Hacker News new | past | comments | ask | show | jobs | submit login
Mailgun: new pricing without plans and limits (mailgun.com)
184 points by old-gregg on Aug 16, 2013 | hide | past | favorite | 64 comments



This is awesome. Aside from the fact that my bill will be cut in half, I hate plans. Plans are just one more thing to monitor (make sure you don't go over your totally arbitrary limit!), and they generally lead to weird inequity.

For example, I've been adding ~75 people per day to my MailChimp list and the price stayed the same. Then one day it went from $75 to $150. Adding thousands of subscribers was completely free, and then adding one subscriber cost $75/month. That just doesn't make any sense to me.

Related anecdote: My company doesn't have plans (most in the industry do) and customers tell me all the time how much they prefer our approach. There may be selection bias going on, but I have tons of feedback supporting my decision to avoid plans.


This is probably a nature of how the MailChimp plans are constructed. I know that the Zencoder plans get cheaper when you level up to the next one (and it’s worth while switching before you hit the minimum included minutes.)

Plans can be done right - that is in a customer friendly fashion while still maintaining reoccurring revenue for the service.


Isn't your company Less Annoying CRM?

I see a monthly plan on your pricing page, did I misunderstand what you meant by "My company doesn't have plans"?

https://www.lessannoyingcrm.com/pricing.php


Chrischen is right. Maybe I'm not using the term correctly, but I think of "plans" as meaning multiple different tiers that a user can choose from. Mailgun previously had plans (i.e. you had to pick which tier you wanted) and now they don't (i.e. there is just one tier that applies to everyone).

Some characteristics of plans that I don't like:

* Passing some arbitrary limit (# of contacts, # of users, # of files, etc) triggers a steep increase in price, rather than having the price gradually increase as usage increase. For Mailgun, usage is the number of emails sent. For us, it's the number of users on the account.

* Because of the steep price increases at arbitrary points, some users get a lot more value out of the software than others. With Highrise (another CRM) an account with 5,000 contacts pays $24, while an account with 5,001 users pays $49/month. Both accounts are getting roughly the same value, but one is paying twice as much as the other.

* Most SaaS companies with plans force users to choose which plan (or tier) they want when they sign up. How would a user know what features are worth paying for when they haven't even tried the software yet?

* Plans are often used to upsell existing customers after they're locked into your service. Salesforce has a plan that is cheaper than my company's price ($5/user/month vs. $10/user/month) but it is limited to the point of being almost useless. I've talked to many people who got suckered into a 1+ year commitment with them only to realize they'd have to pay $65-$250/user/month to get the functionality they needed.

* Customers have to monitor their usage to make sure they don't trigger the next tier. It's just one extra thing to worry about every day, and businesses have enough on their minds without worrying about unexpected price increases.

Having said all of that, I'm currently struggling with how I might go about offering more products than just a CRM without making it look like a tiered system. I don't consider it to be the same thing because they really will be additional products (project management, invoicing, etc.) which are generally completely separate subscriptions if you use a different company's products, but I want them to be integrated/bundled directly with the CRM to make the user experience as simple as possible. Still not quite sure how to pull this off. Pricing sure is tricky.


I took a quick look at it because I will be in the market for a crm system. A question I have is whether it's possible to add contacts to a campaign so it's easy to see the results per campaign.

In terms of offering more products without making it look like a tiered system. If it's invoicing, it's perfectly fine to add that as a feature we can pay for. Lots of SMBs already have their own invoicing system. They know it's something separate. Just give it a name, a different color and sell it as a feature. I think I would like a to-do list as well. Not all contacts buy the same thing but some do, I want to automate it a little bit, so it's a step by step approach that I can go through. Give me a list that I can easily copy to other projects and easily change. Not sure if Basecamp offers something similar. Add it to the crm system as a feature and charge me for it. Sounds fair to me.


Thanks for checking us out! There are two different options if you want to track campaigns. One would be just make a "Group" for each campaign, and then add the appropriate contacts to the group. Very soon we'll be adding the ability to filter the pipeline reports by groups, so that should give you what you're looking for.

Alternatively, you could add a custom field to a pipeline (pipelines = leads for most people) which would track the campaign. It could be a text field or a dropdown list depending on how many campaigns you're running.

As for the task management, it's not quite the same, but you can add a checkbox list to the lead pipeline to track which tasks you've completed. We also have more task functionality coming at some point, but not in the near future unfortunately.

Thanks for the feedback on the pricing. Our reservation with charging more for each feature is that while it's very fair, it seems kind of complicated to some people if they have to pick and choose every single feature they want to pay for. Many of our users love how easy it is to understand the price. What we were considering was offering two products: CRM, and "everything else". So you decide whether you want all the extra stuff (invoice, project management, etc.) or if you just want the CRM, and there are just two prices (probably $10 and $25). I like that model, but it seems a lot like having pricing plans which is why I'm not 100% sold on it yet.


I think he means plans with different tiers instead of one continuous tier.


This seems to be closer to how Mandrill works, and it's exactly what got me to sign up for them only a few days ago. I knew about Mailgun and Sendgrid before, and started looking into them first but their pricing tiers turned me off and got me to keep searching until I found Mandrill.

My app is very small, just bootstrapping without any funding so I'm trying to be very careful how I spend my money. At first I see that Mailgun's minimum was $19/month while Sendgrid's silver plan is $10/month. Easy, Sendgrid wins.

Oops, I want to support inbound email though. Sendgrid will give you that on the trial plan, but not on the $10 silver plan. Suddenly I'm up to $80/month to use Sendgrid. Mailgun wins here.

But before I even noticed that I had already signed up for Sendgrid (believing I'd be paying $10/month), but had to wait like 5 hours and then I received some email from them basically demanding that I tell them what my app is, why I need email, what is the purpose of me sending email. Maybe it's for market research, but it didn't come across that way. It came across as "we assume you're a dirty spammer". Of course if I was a spammer I'd just give them so BS answer, I wouldn't admit it. I can't say call a winner, only a lower: Sendgrid. Mandrill and SES both treat you like you expect; you sign up, they cap you at ~250 emails/day for a bit and then you're fine. Sendgrid fail.

Anyway, then my friend points me over to Mandrill and I sign up. The pricing model fit us well.. up to 12k emails free, then $0.20/thousand after that. In the state my app is right now I probably won't go over 12k for another few months, and once I do go over that limit I'm not going to find myself jumping suddenly to something like $80/month.

If I had been setting this all up just a week later (that is, now) I'd probably be on Mailgun due to this new pricing model.


The same thing happened to me with SendGrid. I gave them the essay they asked for, just to see whether they'd ask for anything even more ridiculous. Thankfully they didn't, and they unlocked my account so quickly I suspect they didn't even read my intentionally extremely verbose reply. My best guess is that the essay request serves as a CAPTCHA.


Inbox providers keep track of reputation of an IP address, so most outbound e-mail service providers guard their IPs, since even a few bad apples can cause havoc for existing customers. Not saying it's justified, but there's some verification that needs to happen sooner or later.


Right, but that's the point of SES and Mandrill limiting you to like 250/day at first and gradually increasing your limit.

Asking you to write an essay telling them what you're using email for is not verification.


If you do a lot of inbound, another thing to consider is difference in pricing practices:

SendGrid and Mandrill inbound is currently complimentary to customers.

To receive an inbound email via Mailgun, you pay once for it to be received by mailgun servers, and once for it to be delivered to your webhook.


>you pay once for it to be received by mailgun servers, and once for it to be delivered to your webhook.

Mailgunner here. This does not sound right. Can you please reach out to me via address in profile? Mailgun does not charge for webhooks, you're probably confsuing this with email relaying (which we also do).

Also it is impossible to have complimentary inbound email of production quality due to the need for quality spam filtering. Handling spam in production setting (throughput, quality of service, uptime, performance) for free would mean wiping out CloudMark and other spam filtering vendors.

In other words, as always, you get what you pay for. Mailgun has always been known for being a true bi-directional programmable ESP and a sizable percentage of our traffic is inbound.


Ev, you and Taylor were on the thread I had last week with Travis about this pricing practice. Will reach out.


I can't figure out the sendgrid pricing at all. 10/mo for 40k emails and 80/mo for 100k emails doesn't fit a normal model of reducing price for incremental usage and there's no clear explanation for it.

I use sendgrid, and I thought I was using mailgun, but now after seeing the pricing simplicity and ease I'm feeling like I should make the switch.

What reason would I rationally have at the <100k emails level for staying with sendgrid?

Additionally, for the lite plan with sendgrid, they don't include email open and click tracking (even though they DO for the FREE tiers).

Would love an explanation for why this is sensible.


Given the higher price, what are the advantages of mailgun over Mandrill worth paying extra for?


Great job on this. What I really like about this type of pricing is that it puts the complex cost calculation onto your side instead of the customer. I use a ton of products, Heroku services being an example where I have to try to guess at usage. Literally every day I have an internal conversation with myself that says, "Do you need to go into Heroku and turn down the Database tier you're on because the service isn't using the capacity that you're paying for? Wait, but you might get a traffic surge and the site will go down if you decrease the capacity."

This seems like for me, the ideal case. Automatically scaling down all the way to $0 so I never have to think about unsubscribing if I stop using it.

I was actually discussing with a friend that it seems like there could be a big opportunity in the cloud server space to have a zero scale server where you only pay for time that it is being used. I believe Heroku already does spin down instances that aren't needed behind the scenes, they just don't give you the cost savings. They keep it.


To be fair, they only spin down people on the free tier, where you can't save over $0.


Is this documented somewhere? Are you sure? How can you tell that you actually have 3 dynos running when you're paying for 3?

I suspect that if you pay for at least 2 dynos what happens is that they keep at least one dyno running and scale up to 3 as needed because as a customer I can't really tell the difference between 1 and 3 dynos unless my traffic requires all 3 at which point the service would have already scaled up to three. We already know from the free tier that they can spin up an instance in a couple of seconds.

I would bet this is Heroku's little secret about making money. If you look at their cost per month you find that they aren't really making much per instance on AWS servers although, they are perfectly capable of detecting when they should auto scale up. Amazon even has a built in service to do that, but Heroku doesn't expose it because that's where they make their profit margin.


Dyno idling is documented, and it only happens to web dynos, when there is only one running. https://devcenter.heroku.com/articles/dynos#dyno-sleeping

I can say from experience they definitely do not scale down the dynos you are paying for - you still see log entries from them and they are immediately available to requests.


Oh interesting. I was wrong.


Our most popular feedback to date is to eliminate minimums and get rid of feature gating in plans. So, here you go.

As always, open to feedback.


I'm not sure if this is feasible or not, but I'd love to read an overview of how you are handling the different plans scenario (previous plans setup vs new pay as you go setup) on a technical level. It'd be helpful for any SaaS owners who want to one day do something similar and as far as I know, no one has ever documented this process.


Before if you had a 59 US $ plan (dedicated IP), the cost was 0.59/thousand, so you could send 118k emails.

With the new plan, for 59 US$ you get the dedicated IP and 10k emails. Sending 118k emails with a dedicated IP would cost 113 US$


For some customers it may make sense to stay on their current pricing, which we allow for.


Love it! I ran into the (free) limits recently during a one-off email blast as part of a new feature rollout. The limit caused a lot of emails to fail.... but MailGun's customer service rectified it in like 5 minutes (at like 2am on a Saturday). Very impressive!

Any idea when the Heroku plans will reflect the pricing changes...?

EDIT: From the article, old plans may be granfathered (if desired)...? If so, how do we switch to new pro-rated on Heroku? Just downgrade the addon?


Unfortunately, Heroku doesn't allow for variable billing so there needs to be a fixed cost and associated limit. Although, we are in the process of reevaluating our Heroku pricing to better reflect our direct pricing and will be adjusting it shortly.


Thanks for doing this. It makes your pricing now fair and easy for an entrepreneur to plan. I'm currently comparing Mailgun to Mandrill and I'm wondering what makes Bakunin better? Mandrill's prices look great. Can you enlighten me on differences I may be missing?

One feature I'm looking for that mandrill doesn't seem to have is multi-tenancy.


Just a quick note: your pricing calculator does not let me enter any number. I use Czech keyboard layout where numbers are typed with Shift key. This happens in both Firefox and Chrome. Once I switch to English layout, it works fine.


I think you should update your FAQ entry: "What are the differences between the plans you offer?"

Otherwise, looks good :)


Thanks for this. Yes, we need to update our other assets to reflect the new pricing.


While reading the new pricing structure, I realized that the Mailboxes feature [1] (allowing programmatic creation of mailboxes via the API) is deprecated. That's too bad; it was the reason I started using Mailgun in the first place. Is anyone aware of an alternate service that offers this feature, short of managing my own mail server?

1: http://documentation.mailgun.com/quickstart.html#using-mailb...


The old pricing was fine with me. Now I worry you will go out of business or get swamped with support requests from people who haven't paid anything. But whatever works for you, I guess.


Mailgun was purchased by rackspace. I don't see them going out of business any time soon.


I was also more than happy to pay $19/month. Now my monthly bill goes to $0? I'm a bit worried what will happen to quality of service and support now.

And the fact Rackspace has purchased Mailgun doesn't mean a thing. Just look what Rackspace did to JungleDisk... I used to be also happy customer of JungleDisk so no wonder I'm a bit bitter regarding Rackspace.


You can continue to stay on the $19/mo plan. There's an option to stay on your existing plan or move to the new pricing.

I'm not familiar with what rackspace did to jungle disk as I've never used it. I was under the impression the service was still running with a rebrand.


I always thought there was a price point to weed out spammers. I hope this doesn't effect the service.


Been using mailgun for a while to send nagios alerts from our server using SMTP to various accounts that need to be notified. Instead of sending it to 3 accounts, you send it to one then mailgun sends it off to the accounts you want added.

It is perfect and worth the $19. Now I'm unsure if I should just downgrade... but it's not really worth it for the value I get out of Mailgun.


I started looking at http://yclist.com/ and realized that many companies I often see listed on the front page here are actually YC companies, but not marked as such.


Mailgun is now owned by Rackspace, so I don't think it's still considered an active member of the YC family.


Ok, how do you explain stripe and priceonomics? I'm sure I'm missing others.


Why do they need to mark themselves as YC companies if they don't want the boost associated with being YC?


When SES is 5 time as less (although less features, I know), there's really no reason why I would go with Mailgun. 10 cents for every 1000 emails is very very hard to beat. And when I'm just relying on you to email, I don't need things such as UI, analytics, etc as much.

I typed in 300000 in the pricing calculator, and got $150.00/month as my price. For any startup, that is just too much. With SES, it's $30.00/month. I know that there's a premium to be paid for certain features, and I know the pricing is better than what it was originally, but SES to me is still superior.


You don't get dedicated IP and custom DKIM with SES. Those affect deliverability significantly.


I don't under the pricing table.

It shows:

Emails Price / 1,000 Total Price

0 - 10,000 Free

Next 500,000 $0.50

Next 1,000,000 $0.35

Next 5,000,000 $0.15

Any Additional $0.10

Does that mean emails 10,001-510,000 cost $0.50 per thousand and then emails 510,000-1,510,000 cost $0.35 per thousand?


Yes, that seems to be what it means. In general, heavy users get better marginal prices, but there's a "free sample" amount for very light users.


Dear Mailgun --

How can I programmatically identify the accounts that are on their first email blast? Please include a header. I suggest it look like this:

X-Mailgun-FNG: 1

with the value set to the number of days on which this customer has used your service to send more than 100 messages.

Then I can filter out all the spammers who are signing up for your service with stolen credit cards, sending until you stop them, and then doing it all over again with a new account.

What, you didn't think of that?


Weirdly, the pricing calculator doesn't appear on the linked page - but it does on the pricing page. (using FF 20, ubuntu)

I think some people might like plan limits, to cap their costs (so they don't accidentally run up a huge bill). It might be nice to have a separate option to also impose a limit. Though this is an extra complication and I'm not sure how many customers would actually want this, if any.


Aside from the cringeworthy meme thing, the pricing seems really, really nice :D.


This is awesome! I've been meaning to use mailgun for some small personal projects, but didn't want to pay to use my own domain. That's the biggest advantage of the change for me.


These price changes make Mailgun much more expensive than Mandrill (especially for dedicated IPs). Mandrill also has inbound now, so I don't see the point of choosing Mailgun over Mandrill.


Also note that mandrill inbound is complimentary to customers right now ..


This is RAD. I love mailgun's service, and now I have even more reason to never use the built-in mail server on my *nix boxes for processing or sending mail.


I had just switched to Mandrill after being unhappy with Mailgun for unrelated reasons. I'm tempted to switch back to Mailgun after this.


If you're gonna promote your big pricing change, why not get closer, or, ideally, undercut your competitors (Sendgrid, Mandrill)?


Why use a service like this in the first place when you can simply send out mail via a VPS?


Many email services block spammers / unknown senders, especially if you start sending hundreds of mails.


Do you happen to know how that works ? I never used a service like mailgun or mailchimp. Do they use IP addresses that are not being blocked by the email services ?


So its essentially paying for the safe haven of knowing that your mail won't get blocked due to stupid policies?


The input box doesn't work with a French keyboard, nor copy/paste.


Just changed my Plan... Now i am saving $66 per month.


Is it just me or is it odd that after the free tier sending an e-mail is almost as much as letter postage?


huh?

$.50 per 1000 is... nowhere near letter postage.


Sending /1000/ emails is almost as much as letter postage.




Consider applying for YC's W25 batch! Applications are open till Nov 12.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: