Magic Button Service - Please Comment

That would be nice . OK

I like it and I will be in it. But, I my trixbox CE is at home and I do not use all extensinos (in kitchen, basement). So, it will be nice if I can tell the system a portion of the extensions will use the Magic Button service and pay for only those a few extensions.

A few weeks ago I asked if you where going to publish the feature list for 3.0 either after or before lock down.

From a community standpoint it would be nice to share in the resource allocation process. You have X number of resources, each feature or fix should have a budgetary estimate next to it. If the community could participate in the decision making process as to what features/fixes can make it into the release based on the resources allocated to the project.

This would also give a tangible driver for donations as the link from dollars to resources should be obvious (I would hope).

Obviously there is a commercial angle to providing this for the "Magic Button" folks. If they are going to fund the inclusion in the CE release then my vote is of course it should be done. I am surprised that more vendors have not jumped on the bandwagon of including installers for ITSP's or limited licenses to services such as Cepstral.

Thoughts???

....Scott

I would like to see this...
--
Dave Ellis
Northeast VoIP Networks
www.northeastvoip.net

Set a fair one time charge for it. Or maybe as a reward for a donation or a contribution to the wiki or something as such. The last thing I need is recurring charges. It makes since as business model from your stand point... however, the last thing I would consider is installing something that a customer had to be responsible for paying (other than their phone bill). As soon as they forget to pay the bill I'll get a ton of phone calls saying their system is broken. The feature sounds fantastic... please reconsider business model.

A "fair one time charge" model would be prohibitively expensive for most customers due to the cost of licensing of speech ports from LumenVox.

Look at it this way:

You are installing a Trixbox for a customer with 20 extensions. In order to ensure non-blocking of speech ports for that customer, you would probably buy a block of 5 LumenVox licenses for $1,000.

Now consider purchasing licenses through the Community Speech Licensing Server. You don't have to worry about blocking because you get unlimited ports. At $20 per month it will take over 4 years to break even on the LumenVox ports you would have purchased. And that isn't even including possible growth in your customer's business. With new extensions comes new speech ports in an unbalanced manner.

These figures aren't even including the cost of the application, which is included for free when you license ports from the Community Speech Licensing Server.

If there is interest in the model of buying your own ports and purchasing the application, the model would be a one time application fee and an optional yearly fee for updates. You would be responsible for purchasing your own speech ports, installing them, and monitoring your port usage to ensure non-blocking to your customers.

I think you'll get more bang for your buck (especially considering the economy right now) under a monthly licensing arrangement. Plus you have the opportunity to mark up the service under any maintenance agreement you should be selling to your customer. I'm not saying we are opposed to this model. I just truly believe you will get more value, more customer stickiness, more revenue opportunities and less headaches under a monthly licensing model.

I'll be honest. I'd rather just sell it outright rather than dealing with monthly subscriptions, licensing servers, billing integration into firewalling, etc, etc, etc. What we are trying to do here is not make things easier on me. We are trying to bring ultra-high-end capabilities traditionally only available to the largest companies to any sized company.

Oh, and I forgot to mention, your licensing requirements will grow exponentially with every new speech application added to FreePBX. Come down the pipeline: speech-enabled company directory integration and speech-enabled IVRs.

Ethan,

Consider a reseller channel so you don't have to microbill small accounts.

Managed service providers like myself would bundle it in the same manner we bundle CNAME and everything else we pay for incrementally.

Scott

The actual billing is based on usage. You could have a 50 station install pay less than a 5 station install based on usage. On AVERAGE it will run most companies around $1/month per extension based on some test systems. The minimum will probably be around $5/month for the smallest installs and most home users will never exceed that. A flat one-time price will not work because of the Lumenvox licensing server costs to deploy this.

i would pay the once off fee of about $1000, no problems i would not pay a monthly licence fee, maybe you need both pricing models , as i understand it the 'cost' is with the Lumenvox license, i would buy 5 licences i have some other uses for them .

Unfortunately it doesn't use licenses the same way, the licenses are divided amongst multiple users from a central licensing server so that individuals don't have the high startup costs. I will discuss it with SchmoozeCom and see if there is a way to do both models.

This sounds like a really good idea to make it affordable.

What about this though;
I would want to test the system first before we decide to keep it. Could I buy licenses for only 1 month, and then if it does everything we want and we are happy with it, set it up on a direct debit or something???

Also what happens if our Internet fails?
Does it have to make a license request every time someone preses the Magic Button, if it does then pressing the magic button could be rather slooow if the license request over the internet takes a few seconds to confirm?!?!

Well done for the great work.

We have spoken before and this exactly the type of service/functionality I believe you need to provide in order to get trixbox CE self supporting. As far as the licensing cost, the price should be set based on what the market will bear which should include research and feed back from trixbox'ers but also research beyond trixbox'ers. Having licensing a monthly or from a one time is somewhat driven by any pass through licensing you may have to pay. If you have to pay for annual licensing cost, then you clearly can't have a one time cost and stay viable. But either way, any reseller should be able to embed the cost so that the customer is comfortable. Many folks bill their customers a monthly per seat cost. This additional licensing will virtually disappear for these folks.

In regards to a 1 month trial. The subscription and billing is automated and is setup on monthly billing. You can cancel the service at any time.

We have worked with Lumenvox on a internet fail over option. They have agreed to build us a special application that would allow the Magic Button to continue to operate with the internet being done for up to 48 hours. This has not been implemented yet since it requires a substantial investment from us to have this programmed but if we get enough request for the Magic Button we will be happy to go ahead with the programming of this and we do have plans to implement it in the next few month regardless.

On the issue of a 1 time fee. If we went down this route it would require you to buy your own lumenvox ports from Digium and install them. If we allowed this with the Magic Button we would need to see what it would be worth to the community to buy the rights to use the magic Button on a per machine/PBX basis. The 1 time fee would be per PBX and it would not let you install the same license code on 2 different PBX's. I look forward to everyones feedback on this.

Thanks

I will definitely test this on my production office box, but I would like to know is there just a one month commitment for testing? Or is it necessary to pay the monthly fee in an annual installment? After that, if the billing will be based on usage and not actual extensions programmed in the system, then I think that $1/mo per license isn't out of the question. My customers who have T1, PRI, analog lines, or pure SIP trunking already are conditioned to a monthly bill for their communication services. And, I can see me inflating slightly the cost of an annual service contract and letting them beleive that the Magic Button is a service free when they purchase or renew a service contract...when I worked for a major telco, it was rough to sell service contracts, as our Nortel systems were fairly solid and the customers didn't see a need when their system was fine for the last 5 or 10 years...but with this model, my company could couple this service and make it easier to get them to purchase a service contract. I already build in offsite backups into a service contract...this is just one more reason for my customers to let go of their money...which is a good thing.

The Magic Button works in Spanish?

The pricing model is still a bit too vague. I keep seeing phrases like "in most cases" and "on average." I'd like to be able to do some calculations. When can we expect hard data on this pricing model? I realize this is a catch-22: you want people to tell you if they're interested before you work out pricing. Fair enough, but I don't know if I'm interested until I know what the pricing is.

A monthly recurring charge may be difficult enough for someone to swallow, the fact that it seems like it might be variable will be an even tougher sell. Why not set it at $1 per extension (with graduated price breaks?)

Kodak

I understand your concerns. The problem with a set in stone pricing model is that if you have a user who uses the application like the Magic Button a lot it requires us to purchase more ports. We have to put a softcap of some type on how many total port request each PBX is allowed each month. The price model brackets that we have come up with is below with some general guide lines on how many extensions you can plan on running with each bracket of ports. Now if you have a 10 user system who does not use the Magic Button often than they might be able to get away with the lower bracket that is recommended for 1-5 extensions but the same is reverse if you have a PBX customer who uses the application hundreds of times a day they might have to be in a higher bracket than what we recommend. Also remember that there is also a Company Directory and IVR that will be released in the near future that will use your same ports below at no additional cost.

a. Tier 1: 1-5 extensions (2,500 ports/month) - $5/month
b. Tier 2: 6-15 extensions (7,500 ports/month) - $15/month
c. Tier 3: 16-30 extensions (15,000 ports/month) - $30/month
d. Tier 4: 30-75 extensions (30,000 ports/month) - $50/month
e. Overage fees are half-a-cent per request, not to exceed the total cost of a tier for the total number of requests in that month. For example, if you are on Tier 1 and request 20,000 speech ports in a month, you will not be charged $5 plus $87.50 in overage fees. You will only be charged $50 for the up to 30,000 requests that fit in Tier 4. If you exceed the total number of port requests in your Tier, you will be automatically moved to the appropriate Tier for future billings and notified via email

Hope this helps clarify things and please let us know your thought.

Also we are working on a 1 time fee solution for just buying the application on a per PBX basis and you would be responsible for buying your own LumenVox Ports our only catch 22 is what would people be willing to pay for this.

Based on the fact that we can run it for a month to decide if we like it or not, and more importantly it is robust against Internet failures we would definitely have it.

I assume the licenses would be cached so it doesn't have to perform a fetch each and every time for performance also.

Is this expected in 2.6.2 and greater or 3.x?

Cheers,

I'm looking forward to seeing this in action in Vegas. Thanks for the breakdowns, that's a bit more helpful from a budgetary standpoint.

As for set pricing: I don't see any technological reason why a hard limit can't be imposed. You've paid for the $5/month option, so your 2,501st request gets an error "sorry, you have exceeded your licensing for the month." This could be controllable with a tick box by the end user, if they wanted. If "dynamic licenses" is checked then you can go up the tiers and be billed (or debited) accordingly. If it's unchecked (which, IMO, should be the default) you'll get errors at that time. It just allows the end user to set hard limits without having to resort to (probably difficult to enforce) an addendum to their acceptable use policy. Just a thought.

i would get about 3 exten license..

but im sure customers would buy it too!

We would not wait for 3.0, this is something we are considering adding fairly soon. We just need to see if people want it and then figure out how to sell it.

Kerry, I am in useranddevice model. When you say "extension", do you mean device or user? Right now, I have 7 users and 15 devices.

You have to read through this to understand. The prices are ACTUALLY based on usage with an average of about $1 per user per month. You would probably be paying $5 - $10 per month.

Ethan i would pay for the once off application code , here is my thoughts
[1] i would expect the cod to be 'open' in as much that i don't have to run a separate compiler or other software security application to make the code run on my PBX
[2] i would expect reasonable documentation , and maybe an email contact if i get stuck
[3] i would not expect any direct help to install or deploy this application , maybe just an install script and the above documentation
[4] i would probably expect maybe a "goodies" file with some experimental IVR's or other stauff and maybe some DEMO code , but i understand that is is all free, unsupported and at my own risk
[5] i would seperatly purchase a lunevox license and install this - an i understand that the lumenvox server is a seperate purchase to the "Magic Button" Application.

how much would i think was reasonable and fair

[1] http://store.digium.com/productview.php?product_code=8ASTLUMLITE5... = $995

[2] Ethans "The Majic Button" Application = $500

So for a total of $1500 i would have the code and need to self install and configure. At this price point i would be able to sign-off on a project like this. if the price point was more i would have to consider it a bit more.

This is only my personal 2 cents worth.

ddwyer,

While your enthusiasm is appreciated, this scenario is unfortunately not a possibility due to a multitude of factors including our pending patents and those pesky attorneys. If you have large-scale desires for this solution, we could certainly talk privately either via phone, email, or if you happen to be going to AstriCon in September, we'll be presenting there and could arrange a meeting time.

As for your request for other speech-related applicataions, I would urge you to take a look at our already released GPL company directory application and look forward to open speech IVR applications that are being developed cooperatively between us and the fantastic FreePBX team. As long as the community stays behind FreePBX, we intend to release a multitude of next-generation speech and handset applications (such as Aastra XML applications) on the platform with point-and-click configuration ease.

Sorry to disappoint, but I think you will find a plethora of cutting edge open source applications that will be coming out of our collaborations in the future.

Ethan

i think we should all just say good job ethan.

it takes a lot of dedication and time to come up with, work out the kinks, program, and actually get running a program/application on this level.

so i say thank you, and that i cannot wait to try this out.

As a fan of Trixbox / FreePbx and Asterisk
One of the issues is extensions versus devices. For example I have 4 extensions per phone and 5 phones. So in license terms that 20.

I use Aastra 57i's and if I used 8 extensions per phone, that would be 40!

Have a guess why I am not a Trixbox Pro user - its too expesnive vs. 5 users / devices which means its a fair deal for Trixbox Pro / or this Lumenvox program as well.

I guess this is holding back a lot of users. Why not have a lot more small users for a fixed number of devices which is probably where the largest parc of users are.

Your view Kerry?

------------------------------------------------------------------------------------------------------
I specialise in Satellite services carrying data, video and voice. I can advise on services, design and requirements definitions...

Norskman, that doesn't appear to have anything to do with the subject at hand, which is the Magic Button software and licensing. You appear to be talking about trixbox pro pricing, which isn't part of this discussion.

If, on the off-chance that you are talking about the magic button: it's been stated multiple times that pricing is based on usage, not on a fixed number of extensions.

Once again, it is not the number of extensions (users+devices) that will determine your pricing tier. The pricing tier is based on expected usage for the number of users. The total number of devices is irrelevant. If you have 15 physical users (people) and 1,000,000 devices, you will most likely still fit in the $15 tier.

jchuby

What about me? Just kidding. This has been Ethan's project for like a year now. He has put thousands of hours into this and has done an incredible job. I use the magic button everyday in m own life and I recently installed a phone system at home because my wife was jealous after playing with it at our office. Now I just need to figure out how to clone Ethan and life would be great.

Just to be clear. I am sorry if i have missed this already but as it is based on usage where 'on average' it works out as being about 1$ per month per extension, exactly how many times can you use the magic button for 1$???

We have developers who would barely use it and a support team who would hammer it, and so I would want to calculate the usage and add a contingency based on a usage trial and not take a broad 'average'.

Also thinking about it, we are based in the UK, and so for performance reasons the licenses would have to be cached as to query a license server in the states every single time would be unusable.

I hope that made sense! Cheers.

ok,ok: lets all thank Shmooze(com) too...
and while we're at it Kerry G...

Anyone else feeling left out??

See Tony's post for details on ports. UK shouldn't be a problem even with our licensing servers in the states. It's about 1 kilobyte of data round trip. If there was enough demand in various parts of the world we would drop licensing servers geographically.

The LumenVox licensing mechanism works by requesting licenses from a licensing server (either locally or elsewhere) and the actual speech recognition processing is done on your local machine, so there is no voice data being transferred anywhere for processing. It's possible to have centralized speech engine servers, as well, but that is not something we would ever consider doing for performance reasons.

Don't suppose anyone is planning on or already has put together an example XML config for the Aastra devices to create a 'Magic Button' :)

I think you are confused on the Magic Button. It has nothing to do with XML. It is a standalone application that we wrote that uses LumenVox for its speech application.

Scott

I would love to have resellers handle micro billing but that is not possible due to our agreements with LumenVox and such.

Hi Schmooze. Sorry if i wasn't being very clear.
When i say XML, I am refering to the code that is hosted on trixbox and downloaded by Aastra phones to control and program the dynamic/programmable buttons on those phones.

So for example when you press one of the buttons on the phone it links you into the Magic Button application and you can start talking your voice commands.

The application will have a feature code or extension number, so you just program a speed dial no different from any other that just dials the code/extension for the app.

Well, I assume. I don't know how else you could invoke the application.

Or maybe the "magic" label is more apt than we think and it actually reads your mind.

I'm surprised that no one asked about the following, because they immediately jumped to mind:

1) Speed. Why would users want to press a button and then talk to do something that takes only one, two or three button presses already?
2) General Accuracy. In my experience, public, untrained LUIs drive users absolutely stark raving mad, because the accuracy is never good enough.
3) Heavy Accents. Here in Toronto, less than half the population, including those in my clients' offices, speak English intelligibly despite living in a nominally English-speaking city. For most, it's a 2nd or 3rd language. If I can barely understand their English, what chance does the LUI have?

While speech recognition seems like a nifty feature, I suspect that 10% of users would love it, 10% would be neutral, and the remaining 80% would hate it passionately (and therefore decline to use it, or struggle and curse me under their breath) . This makes the hassle of licensing (especially monthly) fairly dubious.

Still... I'd love to play with it myself ;)

1.) When you have 50 people in a company, remembering extension numbers becomes more difficult. Push the button, say "call Ethan Schroeder". At minimum it saved you 4 key strokes and a glance at the wall at the directory. When you are out of the office, are you going to have a list of extensions with you? Call into the magic button from your cell phone, say "Call Ethan Schroeder". Done.
2.) Modern speech engines are not only designed to be speaker independent (no training) but also optimized for telephony. If you haven't used one lately, I encourage you to try LumenVox with an open mind. I think you will be very surprised with the level of accuracy you can attain out of the speech engine.
3.) Can't comment on this because I don't have enough data. I do know that the LumenVox engine supports specific dialect subsets of languages (such as UK english, US english and Australian English), which would theoretically be designed to accommodate specific accents in a dialect.

As far as your estimates of user traction, I would say flip the numbers. Our customers love it, and we wouldn't be able to rip it away from them. The only thing that is frustrating for our customers on the beta program right now is our lack of speech ports, which makes the service unavailable to them from time to time. They tend to get rather used to it being available to them because they use it 20 times a day. This problem, of course, is easily remedied by us purchasing more speech ports, which we continue to do as our beta program increases in size.

Take it for what it's worth, but I think we're starting a revolution when it comes to the integration of speech recognition and business processes, especially considering our customers get it for free, and your customers can get it for a ridiculously low price. Time will tell.

I'm beta testing the magic button and it is the best damn thing I have ever seen! I'm very excited to see "easy" voice calling. I press one button and I don't have to worry about remembering anything (i.e ext#s - paging the warehouse - sending voicemails to a group of employees is alot easier - and calling in on my cell phone is fantastic. kerryg you should add this to the dashboard for certain!!!!!!!!!!!!!

Any plans to add this to tb Pro? I'm sure it would be tested in the CE version for some time but hopefully there are long-term plans to roll this out to the Pro versions pending good feedback?