Tell HN: Mailgun lowers free-tier API from 10k to 625 emails per month

422 points by kehphin ↗ HN
Email:

Hi there,

Mailgun is adjusting our plans and pricing to more accurately reflect the value users get from the service and to make room for some great new deliverability features we just released.

Throughout 2019, we were hard at work adding and improving our email capabilities and optimizing our support to help your business grow. While many of these updates were made behind the scenes, the truth is that Mailgun can do a lot more than it could two years ago when we last updated our plans.

What does this mean for you? On March 1, 2020, we will automatically transition your account to the new Flex plan, a pay-as-you-go plan comparable to the Concept plan you’re currently on. You’ll receive your first invoice under the new plan on April 1 if your amount due is greater than $0.50. According to your usage last month, your invoice under the new price per message of $0.0008 would have been $0 for December. It’s a modest change, but we wanted to be transparent about it.

What’s changing with the Flex plan? Flex offers you the same pay-per-use model you were used to on the Concept plan. The main differences are that we are no longer offering 10,000 free emails or 100 free validations per month, and our support options now include limited ticket support as well as enhanced self-service Q&As so you can find answers faster. Additionally, while your existing routes will still be functional, new routes will not be supported on this plan.

What other options do I have? We have several other plans available with additional features and service levels, including a new subscription plan called Foundation that starts at $35 per month. This plan provides access to new deliverability tools like Inbox Placement so you can effortlessly increase your deliverability and email ROI.

Looking for validations, inbound routing, or more support? Foundation is a great starter plan. If this is something you’re interested in, check out your plan options.

271 comments

[ 3.1 ms ] story [ 285 ms ] thread
Completely fine with going to a PAYG model for sent emails, it's a shame that inbound mail will no longer be supported at all on that plan though
I think 10000 was probably too generous, but 625 also seems a bit too little as well for hobbyists.
I was trying to reason how they hit that number this morning. I'm guessing probably average or mean is much lower.
My account is still grandfathered in with 30k free per month.
This takes place in March.
Hobbyist to me means that you are doing it for fun and aren't engaging more than 5-10 people. So given that, I'm having a hard time understanding how 10 people could handle 625 emails a month.
I communicate with family and very few close friends only (less than 10 people) and my Mailgun stats show about 1k emails per month.
That’s 3 emails a day for 10 people... what have you built that requires that level of communication with family and close friends?
As someone who has a noncommercial side project and has been using SendGrid's free 100/day plan, I'm actually thinking of switching to this. I've always been a bit concerned with what would happen if my app got a sudden burst of signups (since email verification is required to sign up), and have been looking for an uncapped plan, but don't want to spend $10 a month or more just for that possibility.

The key here is that with the $0.80/1000 emails PAYG rate, you can still do like 2000-3000 a month for less than $5, which no one except SES will provide. Everyone else has a gulf between "free tier 100/day" (or even 100/month with Postmark) and their first actual paid tier ($10/month for Postmark, $15/month for SendGrid). No one else except SES has this PAYG for low rates as far as I'm aware. I think if your hobby project needs more than 625 emails a month, it's reasonable to spend a dollar or two on a month to handle that.

I run a gitlab for around 30 users and that’s around what I use on AWS to send my emails. It’s a bunch. Rarely do I send emails on weekends so it’s like 20 days a month I send emails just for merge requests and stuff.
Not surprised. Pretty much all transactional email services have done this and I have have used mandrill, sendgrid, sparkpost so far. Perhaps it is not cost effective to provide such high tiers or not worth it due to spammers.
I worked at one of these companies a few years ago in the department that dealt with spammers. The volume of attempted spam from these free (and low tier) plans was absolutely staggering.
Anything that allows sending emails with custom content will attract spammers. We aggressively limited the free accounts, and that still attracted spammers who would just create loads of news accounts to send out emails; we eventually put a stop to it by limiting it to 25 recipients in total, after which they finally gave up (before, it was 25 emails with a max of 10 recipients each, which meant 250 spam emails).
I am trying to figure out where 625 comes from? Nowhere in the email or on the pricing page that I can find does Mailgun say how many free outbound emails you get now.
The OP calculated $0.50 / $0.0008 per message to get 625 messages, based on "You’ll receive your first invoice under the new plan on April 1 if your amount due is greater than $0.50. According to your usage last month, your invoice under the new price per message of $0.0008"

That's not quite how it works out though. Mailgun rolls over the balance till the balance hits $0.50 and when the balance hits $0.50, they cut an invoice for that amount.

Oh wow, so it really is 0 free emails...
We used to be customers of Mailgun, but last month moved over to SES. Not that there's anything wrong with Mailgun, SES's pricing is far cheaper (60K free emails if you're sending from an AWS instance), and deliverability is on par.

But SES is super basic. It's just an API, no fancy analytics.

So of course, I built a small web app that "wraps around" SES, and gives you some insight into what you're sending, some pretty graphs, and other stuff.

If anyone is feeling adventurous, it's https://messageray.com

You give it an AWS access key with permissions for SES, and you can send 60k free emails a month. It's just a side project atm, and completely free.

Needs to be self hosted or otherwise it's just another mailgun.
Mailgun is (was) great for setting up a quick regex to process incoming mail. Maybe I'm misusing it, but I point all my non-email domains there and let this regex forward everything to my main domain:

^(abuse@|hostmaster@|postmaster@|webmaster@).*

I'd gladly pay $.0008 per message x2 to cover both sides of that event because I have very little usage. By putting email receiving into a tier that costs over $400 per year they've priced me out of the service.

It's disappointing the way these companies are happy to have enthusiasts, hobbyists, and side projects light-housing their products for them at the start, but eventually switch to paid tiers that alienate those same types of early adopters. It feels like it happens with every single service I use.

Boo!

I've posted this elsewhere in this discussion but have you tried out Sendgrid Inbound Parse? Does it meet your requirements?

https://sendgrid.com/docs/for-developers/parsing-email/setti...

Technically, yes. Practically, no. Mailgun has a checkbox to forward matched messages to another address, so the appeal was having something extremely easy, fast to configure, and reliable.

Building some type of custom endpoint for it is too much work for what it does. I already pay for email where I can set up unlimited domain aliases, so I'll just wait for my Mailgun email and switch back to aliases.

The incoming mail filtering was the only reason I decided to try Mailgun instead of Twilio or SES. I already have SES set up for transactional mail and my usage is so low they don't even bill me for it.

Ditto, I only just got the email about this today; so now I have <1 month to figure out something else.

The email doesn't even mention that 'email receiving' is switched off as a result of moving me to this plan! Pretty poor, IMO.

Really nicely executed! Love the clean and minimal design. However many of us would be unable use this in a corporate setting due to AWS credentials being shared with a third party service. I am sure if you packaged this service like Sendy (sendy.co) you could find a lot more corporate buy-in.
The problem with SES is that they ignore abuse reports. They even admit that they only block you if you have more than a given complaint rate. That is how gmail works, but for a bulk email service they really need to investigate every complaint.

I've just blocked SES's ip ranges on my servers.

> investigate every complaint

I'm not sure this is valid. Some users will consider anything spam, including password reset emails they asked for.

That's why you need to investigate, rather than simply blocking the user. If set up correctly it should be mostly automated...ask the customer where the email address came from, then take appropriate further steps. SES does none of this, which is why they have this problem.
Unsurprising, given that that has to be a huge spam magnet.
Historically they've been very careful. No DNS, you don't get to send to non-verified test emails. High bounce rate? No more mailing for you until you contact them and explain.

All in all they seem reasonably switched on and protective of their IP addresses.

Is there any alternative to Mailgun for sending / receive mails for free? Let's say 500mails per month? Thank you.
gmail? outlook?

basically anything with imap access.

This is downvoted, but correct. You can programmatically tap into a Gmail account and send 20-30 emails per day, which adds up to the 500/mo.
Whilst technically correct the issue with gmail is it overwrites the from address with the gmail address.
Gmail + a mail merge addon (like Mailmeteor.com or Gmass)
Just curious if it is 500/month, why switch?
From the description, it seems, that you have just 500 mails to send. You can not receive emails anymore. Or am I wrong? Thank you.
Mailjet, even though it is owned by Mailgun, has a higher free tier.

Sendgrid is owned by Twillio, and has a similar free tier.

Ses is pretty cheap, for 500 mails, it is essentially free.

Or get a dedicated IP, use mailinabox or other similar one click setups.

I Own our private VPS where I was sending and receiving without trouble in the past but currently I am not able to setup mail server correctly what is reason why I started to use mailgun.
Disclaimer: I work with Twilio SendGrid.

We have the inbound parse webhook which turns incoming SMTP into an HTTP post to your URL. Our free plan is is 100/day after the trial period which has a much higher limit.

Most large Cloud Providers have an arrangement with at least one email service which allows some free service (at least no extra charge than whatever one is paying for the subscription), with a far higher limit than the typical free tiers - e.g. AWS/SES, Azure/SendGrid, IBM/SendGrid, GCP/Mailgun.

So if you're already paying for some cloud hosting somewhere you likely have a free email service you can use.

There was an earlier discussion with 'Tell HN' title talking about 625/month. https://news.ycombinator.com/item?id=22192543 When I access the page I read "Get 5,000 free emails per month". Maybe they're A/B testing?
5,000 a month free for 3 months. Previously, you got 10,000 emails and 100 validations per month included. You had to have a non-prepaid credit card to sign up as well.
Is this going to affect my current account? I've been using mailgun for years under this 10k limit to send mail from all my domains for free except once when a bunch of alerts sent thousands of emails.
Yes, I got an email about it this morning. You'll move to the PAYG plan automatically in a month.
I'm a happy customer, and will be moving to foundation plan soon.

If I didn't rely on their inbound email parsing, I could get by on the flex plan for less than $10/mo.... it's still super cheap.

We have used Mailgun to power our SaaS emails for years with great deliverability and without paying a cent. It makes sense for them to charge for their service. It was too good to last forever!
batch and switch from 10,000 to 625 is worthy of never using them ever again.
Not much of a bait and switch when the change is coming after several years.
So what you're saying is that if you're offering a free plan as a business, you can never change it?
(comment deleted)
Doesn't sound like you were a customer in the first place.
Is there any other service instead of Mailgun who can also receive/resend incoming mails? I didn't find anything nice. thank you.
I think most of the big names in this space do both, but I can personally confirm that Postmark.app does outgoing and incoming.
I run https://improvmx.com, a service to forward incoming emails to a personal inbox (like @gmail's). Is this what you are looking for?
So under the guise of spam protection, we have created an oligopoly of email gateways, despite sending an email being an utterly trivial, long solved problem, and now those companies are putting the squeeze on.

The best part about this particular item on the list of "things that have only gotten worse over time": they charge companies more to deliver what is essentially spam, and they conspired to make that spam show up in inboxes and "enhance" it with all sorts of tracking.

> despite sending an email being an utterly trivial, long solved problem.

It is absolutely not. Even if you are not considering the whole spam thing.

Sending an email with the multitude of clients, MIME which is very complex in itself, network issues, message queuing, retries, unsubscription, bounces, providers feedback loops, rate limiting and what not is very hard.

If you add the whole spam folder + blacklists situation into consideration, sending emails becomes a more than painful thing you don't want to deal with.

Sending an email is absolutely a solved problem.

I can send a message to my address in 2020 using a 2003 Fedora Core 1 box without a domain using only what's part of the POSIX standard, and I'll receive it just fine.

I just did a few days ago, actually. It even got past my provider's spam filter.

I could probably have done it using earlier software than 2003-era mail/mailx, but that was the quickest way I could find to send an attachment given the software on hand.

You can still do that.

However, I, as a receiver of email, get thousands of spam messages a day. So, I use services that block them before they even reach my view, or get sent directly to the spam folder.

If you want to send me an email, that's solved. If you want to send it to ten thousand people on your mailing list, and not have it end up in the spam folder, you may have more work.

Please explain the difference to me. Mailgun is sending tons and tons of spam email; what is so special about them that they aren't ending up in people's spam/junk folders?
They probably send much, much more valid emails, respond quickly to abuse reports, and since they're large/known enough, they won't get blacklisted because that would be disruptive to large, legit business.
Mail providers like Mailgun live or die based on their reputation. If they as a sender get marked as spam by "The Algorithms", then their product is useless. They almost certainly respond to abuse reports quickly and have enough reputation, and enough legitimate clients, that they have "leeway". But that leeway isn't unlimited.

There was a big fit thrown on this website a year or two ago when one transactional mail provider (MailChimp?) said they were banning crypto companies with ICOs from their service. They literally put "ICOs are not welcome" in their Terms of Service. Why? Because first off, ICOs are scams. And that means they buy email lists, and then spam the shit out of people with "buy into our ICO!" emails, which get immediately shitcanned into the spam folder by 99.99% of users -- they're no different than Nigerian Princes, as far as most people care. That behavior tanks the reputation of the sender, and you cannot reason with The Algorithm once it "recognizes" you as spam. That kind of problem is not a minor inconvenience to companies like Mailgun, it's an existential threat.

They really care with their reputation on the line. Here for example is their response to the only complaint I've had to send Mailgun: "Thanks for sending this over. I was able to tie this sender to a compromised account, which has been disabled in our system to prevent further messages."

Pity we'll have to move off them. It was good while it lasted.

This raises what's really the ultimate question, which is whether and how much individuals should be made to suffer so that advertisers can send spam. Certainly the recipients are not responsible for the advertisers' decisions to spam, or more charitably, to blur the lines. Yet it's individuals who pay, in this case literally, for the advertisers' privilege of spamming them. It's extremely fucked.
I think this is what happens in open systems. The alternative is walled gardens, like FB messages. Even those suffer from spam, although it's much less of the wild west.
The alternative is simple:

For stranger to stranger communication, both business and non-business set up a web form. With a challenge if needed.

For friend to friend communication, use email, friends' address in on a white list.

For business to consumer communication, white list is again used.

For spammer to anyone communication, including business lists that refuse to take you off, not on white list so doesn't get in. Bounce response with reason "non on white list"

It's a cultural change though: an email address won't get through unless you are unblocked, but technically easy.

What's the easy/low-friction way for Joe Random Consumer to manage their whitelists?

An email comes in from a sender address which isn't on the whitelist, but is (in fact) a friend. What's the handling?

An email comes in from a sender address which isn't on the whitelist, but is not (in fact) a friend. What's the handling?

Do I never hear from the first friend? Or do I get bombarded with "hey, read this email and see if the sender is a friend" a bunch of times?

I'm just shooting from the hip here, but I think it's:

Joe manages his whitelist through his mail provider's web UI. Many providers already use the address book as a whitelist; the only difference is they default to filtration instead of assuming spam.

Joe's friend's mail gets routed to the spam box. Joe's friend gets a bounce notification that says "To be added to Joe's whitelist, click here (and optionally solve a captcha/enter Joe's dog's name/submit a blood sample/deposit $0.25 worth of bitcoin into Joe's wallet)." Joe's friend clicks here, and Joe's mail provider adds Joe's friend to the whitelist and promotes the email to the inbox.

Joe's spammer gets the same message and disregards it. Their email stays in spam forever.

Joe's spam box gets bombarded with "hey, read this email and see if the sender is a friend" a bunch of times. He can trawl through them for actual friends if he wishes.

A system with more inherent friction -- whether cognitive or social expectations -- will be out-optimized by one with less.
Actually I think individuals suffer so that less SPAM gets through. The SPAM is flowing regardless.
That cuts both ways.

You might miss something valuable. Perhaps from a long-lost relative. Or a prospective client or employer.

And once you're actually looking through the spam folder, what good is it really?

You're both right, but you're solving different problems. Just because your message got past your spam filter doesn't mean it would get past all spam filters.
It is absolutely not. Even if you are not considering the whole spam thing.

They specified that it wasn't even if you could ignore the spam issue. I said that it was if you ignored the spam issue. The first post in this thread was "We've made sending electronic mail way too complicated under the guise of fighting spam."

I'm aware that it won't get past every spam filter, but it gets past both Gmail's and my own provider (a smaller, non-American one), which is good enough for my own use. My claim was that sending electronic mail is a solved problem, because it is.

> My claim was that sending electronic mail is a solved problem, because it is.

Agreed. I mean, I can _probably_ remember enough to be able to send mail using just telnet without even needing to look anything up. (And that mail, without any mime parts or urls - it very very likely to not fall into any of the spam filtering on my inbound email accounts...)

In the modern era, lots of mailservers run software that looks for weirdnesses in incoming email. I don't know, but I'd suspect that the typing delays of using telnet to send mail are likely to trigger their defenses.
SO I just tested. After working out that on macOS I now need to use nc -c instead of telnet... I _can_ send mail manually, but it _does_ end up in gmails's spam folder saying "Why is this message in spam? It is similar to messages that were identified as spam in the past."

  Iain-2:~ iain$ nc -c gmail-smtp-in.l.google.com 25
  220 mx.google.com ESMTP k5si5682618pls.209 - gsmtp
  EHLO elided.org
  250-mx.google.com at your service, [202.171.181.100]
  250-SIZE 157286400
  250-8BITMIME
  250-STARTTLS
  250-ENHANCEDSTATUSCODES
  250-PIPELINING
  250-CHUNKING
  250 SMTPUTF8
  MAIL FROM: <bigiain@elided.org>
  250 2.1.0 OK k5si5682618pls.209 - gsmtp
  RCPT TO: <bigiain@gmail.com>
  250 2.1.5 OK k5si5682618pls.209 - gsmtp
  DATA
  354  Go ahead k5si5682618pls.209 - gsmtp
  Subject: Test
  From: bigiain@elided.org
  To: bigiain@gmail.com
  Date: Fri, 31 Jan 2020 13:07:00 +1100
  
  Foo baz nah
  
  .
  
  250 2.0.0 OK  1580436443 k5si5682618pls.209 - gsmtp
  QUIT
  221 2.0.0 closing connection k5si5682618pls.209 - gsmtp
  Iain-2:~ iain$
I wonder why google treats this simple email as spam. It contains nothing that should trigger the spam filter aside from where it came from.
Most mailservers don't send emails one character at a time. You appear to have a mental model of how a spam filter works that doesn't include fingerprinting the behavior of the sending server? Even open source mail servers do that, for example postscreen is a part of postfix.
I wonder if it'd fare better if I scripted nc with expect?

(Not sure I care enough to try...)

Some of the things it doesn't contain, though -- like Message-ID and Content-Type headers -- are likely to contribute to its spam score.

If your domain has SPF headers which don't permit mail from the (residential?) IP you're connecting from, that'll be a major factor as well.

Where it came from is one of the most important details!

Is this mail relay actually elided.org or is it impersonating elided.org?

There does not appear to be an SPF record for that domain so Gmail cannot determine if this is a fraudulent message, especially if the machine this request came from does not match the IP of the A record for that domain.

Some years ago I wrote my email in a forum, available to crawlers. I’ve been getting spam ever since.
Similarly, talking is a solved problem. Talking, so that people will listen, is not.

Communication is much more than just broadcasting, just as email is about more than sending emails.

Most ISPs block SMTP traffic, which creates a pretty significant barrier to sending email. Even VPS services send through a secret gateways with all sorts of filters and controls on it.
Sending at scale is the problem. A provider like Mailgun (aka an ESP), is a huge target for abuse. Free and low priced tiers are tricky since they need to draw in customers, but then get heavily abused.

tl;dr: Abuse ruins it for everyone.

I work on the filtering side of the world so I see a lot of the challenges that even purportedly good actors face. Each provider does better or worse at controlling abuse. These things all have a cost:

* IPs -- You need to send from a variety of IPs. They have warmup time before they can be safely used. One bad guy can burn that IP and everyone who shares. Most ESPs aren't going to have dedicated IPs for each (or those are for high tiers).

* Dmarc/dkim/spf -- Passing validation can be hard since it requires your client to work with you.

* Abuse reports -- Likely huge volume

* Bounces -- They can count against the customer. How much is OK (e.g. AWS SES will cut you off if your bounce rate gets too high)?

* Throwaway account abuse -- Huge issue, this threshold is a simple hammer (so now bad guys need ~10x more throwaway accounts).

* Account takeover -- That good guy is now sending out crap. Now what?

* Post-abuse cleanup -- Good luck going around and working with major providers to get yourself unblocked. Its a huge time sink.

By using Mailgun, etc all of the above becomes their problem. On top of it they'll offer analytics, help crafting content that works in varying mail clients, etc.

A small business type sender doesn't need much of this (other than maybe the technical help). But at some point, the scales tip.

^ Couldn't agree more, especially "Abuse ruins it for everyone."
> IPs -- You need to send from a variety of IPs. They have warmup time before they can be safely used. One bad guy can burn that IP and everyone who shares. Most ESPs aren't going to have dedicated IPs for each (or those are for high tiers).

I'm curious, is that specific to IPv4 or you see the same phenomenon with IPv6 where providers allocate /64 or /48 to customers? Or there simply isn't enough sending volume from IPv6 at the moment to warrant specific filtering?

It’s mostly IPv4. The volume over v6 isn’t there.

Blocking via IP on V6 is tricky since it’s not clear what a safe/useful block range would be.

I've sent a lot of mail in the last 14 years, and it's not hard.

I spammed with invite mails to users friends back in 2008 and mail started to end up in spam for a while, but after switching to double opt-in for all mail and making unsubscribing easily available it went back to good. Add some periodic list cleaning of long time inactive subscribers and it works perfectly!

It does get harder if you have to account for bad users, but if you're the only user then you can make sure everything is good.

So, is it hard? That depends entirely on how wanted or unwanted the sent mail is.

It is clearly not trivial though. I get that it feels like it should be and probably is if you're sending a few emails, but "at scale" (pardon the buzzword) it is not trivial hence why people use these services.
I've been running a SaaS in the email space since 2011. It is not surprising at all, to me, that Mailgun decided to do this.

Sending an email _is_ trivial - you're right - but getting it delivered to the inbox, instead of the spam folder, is far from trivial. At scale, it is incredibly difficult. Even worse, spammers and fraudsters are utterly relentless. They hammer services like mine and Mailgun's constantly with fraud and phishing and porn and other garbage, inbound and outbound.

Yes, sending email is indeed a solved issue, but that isn't what mailgun (re?) solves.

Asking someone's permissions first, tracking their desires and immediately updating your response is what they solve.

Seems trivial? Try doing it for (looks at note) 626 people.

I think their point was that it's been drummed into our heads that ISPs and mail services only accept email from 'trusted' IP blocks, so you can't do it yourself, better use 'the experts'.
>despite sending an email being an utterly trivial, long solved problem

AFAIK, SPAM is a still a hard problem, and getting emails delivered isn't trivial.

"Under the guise of Spam" is a very reductionist viewpoint. It's like saying "Under the guise of the measles, we now have an oligopoly of pharmaceuticals." The latter might be true, but its important not to understate the threat of the former.

No, we've created something far better. ~50% of the Spam that makes it to my inbox is sent via Amazon SES. Amazon does not give a fuck, I've reported it to abuse, they ignore it.

I can't just block Amazon SES though, because there's also a ton of legit services that use them for mail transport. We've really helped spammers evade blocking and decrease their infrastructure costs, and because there's so much "normal" traffic on SES, spammers will blend right in. And as long as they pay for services rendered, Amazon couldn't be happier.

These providers are just as responsible for spam by running opt-out mailing lists and, even worse... spamming you with surveys and confirmations to see why you unsubscribed. With tracking pixels and rewritten URLs that go through their system before you get to the real destination.

You don’t get a choice. That was made for you.

If it's an oligopoly and presumably throwing off cash, what's stopping anyone from creating their own transactional email service?

Answer: It's very difficult. Email is adversarial. Getting to the inbox is incredibly difficult to do _at scale_.

A private service where you are very sure that subscribers wants and opens the mail? Easy.

A public service where a lot will be spam and unsolicited emails? Must be hell to run.

> despite sending an email being an utterly trivial, long solved problem

Then why do companies need transactional email providers?

We have our own domain in G-Suite for our mail. I just checked my GMail Spam box. In it are 6 emails from this week. All 6 of them were sent through Mailgun, from and to the same domain (the one in G Suite). One was a cron email from one of our servers and the other 5 are transactional emails about people signing up to our service.

All of them passed both SPF and DKIM.

How. Just how did these emails get spam filtered by Google? Everything about them screams that they're legit emails from ourselves to ourselves.

Wow, sounds like even if you pay for a company to do then still might not go to inbox 100%... I noticed marking things not as spam still sends things to spam, I figured be doing that would tell it it's not spam...

I was recently thinking about the email problem myself, and was reading about IP Warmups. Seems like every provider has different recommendations too. So sounds like not an exact science either. But maybe if it was an exact science it'd be hard to tell the good and bad guys apart. I guess one of the tricky problems with decentralized systems.

You want to try an impressive open source software project? Create a framework that takes an email forwarded to a given address, turns it into an HTTP request with a JSON body, and sends that HTTP request to webhook.

It’s easy to make a proof of concept. Production grade is an entirely different story.

Remember to preserve conversations and forwarding. Make sure file attachments are handled and scanned for viruses. Make sure you watch out for recursive automatic forwarding. Make sure that any html is secure. And make sure it will scale well ... and I mean billions of emails a day well.

You are 100% right. It started the same with our service; we bought back ImprovMX (I don't put the link as I'm not looking for publicity - just for reference) when it was just a 5$ droplet at Digital Ocean, and now we have 10 big servers, more than 700$ monthly cost (and I honestly think it's not that much) and have to handle all the messy parts of delivering an email.

It's easy to send an email, but it's way harder to ensure that the emails you send are delivered properly.

Maybe a better idea for them would be to say charge you $10 upfront for 12.5k email credits which should last you a while. Seems like a better deal for "hobbyists" vs a bill at the end of a couple of months. For MG, they can collect payment upfront. I guess maybe for some people, going from free to $10 might be too much of a leap? We were on the "free" Mandrill plan which became PAID (which we were fine paying) but you have to pay for Mailchimp before you get the privilege to pay for Mandrill (we don't use Mailchimp at all), all our stuff is transactional. Hard to get out of it as we are using both inbound and outbound
This is what twilio does right? you buy credits ahead of time and then consume them via service usage?
Correct. Purchasing phone numbers is recurring ($1/month for residential and toll free numbers).
Postmark used to take this approach and it was awesome. Now they've switched to a subscription model for new accounts :(
I submitted this under the title "Mailgun eliminates 'Concept' plan, no longer offering 10000 free emails/mo(https://www.mailgun.com/pricing)" but it was a dupe. I think mine might be a bit more descriptive title, though. They are not offering 625 for free. There is a 3 month trial that gives you 5,000 emails.

Does anybody know the $0.0008/email is the real pricing, or if you have to pay $0.80/1000 up front like the footer of the price table says?

If you're staying on the Flex plan, it's true pay as you go, no need to pay up front.
I've been happy with Mailgun for free, super low volume sites (i.e. single digit emails per month.) However, I've been shopping around for alternatives because I have a domain that won't pass their dashboard validation of DNS records. I checked DNS from a wide variety of other tools, and everything works. I waited weeks but the dashboard still reports problems, but it doesn't tell me _what_ the problem is, and their support, naturally, isn't the best at the free tier. So I couldn't get any help from them on what the problem was. I'm open to the idea that I misconfigured DNS, despite it matching other domains, and the DNS tools validating my work, but without someone on their end giving me debugging information, I haven't been able to solve my issue, and have to use an alternative.
Whoa. Thank you so much for telling me this. I am literally in the middle of a one-day project to send nearly 10k emails and have been using the free tier for years with no issues. I was just thinking today about how good the free tier is. I am not surprised they are making this change. I will gladly upgrade, Mailgun provide a great service and I have never had an issue.
I really hope this means every dang blog/recipe page/random article will stop asking for my email so they can spam me.
10k emails per month is generous. Maybe too generous. Selfishly I wish they didn't increase the price, but as a aspiring SaaS founder, it's totally fair that they did. They do provide value.

But 625 emails per month amounts to around 20 emails per day. If you have 5 customers, then you can send them 4 emails per day. So not much.

I wish they found a middle ground here.

Who cares though? If you have customers the cost to send an email is still almost zero.
Sure the marginal cost of sending another email might be zero. But typically you price based on what people are willing to pay -- not based on costs.
I may have confused you. I mean who cares if you have to pay for mail gun, it’s basically free.
I thought you were referring to Mailgun.

I thought you were saying that since Mailgun has customers, and has built their email infrastructure already, the cost of providing a free service to people who aren't using the tool much is zero.

Thanks for clarifying.

Yeah I see how it sounded that way. Right the reverse is true, they are charging so little that if you have a customer it’s still almost free.
What service sends 4 e-mails to your inbox everyday?
Heroku, Sentry, Jira, GitHub, any service that's notifying me of events via email regardless of company size.
I have some free v1agra for you to buy from n1g3rian princ3.
It's seems to be working though, everyone knows about the Nigerian Prince. He's really good at marketing.
Interesting thing about that. One might ask why any scammer still uses Nigerian prince tactics since it is so we'll know and so obviously a scam.

The reason is that they are selecting for the most clueless marks. If you don't get a bunch of red flags from an email like that, then you are likely going to be an easy mark.

> But 625 emails per month amounts to around 20 emails per day. If you have 5 customers, then you can send them 4 emails per day. So not much.

That's if you're sending email every day of the week.

If it's just Mon-Fr then it's 30 emails per day.

If you have 5 customers ... then you can hopefully afford USD$0.80 for 1000 emails / month.

As a hypothetical customer please don’t send me 4 emails a day. Or week for that matter.
Mailgun is transactional mail though. Not marketing mail. Mail that you want to receive like password resets, email confirmations, notifications, etc
(comment deleted)
I don't know about you, but I usually don't want to receive notification emails. Did you explicitly ask for it? Good. Is this some stupid notification that I won't care for and was enabled by default? Congratulations on getting flagged.
> Did you explicitly ask for it? Good

That's what 'transactional' means... it's supposed to be used for the end or beginning of a transaction (reset password, receipt for a purchase, things like that...)

No, that is not what transactional mail means, you're wrong.

Things like friend requests are also transactional email by definition. So are those "you left your cart behind, want to complete the purchase?". That doesn't mean that you asked to receive it.

Instead of being pedantic about it lets give it another name then. Whatever we call this new category of email that does not include password resets and does include friend requests and abandonment emails - don't send them 4x per day.

Come on, sometimes it's better to be understood than to be right. Fuck the definition.

Would you like to be notified your account has been breached? Perhaps a notification that a login was attempted from another country.

There are plenty of good notifications besides social media updates.

I send between 500 and 1000 per month through mailgun. On high-usage months, I will pay around $1.

Even if I sent 5000 messages one month, I'd pay around $4.

This is not even close to a significant cost.

(comment deleted)
Any chance they are A/B testing this?

I'm a MG user and haven't received that email and their page is showing me 5K/month free tier.

Are you looking at the part that says, "Get 5,000 free emails per month for 3 months?" I would interpret that as an introductory quota for new accounts; after three months the monthly quota will be reduced.
I use Mailgun with various client sites. Where I find these sorts of services awkward as they become paid (Google Maps APIs are similar) is that for each client I have to put my credit card on file, monitor charges and bear costs in case of a blow-up OR I have to go through the hassle of coaching technically hapless clients through logging in and adding a credit card.

“I thought you said it would be free?” “It probably will be, but we still have to add your card.” “What do you mean ‘probably’?”

The problem is having clients who expect free...or rather delivering free to clients who expect free. Your job isn't to save the client money on your services. If having your credit card on file provides value to the client, then charge the client. If it doesn't, charge them anyway because it has value to you.
We started with Mailgun when the free tier seemed established and permanent. No card required. Our CMS taps into their API. Then they required a unique phone number for each account and even that was a time sink getting clients to comprehend what was going on.

I think some people underestimate how cost-sensitive some clients are. And how time poor some small businesses are.

> I think some people underestimate how cost-sensitive some clients are. And how time poor some small businesses are.

Those aren't "clients", they're "moochers".

If they're prepared to walk away because it costs them $35/month to send to their mailing list, you probably owe it to yourself/your business to spend your time seeking new better clients rather than talking them thru how to get stuff/service for less than $35/month... Your business should be "cost sensitive" as well...

Exactly. My business lifted off when I realized my market isn't the people who benefit from what I'm selling. My market is people who benefit from what I'm selling, and are willing to pay for value.

Freeloaders are simply not in my market.

And with that focus, I've been able to exponentially improve the value of what my company provides.

They are clients if they found us to provide a service they needed, and we can provide it in a profitable way.

Our business found Mailgun to be a great solution when we anticipated clients would stick under 10k/mo and not need a card on file. With Stripe, every time they pay their cut, they've had a sale. Every time they mail-out through MailChimp, it's to a list of customers.

We are usually using Mailgun to send transactional emails. I can't promise a client that they won't get a flurry of junk signups or password reset mailouts that hit their credit card.

The bottom 50% of the market is heading to self-serve site-builders and it's savage for a small web business.

Mailgun does not owe you or anyone a free tier.

Start charging your clients money to cover costs that comes with their business.

They literally offered a free tier!

My original point was not even anything to do with cost but hassle!

In most businesses "hassle" and "cost" are synonyms...
I think the ask is that we wished Mailgun did that from the start instead of creating messes later in the process.
Never assume free tiers will last forever. Recognize they are often a way to entice people into the system. And that often they change as the business grows. Thus don’t build your business that depends on free tiers.
Unless the vendor has more to lose than gain if they started charging.

Discount stocks brokers make more money on float than trades.

Search engines make enough off ads instead of charging per search.

Photoshop didn’t have a free tier, but rampant piracy from home users created proficient corporate users.

I don't mind them having a free tier and tweaking their offering. But I think grandfathering in older accounts is nicer, and I wish it was easier to work with accounts in a technical role when you are not technically the client, but are acting for them in most ways.

e.g., I can set up their account and other technical parts after that, but I can't verify their mobile number (added since we started working with Mailgun), don't have their credit card details to put on file (same), and often end up agreeing to legal notices for them rather than asking them to complete signup (and stalling the project).

Mailgun also wished that you'd have signed up as a full paying customer when you first wanted their service. But they knew they weren't gonna get what they wished for either...
(comment deleted)
(comment deleted)
I think it's possible you're both right. There will be cases where you need to drop a client because they're not willing to pay you enough for keeping them to be profitable (parent's point). On the other hand, there's a whole "low-end" market available to some people who have the resources to band together a large number of very-low-profit customers into a product that ultimately still makes you money (your point). If you can find such a market, then providing a service to these people is economically justifiable.
I'm completely with you on this.

The responses to your posts are disgusting. I don't think a lot of people around here understand what it means to provide services to small local businesses.

I don't know about you, but I'm in a smaller community, people here still earn $7.25 per hour, and I just don't have access to deep-pocketed big businesses. I work with small businesses, individual business owner/operators, and mom and pop type places. $35 here, $50 there - it all adds up fast, and they are very conscious of these costs, and so am I.

Going to a small business owner and trying to explain any kind of price hike for a service they thought was free is just another burden to them, and it makes MY recommendation of the original service seem bad .... and I'm the guy they came to for GOOD advice.

> it makes MY recommendation of the original service seem bad

If your advice was "you get 10,000 emails per month free via $service" without qualifying "right now, but that isn't guaranteed forever, their paid tier is $x per 1000 emails and switching mail providers in the backend of your website will take approximately $Y hours at $chargable_rate", then it _was_ "bad advice" (or at the very least "incomplete advice").

I made this mistake way to many times before I learnt that lesson. (Most recently with Google Maps on websites...)

(comment deleted)
FWIW, as someone who built their career from Ogaki, Gifu, Japan, you're on the same Internet as most deep-pocketed businesses and can choose to pursue their custom at any time. Strongly consider doing so. Most technologists overrate how difficult it is.

I think most HNers currently serving mom-and-pops in a bespoke fashion should strongly consider exiting that market tomorrow. Those customers will, over time, gravitate to a Shopify or a site builder or similar because they can amortize engineering costs over 100,000 similarly situated customers and consultants can not.

I understand there are aesthetic reasons to prefer that non-tech-forward people in your local community have someone to ask questions to and help navigate options, but if you want you can throw free Set Up Your Shopify office hours every Friday as a pro bono gesture, underwritten by the piles and piles of money from the many businesses in the world that can afford professional labor.

(comment deleted)
A lot of my clients fit this. What I’ve done to “solve” it for us, is to stop charging for hosting and “services” each month and realistically price it in to the initial contract. A lot of our projects are between £5k and £20k. I don’t want to be raising invoices for £20 or £30 per month. Their corporate finance department doesn't want to be processing small amounts each month either. All these projects have limited lifespan - 1 year or 2 or most. It makes sense for all involved to just price it in early. They like it, I like it. Sometimes projects run longer, sometimes they run short, on aggregate, it seems to work for us and them. Yes, if something goes nuts and costs thousands a month, that’s a conversation you want to have quickly, but in that scenario, the traffic they are getting is leading a different conversation anyway. YMMV.
You could align incentives this way by structuring the deal to include a three year "service and support" component. You can build that in to provide a fixed cost yet at the same time have the flexibility to charge more if your cost structure has changed after the time passes.
I don't underestimate how bad clients can be. I don't over-estimate how hard it is for some self-employed people to ask for more money. It's ok to be fired by people who aren't paying you. It's even better to fire clients who aren't.
These clients ARE paying me. I like them. They just have an overloaded plate of issues and it’s no fun having to tax them further when a third-party setup changes. Initial point wasn’t even about cost but hassle.
(comment deleted)
I don't underestimate how small businesses operate. The issue here is that the relationship has B2C connotations. To me, that's unhealthy. B2C relationships are about bargains and lowest price. B2B relationships are sustained by each party wanting the other party to stay in business.

Your problem cascades from a B2C approaches to a B2B relationships. Not just yours to your client, but Mailgun's B2C approach to its relationship with your business. Free commercial email is likely to be unsustainable. At best, it's money left on the table because at 10k emails a month, there won't be many conversions. They were delivering free to a customer who expected free with the same damage to the customer's perception of the relationship as between you and your client.

The only part under your total control was whether or not to charge your client money. If you had, then the bad part would be Mailgun raising its prices. You would have had the option of eating the cost. Or having the unpleasant conversation with your client.

Maybe there's something slimy about charging a client for a free tier mail service. That's also under your control. Just pay for a paid tier and pass along cost at normal markup. In the long run, and that's what B2B relationships are about, it will tend to be simpler and simpler tends to be better for everyone.

(comment deleted)
Perhaps the better answer is to never pitch or put your clients on any "free" plan. Explain that these sorts of marketing and outreach campaigns cost money, just like getting a Yellow Pages ad costs money, and perhaps work with them on figuring out how many customers or leads these efforts generate so they can more accurately evaluate the cost.
I have a lot of clients using MailChimp's free plan which has worked well for them. I am grandfathered into using all Google's services for my business for free (ignore usual criticisms of Google) because I got in early.

I've worked in this industry for over 20 years and found that the more complicated your explanation, the less likely clients are to want to work with you. Being a stick in the mud on usability is another example - they'll leave for someone who'll make them a splash page without question, etc.

I certainly understand that. And grandfathered or not your use of Google's services for 'free' for your business is coming to an end.

As long as you understand the risk you are carrying you can price out the externalities of it going south.

To explain that, consider this artificial example made up based on what you just wrote; "I have a lot of clients using MailChimp's free plan ..."

Now lets say MailChimp has a tough few quarters, maybe they get a new CEO or maybe they just need to get more money for their services than they currently do and so they revise their "free" plans into non-free plans in ways that prevent some chunk of your customers from using their free plans.

Your customers are forced to change their previously working system for a different system. Four outcomes pop out of that event; They can use their current system but now it costs them more money, they can hire you (more money) to come up with another free system which they now know will only last for some random period of time, you can design a new system for them for free (costs you money), or they can go with a new contractor who will design a new system for them.

As the original provider of the system, you cannot predict when this event will occur. It may occur at a slow time, it may occur on black friday when it the customer's busiest time of year. And as the original provider you cannot predict the new cost of the system, or the cost of a replacement system.

So the risk you are carrying here is mostly reputational (some of your customers will say bad things about you if this event happens at a bad time or costs them a lot of money they were not expecting to pay.) Depending on what type of warranty you provide or imply for your services (most people disclaim these so the common case would be to leave the customer with the costs) you may be carrying some financial risk to correct the future situation.

I will reiterate that there is absolutely nothing wrong with carrying this risk, as long as you are aware you are carrying it. Because you can plan for the various cases (for example by continually researching equivalents to MailChimp so that in the event this happens you can quickly and easily move your customers over to the new solution)

> And grandfathered or not your use of Google's services for 'free' for your business is coming to an end.

When are they coming to an end? I'm also grandfathered in, and haven't heard about this.

Do you as a consultant have a free tier?
Yes, I have worked free for charity, free for friends, and free/speculatively for or with people on projects hoping that some pay off. Some do.

I have a SaaS product with a free tier (like Mailgun) and I have had an app/game which was free with an upgrade.

(My initial point was not about the freemium model but about other changes that have made working with Mailgun difficult.)

Oh. I haven't done consulting in a couple of years, but in the decade I did, I always insisted clients sign up for anything important on their own. Not just for this reason, maybe really not for that reason at all, I sold it as, "If we decide a year from now we hate each other, I want you to be in complete control of your site/ tools."
Have your clients created a virtual card number using Privacy.com and allocate $X of spend to it per month. They can then give you that card number to setup their services.
We're not in the US. They would also question this as something they haven't had to do for any other service. Plus, painful to repeat for dozens of small clients. I appreciate the idea though.
Hi! So, for the past year I’ve been working on a self-hosted service to allow people to switch between different email providers at runtime. Basically, how can we make it easy for people to configure how, or when, to use a given SMTP provider, or do failover, when changes like this happen.

We don’t have a landing page or anything- just working code! Would anyone here like to see it working or be willing to share feedback?