There are tons of email service, free or premium or custom domain hosting. Most popular ones are Google, Outlook. The most recent new player in the market that I know was Hey.com. I wonder, if the market has room for new email service provider in the market. If yes, what sort of innovation should we expect, or email problems to solve?
There is always room for a new product if you provide value that established entities don't offer.
Proton Mail for example built it's business around privacy.
There is alot email is capable of that has been forgotten.
1. For example I am part of this mailing list that is like 30 years old. And it feels sort of like a small niche social network. There is alot of improvement that needs to be done in this space.
2. Alias generation. I really would like to subscribe to some news letters or place my email out in public but I don't want to do it with my real email for fear of being spammed.
3. Cataloguing & curating emails. It would be really nice if I could book mark different messages into different folders.
4. Read receipts. All email hosts claim to have this feature but I have actually failed to use them for every single service I am signed up to.
5. Make it easy to configure custom domain names.
6. Tagging doesn't work on Gmail's mobile app. It's so badly implemented on other services.
7. Have a good API why do I have to go to sendgrid?
Fastmail has Masked Email, it even integrates with 1Password so you can generate a new email as you're registering for a new site.
It's easy to unsubscribe from a shitty service or a newsletter, just delete the email and that's it.
They also have custom domain names, you do need to do the requisite DNS magic, but they give you step by step instructions where you just copy-paste stuff. And I think you can have them host the domain so they'll handle everything.
Fastmail labels ("tagging") works pretty well too, you just need to write the rules yourself, there is no fancy magic doing it automatically like Gmail.
> Alias generation. I really would like to subscribe to some news letters or place my email out in public but I don't want to do it with my real email for fear of being spammed.
+1 for this.. I'd like to give out specific emails while filling forms and be able to deactivate/delete it as soon as i don't need it.
> Alias generation. I really would like to subscribe to some news letters or place my email out in public but I don't want to do it with my real email for fear of being spammed.
If only one could give out an e-mail address, which would be a valid destination from a single, intended source ...
Yes I just did, in a private window, and all I needed to do was complete a captcha.
But even if your IP is on their naughty list, you can delete that data once you have registered, and I would be very surprised if Proton secretly kept it after you delete it.
I recently launched a new email service called Port87. It's very different from how current email providers work. It's all based on tagged addresses, aka subaddressing.
My email is hperrin@port87.com.
When I want to give my email to a service, like Netflix, I wouldn't use that though. I'd use something like hperrin-netflix@port87.com. Then when Netflix sends an email there, Port87 will automatically create a pending "netflix" label for me to approve. All of Netflix's emails will go there. Each email address is separated from all other accounts, the email is automatically organized, and you can control whether you want notifications and to mark new email as read per label.
Then if I want to give my email to a friend, I would use something like hperrin-friends@port87.com. My "friends" label has screening enabled, so any new sender gets an email back asking them to click a link to prove they're human before their email is delivered.
Finally, I can share hperrin@port87.com publicly, because any email sent there gets an autoresponse with a list of addresses for my "public labels". The sender can choose which label is best for their message. For example, I have a lot of open source projects, so I have a public "opensource" label.
I've been using it for over a year now since I launched it publicly, and it's really a night and day difference over the more traditional style email services.
I agree with you for 99% of cases but there can be unexpected domains in the mail, or more interestingly, having your data sold. If hyperrin-nexflux2@port69 starts getting random spams for example, then you have a good idea who fucked up
Because senders don't always use the same domain, and this way lets you block the entire address. If a service leaks your email address, you can just block that email address.
It still all just goes to your inbox. It won't put it in a label for you, you can't control whether you get push notifications per address, and it won't autorespond with a challenge on ones meant for real people.
Personally I'm using Migadu at the moment and for me the killer feature is that the price per month is not bound to users and/or domains. I have several tens of "users" (many are myself but for reasons with completely separate identity as far as the server is concerned), and 8 or so domains. There is barely any traffic (in the order of <10 emails in and out per day).
Most other providers I would have to pay 10-30x as much, just because I have several users and domains, at a grand total overhead cost of 0$ to the provider.
They have a lot of positive points, but I can give a few negatives to try to balance out the opinions:
* When sending emails to others, in my experience they usually arrive at spam. This could be because in those cases I was sending from domains less than one year old, but I don't think I've had this issue with other providers (e.g. Protonmail).
* Using them to send emails in any automated way (e.g. scripting) is against their TOS. I understand their reasoning, but I would have liked this rule to be relaxed a bit because I would really like to use it for sending notifications to myself, so I would be happy even with a restriction like "if sender and receiver are the same address (or part of the same mailbox), you're allowed to use scripts to send those".
* They don't guarantee any kind of SLA. If they go down, they go down. Not a problem most of the time, but during the start of The Pandemic they had a lot of downtime and that made me realize how painful it is to try to use email when the provider has flaky availability. They fixed it after ~3 months or so, but that was bad enough to make me stop using them for some time. When I heard they fixed it I gave them a second chance and haven't had any issues since then (it's been 3-4 years already and I haven't noticed issues). But this single experience is hard to forget, so I don't use them for anything critical (stuff I would need to read in less than 24 hours or so, much less respond), just in case. Even if SMTP should mostly ensure that you eventually receive emails, and that your sent emails eventually reach their destination, I still have the critical stuff sent to a Gmail or Outlook address.
Despite those points, I like their service enough that I still recommend them from time to time.
I've been using them for years and did not realize they don't allow automated emails. At some point I was sending some automated emails from one inbox to another, but they were very infrequent, less than 1 a day. If you're looking for email-based notifications, I'd recommend Pushover [0] anyways. It's more flexible (can do email, curl, etc.), free, and won't violate ToS.
I'm going to give you a niche idea.
An email provider / hosting / client focused on a single user owning a whole domain.
I don't care about individual email addresses. What I really want is to be able to handle asterisk@asterisk.example.com both for receiving but more importantly for sending.
Receiving is fairly easy with aliases and catch-alls. But for some reason all clients / providers make it really hard to send emails from any address you own.
This would be nice. I am using Proton right now, with my own domain. I can make up any address and have it go to my catch-all inbox, which works decently well. However, every so often something will want me to send an email and that's a problem.
I was buying a pair of shoes online and some kind of fraud flag was raised. They asked me to reply to their email to prove I owned it. I ended up replying from my actual account, as I didn't want to burn one of my actual address on a one-time thing. I never gave the email I sent from to anyone, and it has received 0 emails, I really didn't want to risk having it added to their email list by using it, but didn't feel like I had a lot of option. I went on a bit of a rant as a reply, explaining my whole setup, why wha they were asking was problematic, and warning them how irate I would be if the email I was sending from ever got spam in the future, I'd know it was from them.
Of course, the more I try to solve this email problem by using additional addresses, aliases, etc, the more I think I'm just creating a larger problem. If I could go back in time, I would probably not do any of it, and instead just have the 1 address that I use for everything to keep it easy. Instead of investing time in making new accounts, actually invest time in clicking unsubscribe, blocking, and training the spam filters... also, being very selective about where it's used, although that can be difficult these days.
This is part of a feature I'm working on for my service, Port87. Single user domains would work the way labels work in the @port87.com accounts. So all you'd need to do to send from an address like "myaddress@example.com" is create a label called "myaddress".
I don't have the custom domains feature ready yet, but once I launch it, it sounds like it would fit your needs.
- protect my data from overzealous law enforcement
Right now, I'm running my own mail server on a dedicated machine. I wouldn't mind getting rid of it. One feature i need is a configurable sub-adressing separator (not just "+").
> One feature i need is a configurable sub-adressing separator (not just "+").
After years of training my family to use "U-C@D" to help identify where spam is coming from (because there's still an astonishing number of sites that refuse '+' as a valid email character), I realised I'd trapped myself into (probably) having to provide the email server forever because I've yet to find a hosted alternative that provides configurable/additional subaddseps.
Migadu has a feature called "Pattern Rewrites" which would allow creating a wildcard rule to automatically redirect mails to "U-*@D" into a "U@D" mailbox.
Do you have your own domain? I use email alias (eg WifeLinkedin@domain.com ) with a single alias per site. I send all alias minus the 2-3 I care to a folder.
General aliases are bad imho because it tends (in my experience) to attract spam... One way I get around that, is instead of aliasing @domain.com, I do .a@domain.com, so I use my emails when i subscribe ServiceName.a@domain.com
I think so. Another recent entry is purelymail, which I love because it's a flat rate for unlimited domains. I'm trying to spin up multiple "small bets", and per-user pricing is unreasonable, and for better or worse, I don't want a single catch-all email.
A secondary benefit for me is that Purelymail is cheap ($10/yr). But a major downside is that the CalDAV implementation doesn't yet support Accepted/Rejected scheduling responses. It's fantastic that they have CalDAV at all, but this one feature would seal the deal for me.
I don't mention it to suggest you could directly compete on that unlimited-users value prop, but it's an example along the lines that max_ suggests.
One that comes to mind is integration with recent ActivityPub / fediverse stuff. There might be cool design space there
I think email today "just works," esp. since spam filtering became decent and auto-categorization. I don't think being a new email host is where the win could be but at the client end. That's what recent successes like Superhuman, Hey have been.
HEY is not a client but a full email hosting service. I disagree that "email just works". As other commenters here have pointed out, there are many niches and improvements to be made that match with evolving trends in society, e.g., privacy-conciousness.
Until Thunderbird added support for Exchange there was a slight hole on Linux (no FOSS mail solution with support for it) but right now I'd say everything works more or less out of the box. SPAM filtering and categorization could be a bit more automated, but in general I'd say everything is covered.
Mail server projects are a different story though. There's no simple way to create a fully fledged mail server without weeks of pain to configure spam, routing, calendars, security... there are some preconfigured images in all major cloud platforms, but nothing like `apt install mail-server` that pops out a simple config assistant in an already existing server.
I've been using Evolution EWS for well over a decade.
I've been running small mail systems for around 25 years. Just like learning to drive - lots of stuff to learn up front. Just remember to fill up every now and again and keep on top of maintenance.
I think maintaining the MUA access to the mail server is one of the more complicated issues. A globally accessible IMAP or POP3 service faces a lot of abuse. These days, I use MUA's the that read directly off the spool. Old fashioned, but heads off all kinds of badness.
No listener required, or the services, authentication and encryption that go with them. In this case I run email for myself. Email gets pushed from an Internet MTA with global addresses, to a local MTA that happens to be my workstation. I either read the email from the console, or ssh in through a jump host.
Does anybody remember Eudora? Not the Thunderbird version, the one before that. It was fast as hell with cool shortcuts (eg Alt-clicking on from groups all the emails from that sender) and therefore super easy to find stuff and organize your email even if you had way too much of it. My M1 MBP is probably 2000x faster than last computer I had running Eudora, but modern email is still slower and clunkier. A web based email service that is blazing fast and makes sorting/filtering/searching my email easy is something I would be happy to pay for.
Another email service I want would be one that I have to authorize the sender before their mail gets into my inbox even if it's not classified as spam. Sure, that email can be routed to some dumping ground so I can sift through it if an expected message is missing, but by default nothing goes to my inbox unless I've routed the sender there. And when I "accept" a sender it should be trivially easy to route them to any mailbox or make a new one. And it should be trivial to manage those routing rules later. All this is already possible with existing services, but enough of a pain that I don't maintain it.
The last thing is a tough one: a way to port my existing email into the system and a guarantee that I won't lose my email if you shut down in a few years. Trying other email providers is not worth it if I won't be able to find anything from the past and I might be at risk of losing them in the future.
An email provider with a high quality, efficient native app that has first-class offline support, threading, some well-thought-through AI-based categorization of emails, respects my privacy, and does not issue network requests based on email content would be quite welcome, IMO.
Look at Fastmail for a surprisingly popular service that gets most of this massively wrong.
Their apps are entirely lacking threading support, and IMO more egregiously, offline support.
Fastmail likes to brag about their JMAP protocol, and maybe JMAP is great and maybe it's not, but it's certainly the case that their iOS app is 100% nonfunctional when offline. I've been using various forms of email for quite a while. Microsoft's MAPI crud in all its incarnations works offline. POP3 works (albeit poorly) offline. IMAP works offline. Fastmail does not.
I meant not issuing network requests for resources referenced in the email. Kind of like how hey.com goes out of its way not to load remote resources that can track email reads.
I would expect: fair pricing, local (as in the country you live in) hosting (for legal reasons), features like using your own domain, maybe service for calDAV cardDAV etc. to tie up the bundle.
Also no user data aggregation, no excessive logging, etc.
Really, the one thing I would look for to recommend this to others is having that _and_ ease of use for non-techies.
I could be wrong but I think a good way to have an edge is pricing. Email is generally a low bandwidth service with lots of cold storage, yet for most personal email services the price doesn't seem to reflect this.
Like for example Google One is $2/100GB and $10 for 2TB yet you can definitely get storage much cheaper than this, and bandwidth is close to, if not free on bare metal clouds.
I've been considering this myself. That or some IRCv3 offering.
If you go the commercial route you'll almost certainly end up building an ERP/CRM, as it's what a lot of companies seem want: a personalized ERP/CRM (not sure why, though it does kind of feels like just outsourcing an IT dept whenever this has come up in the past)
My wish list:
1. iOS native sync, preferable without Active Sync. This includes syncing of read states.
2. Custom domains for families and small business without shelling out $10 per user and month.
3. A shared mailbox within the custom domain as „family“ contact mailbox. At least two people should be able to read it, possibly sync read states AND answer from that address, including sent mails need to be stored in this mailbox.
4. A very simple way to setup the accounts on different systems, especially iOS.
AFAIK, except 4 those are already solved, but not within one provider - without paying those mentioned $10/u/m.
> 4. A very simple way to setup the accounts on different systems, especially iOS.
On iOS, at least, you can do this using a `.mobileconfig` profile (which is just an XML plist file). Create that, get your user to access it via HTTP, messages, etc., and iOS will configure the mail account for them. I've used it when family members have lost/forgotten/wiped their details.
Last time I checked it wasn‘t possible to configure multiple aliases via this method and it wasn‘t possible to add aliases after adding an account this way.
This is a few years back, so it might have changed.
I get that subscriptions pile up, but email is the cornerstone of your identity online. Your login and recovery for most services. Getting your mails through requires work these days. I think overall is good value.
Protonmail is increasingly Gmail-like in every respect -- with its AI "tools," with it asking you for personal information upon account setup, and in obliging cooperation with foreign courts -- so a new service which doesn't deviate from a focus on true privacy would rapidly find its niche.
I want a "Mail-in-a-box"/Plex mashup... that lets self hosters setup a domain, with a mail server than handles all the spam and "trust" shit that comes with hosting a mail server.
The new hosting service does all the trust and anti spam/virus/phishing stuff, and maybe does or helps with the neat stuff you get from gmail (indexing and clever integration to calendar and travel/shipping integrations.) We'll either buy VPS or self host on hardware in our house the actual data.
I replaced Outlook with Obsidian + syncthing. It would be nice to add email into mix. Some sort of service that fetches mails from multiple sources, and exposes them as filesystem. I want to link individial emails from notes.
Am I the only person who thinks email search is completely useless? I’d happily pay for a service that lets me search selected fields for a substring (eg title only, title and body). Give me regexes and I’ll pay more.
I need server-side filtering rules that I can configure with a web UI. That basically leaves only Exchange and GMail. AFAIK, there's no privacy-respecting small provider offering server-side filtering rules.
I suppose it depends on your definition of "privacy-respecting" and "server-side" but, for example, Proton offer SIEVE filters you can configure with a web UI.
Email is root node for almost all services. Some future improvement may be a root node and also email -- but not necessarily bundled as a unifed item. Maybe it will have some web3 stuff in it or similiar.
I think there could be, sort of. I am currently using Migadu, which works great and I don't plan on leaving. Their selling point is that the amount of domains you add is not limited. It's affordable and reasonable, and a great option as indie dev with multiple projects. No fictional per-seat pricing that does not correlate with costs to the provider.
What I think is outside their target market, but what would be a great next step for projects with slightly more needs would be a Google Workspace/Outlook sort of service with a similar approach. Simple CRM features, shared contact management, calendars, signatures, etc.
Roughly what Proton is doing without the Silicon Valley pricing, or maybe even Hubspot.
I use Migadu, mostly for people who just want a couple of domain emails when we host their site. It works good for that and I like the accessibility of the logging. Needs a user-accessible out of office I don't want users emailing me to be their personal assistant. I did try to use it for transactional type emails, sending forms from Wordpress or scanners etc but it wasn't so good for that because the emails were too spammy and the accounts kept getting deactivated. autodiscover was a bit hit and miss, nobody likes talking people through imap servers over the phone, but it seems a bit better now.
Yup. Everything left and right of being an email inbox is a bit iffy, and that makes it feel a bit incomplete at times. Like I said, I like it a lot, and it provides a lot of value for the price I pay for it, but I wouldn't mind an option with just a slightly wider feature scope.
The big advantage with Google Workspace/Outlook is its accompanying products - calendar, word/spreadsheet processing, cloud storage, etc. Problem with starting only email service won't make the cut, but starting with email + other services make the project way more complex. Not forgetting answering the question "how each product (i.e. word processing, cloud drive, etc.) is better than Google/Outlook's?"
Yeah, you're right. Just a hosted inbox alone would be hard to differentiate in the market and I have no answer for that. For hosted email with auxiliary services there is room to carve out a market. At least that's where I see some value. As for being better: It really doesn't have to be. It just has to appeal to a different audience. I think non-VC funded SMBs that don't have Silicon Valley funds to spend on a dozen different per-seat products is a pretty big market.
What I am thinking doesn't have to compete with the whole Office or Google Docs suite, or HubSpot. There is imo enough value in being an economic all-in-one email solution for "the rest of us". Point DNS there, get email, CRM, maybe even a small transactional mail quota and some marketing email features.
I have two projects I would onboard yesterday if this existed, and I have spent more than a hot minute thinking about going for it myself, just to scratch my own itch. The main reason so far that I haven't is that Migadu + Resend is enough for my needs at this time, and I can do without CRM for now.
Proton Mail for example built it's business around privacy.
There is alot email is capable of that has been forgotten.
1. For example I am part of this mailing list that is like 30 years old. And it feels sort of like a small niche social network. There is alot of improvement that needs to be done in this space.
2. Alias generation. I really would like to subscribe to some news letters or place my email out in public but I don't want to do it with my real email for fear of being spammed.
3. Cataloguing & curating emails. It would be really nice if I could book mark different messages into different folders.
4. Read receipts. All email hosts claim to have this feature but I have actually failed to use them for every single service I am signed up to.
5. Make it easy to configure custom domain names.
6. Tagging doesn't work on Gmail's mobile app. It's so badly implemented on other services.
7. Have a good API why do I have to go to sendgrid?