Where are you buying your DIDs and Routes? What switch are you using?

I'd love to help, I literally eat sleep and breathe this stuff. There's no way your margins should be skinny unless you're playing LCR games, in which case you make it up on volume.

Source: I work at 2600hz, the bootstrapped open source telecom cloud company. We've bootstrapped to 30+ employees on a pure telecom business so I have some experience here.

I'm currently on Twilio's first tier volume discount (I assume there's further volume discounts that aren't on the website), paying .8 cents/min IIRC.

I looked into buying voice minutes and DIDs and running my own freeswitch boxes. Any provider that had listed prices charged upwards of .4 cents/min or had unlimited minutes per DID with a $4 or something fee per channel. A lot of them also limited the number of channels on a DID.

Another reason I have low margins though is because my free plan and free trials also cost me real money. There is no way I can get around that.

While we're at it: Another reason I looked into self-hosting was because I'd like to host conference calls with up to 5000 participants (think earnings calls, public forums, classes etc), most of which are mute. THAT is a high margin business, with companies charging up to 20 cents/minute/caller for those kinds of calls. I read this may be possible with freeswitch on beefy hardware. What do you think? Do you know any way I could get around hosting my own freeswitch servers, load balancer, redundant backups etc?

Have you considered rolling back or sharply limiting your free trials? With the exception of people you've made commitments to, if they aren't accomplishing a business goal for you, you don't owe the world them. There is nothing intrinsic about SaaS that says there has to be a free plan. (Additional options: aggressively using the free plan as a viral spread mechanism. Your free competitors do this, as I learned the other day when on a sales call organized by somebody who would have happily dropped $50 to remove the external branding for that single phone call if they were aware of your option existing.)

We also run on Twilio, and have a 30 day free trial with CC required upfront. While we still have marginal expenses to service free trials in a way that many SaaSes do not, the costs are not terrible and easy to justify as a cost of customer acquisition. (We shoot for, and mostly get, 80%+ margins on the paid plans.)

If you're ever in the mood to chitchat about this, drop me an email. I'd be happy to share about how e.g. moving a bit upmarket and doing a bit of enterprise sales has worked for us.

Thanks for chiming in!! :) I always find a lot of value in your advice.

I started out asking for a CC before the trial, but on a small adwords test campaign that converted exactly zero out of a few hundred people. Then again, that may have been because the entire site was a single signup/cc form and had very little information otherwise.

Putting my marketing hat on, I'd strongly suggest Patio11's blog on inbound marketing. He's really good and the points he make are dead on (the big marketers are basically doing the same thing except paying Marketo and Hubspot for automation).

I think it would be difficult to do 5000 participants in a conf call on stock freeswitch boxes, but we've done over 20k using Kazoo (Which leverages FreeSWITCH as the media server). Our secret sauce is an abstraction layer and custom freeswitch modules for distributed scaling. When you get over a few hundred participants in a conference call you're always doing some sort of bridging between boxen and when you chain lots and lots of boxes together keeping state across the cluster becomes a challenge. We've done a lot of work handling distributed state in conferences; it was a pain.

If you want us to host something like that for you, drop me a line in my inbox. Contact info in my profile (we're also open-source if you wanna get your hands dirty).

Please note: Kazoo (our stack) is built for carrier deployments, and requires a minimum of 8 servers for full redundancy. You can get by with as few as two but full redundancy requires 8 (and is thus our recommended minimum configuration). But it does have all of the stuff you've mentioned and you don't need beefy hardware to do it either.

I'd also like to note two things:

1) if you want to charge more, tell a better story

2) if you want to charge more, take features out of your product

Those two axioms are usually true and apply pretty generally to SaaS businesses.

At 2600hz we started with free plans and then realized the headache, now we mostly do paid support (there's a lot of free-ness because we're open-source anyways).

Generally though, I'd recommend against free trials. They just waste a lot of your time with tire kickers.

Do you know how easy/hard it is to start a "international calling card" service?

I see lots of random ones popping up every day and they mostly look unprofessional so they remind me of email/web-hosting resellers.

I was wondering if reselling "international calling card" is a thing? or do you actually need your own infrastructure?

I was hoping that I can resell these services and make some passive income that way (I can directly market to low-competition markets in my home country).

The typical way this is is done is to get a carrier route to a specific country and then arbitrage between that rate and the expected rate from say AT&T. There's usually a lot of margin when you compare it to AT&T costs, but you'll find the business is ultra competitive. We call voice the race to zero.

Erlang was basically built for telecom so it made sense for us to build kazoo on top.

Keep your eyes open over the next 60 days, we've got some fireworks coming out.

It all depends on volume and cost to deliver. Some of the biggest providers have the highest costs to deliver service.

Does that help?

Edit: for example, a hosted PBX provider, pricing at $35 all inclusive per seat, should have gross margins north of 150% and net margins above 40%, assuming at least 10,000 seats. There are, of course, huge variations because of management style, frugality and a myriad of other factors, but there's money to be made.

The easiest money in telecom, with the highest margins, are the Analytics, billing and operations tools (of which there are many but finding quality ones is a YMMV proposition).

Your lowest plan also feels too inexpensive to me, especially since those customers are probably responsible for a disproportionate number of support requests.

Those plans are far too lean.

Also, I don't understand what a "local line" is when I go directly to your pricing page. I also don't understand what a "toll free line" is. Specifically I don't know why I need 5 lines.

I don't run a pure SaaS (http://gridspy.com) but $9 / mo? These people are not employees. They have profits. We had to fight the same tendancy to offer unrealistically low prices.

Your typical "professional" will bill $80-$200 / hr while on the conference call. They'd swallow $40 +

You probably don't need a free plan, though you might want to consider a prepaid one. Bill on minutes or something. I'd consider eliminating it entirely.

Team should be about $200 and business you should be thinking $1000 / month plus.

I also think that segmenting based on max people per conference makes a lot of sense. It would also be simpler to bill based on "number of conferences" rather than "number of minutes" and trigger a new "conference" when a call extends beyond 1 hour.

Higher prices should support more marketing work. You'd be surprised - higher prices will make your clients think that your product is more valuable and thus more likely to try it out.

You've also got a viral element. Your users are telling others about you (to invite them to the conference). Ask them to put each person's phone number and email addresses into the system so you can send reminders to each participant before the meeting.

In those reminders, I suggest you offer a "coupon" for 1-5 free conference calls to that guest. You might only want to do so if the guest has an email from a different domain. Monitor for abuse. This will help you get natural viral marketing underway.

Another alternative is to build a followup email after the conference ends that has details on who joined the meeting, how long they talked, a downloadable recording (?) and their contact details.

I hope this helps. It looks like you have a nice product.

1) Most BIG conference calling companies are registered as competitive local exchange carriers.

2) As CLECs, these companies get to charge for inbound traffic.

3) Charge customers per minute in addition for insane profit.

Companies like FreeConferenceCall.com are basically ignoring the subscriber revenue in exchange for only the inbound tariff (that's their industry disruption).

Example: I setup a conference call bridge on FreeConferenceCall.com and I call in. One other person joins me. I call in on my cellphone and so does the other person. Neither one of us pays for long distance so we don't care, but on the backend, FreeConferenceCall.com is billing our cell providers $.30 per minute.

That's the reason GoogleVoice blocks those numbers, because Google is already eating $.004/min for US numbers, they can't afford to eat calls that cost $.30/min; that's nuts.

This is the same business model that powers MagicJack as well. Just a quick FYI. Paying people for inbound minutes is only for folks that don't have enough volume to get a CLEC-level agreement (you can get CLECs to sell you inbound for $0 or even to pay you for the traffic if you have enough volume). The beauty of Twilio's business model is that they get all of the aggregation benefits of your traffic.

Does that help?

Edit: Also the comment I'm replying to has a ton of great advice!

BTW, FreeConferenceCalls know their model is doomed and are expanding as fast as they can into countries with phone systems modeled after the US one, where the loophole still persists. It's pretty clear, though, that their business isn't sustainable. My guess is that somebody will soon gobble up FreeConferenceCall for their customer database.

The real question is: Who's gonna get more spam, The old Acme Packet customers who are now getting spammed by Oracle or the soon-to-be-potentially-sold FreeConferenceCall.com clients?

It has to do with treaties and international tariffs versus state to state. When you hit a certain scale, say Verizon scale, you get to have your traffic carried for free by the other major telcos. It's all a volume game.

In short, people without volume maximize the rules of the system in order to increase their margins. One such method was to pretend calls were coming in from Canada when they were actually state to state.

Another odd fact: Calls Intra-state (from one county to another) are just about universally more expensive than calls Inter-state. The reason is, when you call out of the state, more often than not, someone is paying for receiving that call, but when you call intra-state all of the cost is born by the local operator.

Cute, right?

I don't have any data to suggest that larger companies have bigger conference calls, but it may still help the value perception of more expensive plans to limit by number of callers. In fact, most of our $9/month customers aren't Professionals that charge $80-$200 per hour, but people that would otherwise use FreeConferenceCall et al and are already stingy about spending that much. A lot of "randos" signing up from Google or ads are very small companies, non profits, organization committees for various things etc. I like those customers and don't want to lose them, but I agree that I should find a way to separate them from the lawyers that sometimes get the $9/month plan as well.

The reason you'd need multiple lines is if you're multiple people. The premise of HipDial is that each person has their own, dedicated conference line and phone number. E.g. sales teams with a few people will sign up for the Team plan.

The viral mechanics actually work out pretty well like you said. That's the reason we have the free plan, too. On the free plan we used to even send text messages to the participants after the call, asking them to use HipDial as well, but those converted so badly and gave us such a bad rep that we canceled them.

The obvious thing to do is have a announcement to each person who dials in which advertises your service, and pay to allow people to customize/avoid that.

Alternatively, you could limit the length of calls on the cheaper plan. Poorer customers will just redial, richer ones will pay more to avoid it.

These companies break down into a few broad categories:

1) it's new technology (this is the holy shit category)

2) it's cheaper (IMHO boring)

3) it's higher quality (think uberconference)

4) It's a different experience (Grouptalent using conference bridges for interview screening)

Most of these services are of category 2 or 3 and aren't that interesting. The 1's and sometimes the 4's have the potential to be really cool.

I don't know where speek fits but I think it's a 2 or 3.

Edit: not to say 2 or 3 businesses can't be hugely profitable but they're less interesting to hackers. There are still Multi-billion markets for businesses in category 2 or 3. I happen to love the guys at Uberconference and think they have a great product as a sidenote.

There are significant fixed costs associated with the development they did so far. I doubt $5,000/month would break them even considering just the salaries and office expenses such as rent.

- Fees to receive payments

- Hosting

- Support services (github, zendesk, etc.)

- Your own salary

I'll bet that not much of that $5k is profit.

   > The best part about their $5,000 is that most of it will be profit.