful volume through Mailchimp, that is the line item to watch over the next two quarters, ahead of the roadmap.
Agents Aren’t Choosing Mailchimp #
There is a second buyer now, and it doesn’t read your homepage.
Start with where apps get born. When someone builds on Replit, Lovable or Vercel and needs email, here is what the platform hands them:
Replit. Its integration docs list two email options, Resend and SendGrid. Built-in test sending inside Replit agents runs on Resend, and builders graduate to the Resend connector to add a domain and send to real users.Lovable. Resend ships as an app and chat connector, a workspace-level connection that works in the chat while you build and in the published app.Vercel. Resend joined the Marketplace on July 1, 2026, provisionable with one CLI command. Installing it also installs skills that tell agents how to work with Resend.
Mailchimp is in none of those lists. Not as an option, not as a connector, not as a skil l. Some of that is category, since Mailchimp sells marketing rather than transactional infrastructure. But Mailchimp owns Mandrill, it sells to exactly the small businesses these platforms are minting, and the app being scaffolded today is the marketing list of next year.
The deeper test is whether an agent can operate the platform without a human clicking through a dashboard. Mailchimp’s only official MCP server covers Transactional Messaging, the old Mandrill product. The marketing platform, meaning audiences, campaigns, automations and reports, has no official MCP server. Everything on the registries labeled “Mailchimp MCP” for marketing is community-built against the v3 API with a key you paste in. The credentials make it worse: the official Transactional MCP uses a static account-level bearer key with no per-user identity, scopes or revocation, and Marketing OAuth tokens never expire and have no refresh token. Nobody hands an autonomous agent a permanent, unscoped, account-wide key to a subscriber list.
The detail that says the most: one community server’s docs note its automation coverage is for classic automations only, because automation flows are not available in the Mailchimp API. The newer version of Mailchimp’s own core feature cannot be reached programmatically at all. Resend, by contrast, ships an official MCP server, a Skills layer in its CLI for agents, and 50+ commands across the full API.
Two things get conflated here. Mailchimp has shipped real AI, including a conversational analytics agent in May 2026 and integrations with ChatGPT and Claude. All of that brings AI into Mailchimp. None of it makes Mailchimp drivable from inside someone else’s agent stack. Being AI-powered and being agent-operable are different products, and only the second one protects you when the thing choosing your category is a model.
Eight Things B2B Founders Should Take From This #
1. Focus is a moat with an ongoing maintenance cost. Mailchimp didn’t lose because it stopped working. It lost because it kept adding, and every addition was a small tax on the one job it was hired to do.
2. Ease of use is a retention metric, not a marketing claim. When your own CFO diagnoses “harder to use” as the cause of churn and weak expansion, that is a product decision from five years earlier arriving on the P&L.
3. Your free tier is your pipeline. Cutting it is borrowing from next year. A 96% reduction in free-plan capacity since 2022 books revenue now and unbooks the paid cohort that would have graduated in 2028.
4. Big user counts hide bad businesses. 11 million users with zero growth and declining revenue is a worse position than 196,000 customers at 110% NRR. Report the denominator you actually monetize.
5. Watch for “ex-[product] growth” in a parent’s earnings release. When a public acquirer volunteers its growth rate with one asset removed, the asset is on notice. Intuit did it for four consecutive quarters before the cuts.
6. Integration synergy is not a buyer benefit until the buyer feels it. Combining purchase data with marketing reach was a sound idea that took years and produced a harder product. The synergy has to show up as a job done faster, or it’s just architecture.
7. Switching costs are thinning while acquisition costs are shifting. Mailchimp built migration tooling to pull customers off Klaviyo. Brevo runs free concierge migration off Mailchimp. AI is turning list, template and automation migration into a task rather than a project. At the same time, new send volume is increasingly provisioned by developers and agents through an API rather than by a marketer signing up for a UI. Incumbency used to survive mediocrity for a decade. In 2026, plan on far less.
8. Agent-operability is the new evaluation nobody told you about. Publish an official MCP server that covers your real product, not a side product. Scope and expire your credentials so an agent can be trusted with them. Expose your newest features through the API, not just the UI. If an agent cannot complete the job through your platform, it will complete it around your platform, and you will read about it later as churn.
The Seven-Year Gap Between the Decision and the Disclosure #
A lot of category leaders from the 2010s are sitting somewhere on this curve: a product that won by being best at one thing, an installed base large enough to make price increases look like strategy, and an owner whose incentive is now cash rather than compounding. The financial engineering works for a while. It doesn’t produce a product anyone new chooses, and it doesn’t get you into the default list on the platforms where the next million apps are being built.
Mailchimp was the answer to “what do you use for email” for two decades. What’s worth taking from it is how long the lag runs between the decision to become several products and the quarter the numbers say it out loud. About seven years here. Nobody gets a warning in the year they make the choice.
Sources: Doug Dennison, “The Fall of Mailchimp,” and his Email Love interview; Intuit Q3 FY26 earnings materials and call (May 20, 2026); Intuit Q4 FY25 and Q1-Q2 FY26 disclosures; Reuters (Aug 2025); emailexpert (May 2026); MarTech (May 2026); EMARKETER (May 2026); EmailToolTester ESP market data (Jan 2026) and Mailchimp review (Jun 2026); Klaviyo Q1 2026 results (May 5, 2026); Mailchimp Developer documentation on the Transactional Messaging MCP; Scalekit and community MCP repository documentation on Mailchimp Marketing API coverage; Adobe Experience League documentation on the Marketo Engage MCP server (Apr-Jun 2026); Mailchimp pricing and plan documentation; Resend company blog: 1,000,000 users (Dec 18, 2025) and Launch Week 6 wrap-up (Apr 20, 2026); npm download comparison via Zeno Rocha; Replit integration docs; Lovable connector docs; Vercel changelog and Marketplace listing (Jul 1, 2026); Twilio SIGNAL 2026 announcements (May 6, 2026) and Q1 2026 earnings call (Apr 30, 2026).