“Make something people want.”

What a great motto!

But, how do you know what people want? for us, the answer has been to build something good but initially quite simple and get it to the market as quickly and possible. Now we try to iterate fast on improvements and additional features based on actual customer feedback.

In order to do this, we spent the time since the launch, beside supporting our already existing customers, in a series of meetings, talks and email exchanges with potential customers.We engage the customer based on the offering that we already have and give them a demo. If they are interested all the better, but if they are not, which is often the case with any new product which has some innovative aspect to it, we ask them the Three Questions:

1. What is lacking in what we showed you? what additional features would make you consider buying it?
2. What bothers you most in your technical support operation?
3. What bothered you most in the past and how did you resolve that?

Our purpose is to implement some sort of pilot implementation based on the most interesting answers that we’ve heard and get back to the responders with the results.

And while this process is not nearly even half finished, the results thus far were truly eye opening and we thought it might be a good idea to share some of them here and get your feedback.

So, here they are: what are users and would be users really want, also known as the SupportMatrix Totally Non-Official Roadmap:

Product Knowledge Base

Enable the user to see variety of information regarding the hardware/software setup of the current session, within the browser and side by side the DriverMatrix session, using a simple tabs interface.

The tabs page replaces the existing operator view, and the session tab includes the existing functionality:

Click on the thumbnail to see the full screen shot

When the operator clicks the “Description” tab, he or she gets information for the product or application.

For example, here is a would be page for the a printer driver, including a photo of the printer, textual description, and a link to driver page:

Click on the thumbnail to see the full screen shot

The description is inserted by administrator in a normal administrator driver view using a WYSIWYG html editor, which is embedded in the administration interface:

Click on the thumbnail to see the full screen shot

Statistics module

This feature adds a wealth of statistics and reports regarding the system usage for the support center managers as well as system administrator. The statistics covers sessions lengths, type of drivers, products and operating systems and system performance statistics.

Data about the activity is gathered constantly during the systems operation, and the statistics are periodically updated.

Some statistics are targeted towards the support center manager and record such things as which product or operating system takes longer to support:

Click on the thumbnail to see the full screen shot

Other statistics are targeted towards the system administrator and can be used as a tool to fine tune the system to run better:

Click on the thumbnail to see the full screen shot

Reports export

Enables the system administrators to export data from the system to other systems in different formats, such as CSV, for further analysis.

Session Preloading

Enable shortening the perceived delay time of a session start: the CRM system can signal DriverMatrix to preload the server side of the session, once it knows that there is a high probability that an operator will require this specific session. For such a preloaded session, the operator browser will start with a much shorter delay.

Screen Shots Based Instructions

Enables the operator to record his activity with the DriverMatrix sessions as screen-shots or video, for sending the end customer as a visual instructions when communicating by email.

So, which of the above features did you like and would like to see in the next version? or better yet, which feature is missing? do let us know.

Luckily, DriverMatrix has been designed to be a good team player: it is essentially effort-less to integrate into your existing infrastructure, whether a trillion dollars CRM system from behemoths like Oracle or an in-house developed solution.

To prove the point, we went ahead a created a module that integrates DriverMatrix™ into SugarCRM ™, the world leading Open Source CRM solution.

The connector allows a support center operator to assign a DriverMatrix™ profile to each support ticket opened via the SugarCRM™ system and start a DriverMatrix™ session directly from the SugarCRM™ web interface, thus providing seamless integration between the two systems.

The connector is implemented as a native SugarCRM™ module and comes with a user guide that provides detailed installation and usage instructions.

We proudly present: The DriverMatrix™ SugarCRM™ Connector.

SugarCRM is a trademark of SugarCRM Inc. in the United States, the European Union and other countries.

In his post, Matt raises the question whether current day commercial Open Source ventures, SupportMatrix being one example he gives, are focusing on niche markets rather then CRM, ERP, ECM and the likes and worries that such niche markets are not big enough to sustain Open Source development leaving it prey to un-software service companies such as Google, Yahoo and friends. Matt also ties this to a recent drop in VC investments in Open Source companies.

Well, Matt, first of all, thanks for the mention. You are absolutely right - we are certainly not “… the Open Source X”. There is no such X quite yet in the proprietary world. We’re rather proud of that, but it’s quite independent of the really interesting question - how big is the market for what we’re doing, that is, software that helps making big support center manageable and technical support better?

The truth, and please don’t ever mention that we said this to our investors, is that we haven’t got the foggiest idea.

Oh, we are well rehearsed in reciting the numbers in the harsh light of the slide projector: “… there are 100 big support centers per region, we have 4 regions (EMEA, North America, Australia and Far East), at an average 500 operators per big support center and a price tag of 5k$ per user we get a total market cap of One… Billion… Dollars! ” [insert little finger in side of mouth here for full effect].

But as old Field Marshall Helmuth Carl Bernard von Moltke used to say: “No battle plan ever survives contact with the enemy .” and this is true today or marketing plans of start ups just as it was true for the Prussian army back in the old Field Marshall days. We’ll find out how big the market is once we manage to sell enough of our stuff to actually have enough meaningful dots in the graph. Until then, it’s all hearsay and wishful thinking.

You see, because we have no idea how big the market size really is or how much money we’re going to make, it makes perfect sense to try and throw as little money and resources on the problem to start with and the best way for a software start up to start small and remain frugal is is to use Open Source software. That way, if it turns out the market really is smaller then we thought, no harm done. We’ll just move along to something else.

If we do succeed however, we’re completely based on Open Source software and committed to the Open Source movement for good, and through the magical pixie dust of the GPL, even if we get bought out, this will not change in a major way.

In a way, Open Source works like an angel, except it doesn’t invest money - it invests IP. If the investment went sour, no big deal. But if it went well, we’ll pay back, with interest and a premium, IP to feed our Open Source angel next investment.

So don’t worry Matt. Open Source is fine and plays Y-Combinator with the coolest startups of the world. Losing doesn’t cost it a penny, and scoring produces hordes of paid developers to maintain it’s riches. World domination is proceeding according to schedule.

RAMAT GAN, Israel, November, 27, 2007 – With too many new software releases and gadgets on their hands, technical support centers around the world are having a hard time coping with the ever increasing complexity of current day computing environments. A new open Source software called SupportMatrix™, which might be just be what the doctor ordered for those support center growth pains, was announced today by Codefidence ltd., an expert in Open Source commercialization.

Walk into the operations room of any big technical support center and you will find a large plaque on the wall with the listing of all supported hardware and software combinations. An HP LaserJet 6000 printer on English Vista? Canon XH H1 camcorder attached to Windows 2000 SP3 Japanese? Each of these is a single combination out of hundreds which technical support center operators need to become knowledgeable about. Lately, however, there are just too many of them.

Solutions to difficult problems crop up in the darnedest of places, it seems. New software called SupportMatrix™ and released under an Open Source license by Israeli Open Source commercialization specialist Codefidence Ltd., may prove to be the answer to the technical support industry scalability pains: SupportMatrix™ uses virtualization to let technical support center staff interact with a simulation of the customer’s computer, right down to the operating system version, language and peripheral devices.

“SupportMatrix™ is innovative because there is no software that needs to be running on the customer machine and no connectivity is required. So an ISP, for example, can use it to solve customer connection problems, unlike traditional remote control software. ”, tells Gilad Ben-Yossef, Codefidence CEO and co-founder. “In fact, half of our current user base support customers via email. And since it is based on virtualization technology, teaching the system about new software is an easy as installing that new software on a regular PC.”

Released as Open Source software, the first version of SupportMatrix™ is available as a free download on the web site www.support-matrix.com, but Ben-Yossef does not expect many support center operators to download it and put it to use by themselves: “The IT industry have grown up: It’s not about the bits and bytes anymore, it’s about servicing customer needs. SupportMatrix™ is enterprise software. Our customers need guarantees and SLAs, they want it to talk with their CRM systems and there must be someone to call if there are issues. IT departments can download the bits and bytes from the web site for evaluation, maybe run a pilot, but for production use they will want to get a support contract with us. It’s the way the enterprise IT market works now. We’ve grown up. ”

SupportMatrix™ is already being used in support centers in Asia and is immediately available for purchase via resellers in EMEA, with additional regions planned to be added during Q1 2008.

Codefidence Ltd. is a privately held company specializing in productizing of Open Source components. The company was founded in 2003 as a response to the growing need of OEMs which make use of Open Source components as part of their products and services for a independent and professional support organization.

###

For more information, please check out our web site at http://www.support-matrix.com/ or contact:

They didn’t do it because the boats were not valuable any more. In fact, when landing on enemy shore boats become more valuable, not less. But the value of these boats where in their ability to carry you back home, even if you didn’t manage to slit the throats of all those other people.

That value proved to a bit too distracting. Better burn the boats and make sure all the men understand that there really isn’t any way to turn back, unless it’s on the broken backs of the enemy soldiers.

Releasing your commercially developed software, the one you spent the last two years designing, writing and debugging and you really really hope to be able to generate revenue from again, as Open Source software, works in almost the same way.

We destroy the value of the bits and bytes, the precious intellectual property so dear to companies of old, venture capitalists and lawyers. We make a big mound of it and torch it all up.

Because the value in the software itself proves to be too distracting. When you’re proprietary you can kid yourselves that all you have to do is be better then the competitors in your niche. You can lull yourself to believe that the customers will buy because they don’t have a choice. You drink the KoolAid.

But the IT industry have grown up. It doesn’t work this way any more. If you want to succeed, you need to understand that bits and bytes don’t matter any more. Only customer satisfaction does. Competition does not matter any more. Knocking your customer socks off their feet with amazing service does. And you have to do it again and again and again.

When you know that you have to build your business so that your customers will pay even though you are giving the bits and bytes for free, when your worst competition comes not from some other company down the road but your own creativity and genius and sweat and blood and tears, when you are faced with all that - you realize you have no other choice: excellence is the only option and so succeed you will or die trying.

Open Source does that. I’m not the first to realize it, I know. Matt Assay from Alfresco said it before me. That is how and why RedHat wins CIO Vendor Value awards two years in a row, way ahead much bigger and stronger companies.

But you know what? it’s something else entirely when you can smell those burning ships and hear the locals shouting.

Burn, baby burn. We launch in two days.

Gby.

Burning ship picture courtesy of chatigirl, the Creative Commons and flickr.

Before the summer began I took care to have good connection in both ends of the pipe: The Israeli office (upgraded to 5M/0.5M which was the maximal ADSL in Israel then - thanks Limor), and the Austrian alps office (upgraded to unlimited 3M/384k - thanks Martin). That was already enough to overcome the distance and feel like next door. OpenVPN eliminated the fact that I was in a foreign network, and I just had the whole office, including the printer and the phone system (thanks to asterisk) at my fingertips.

This luxury followed me also to internet cafe (where I met other nice people working with django while drinking good coffee), when I was away from the office. The laptop connected via wireless, activated the VPN and using a cross cable, a VoIP phone just connected normally with SIP over VPN to the office phone system. I had the local extension, so I could call using 3 digits number to all the other colleagues, and this way managed the project like I was just there. Of course I was reachable also via the Israeli office number for all those left behind in the boiling country, so some people don’t even know that I was away.

Being far had absolutely no disadvantage. Even when a need for technical meeting popped, we just did it over the phone (the VoIP one), while sharing the screen (using  GNU screen). This way, all parties could see the screen and type on it at the same time. Such a virtual meeting was also held with non-Codefidence people, like with Meir Kriheli which consulted us regarding some Django issues that popped up, and they got solved over the phone (well, 4 hours phone call). Thanks Meir.

Whenever I needed a break, instead of going to drink a coffee (what I usually do in the Israeli office), I took a short walk breathing mountains air and watching the scene around, drinking fresh alps water, and practicing my German. This combination generated a far better code, with much higher efficiency.

A challenge was to help test the system with massive amount of sessions as part of the development. This could not have been done directly over the VPN, because of the bandwidth limits. For this we used many virtual machines with many sessions each, and they were controlled using a single vnc session at a time. It was encapsulating multiple vnc sessions within a single one, with time slot modulation. Each time few such virtual machine were active in the foreground, while others were doing their job in the background. While we slept, automatic test scripts (written using xmacro) were on shift, so the 8 CPU cores machine with 16 GB memory were almost constantly busy. There are better ways to do it, but this one did the job simply enough.

The bottom line, it worked and we delivered the beta on schedule. This coding environment is highly recommended.

In order to install the beta version which was adapted to this specific OEM, I flew on late September to Japan, and worked together with our reseller there.

It was the first time for me in Japan, and the cultural difference gives the feeling that it is simply a different planet. Working all day in a suit and a tie is something I not used to. Doing that with slippers was definitely a surprise, although a good one.

Understanding the wonders of computerized toilet were amusing, but this probably happens to every Western visiting Japan.

The first install was at the resellers office, and it went smooth, as expected. We were a little more fearful for the second installation at the beta customer site, where issues always pop up. I was prepared to deal with for major issues with heroic efforts, but boringly, everything seemed to pass smoothly, both during installation and during the training. We must have did something really good in a previous life.

To succeed with an Open Source product, we have a policy of “Promise less, Deliver more”, and we stuck to our rules and delivered much more then the agreement promised.

This was possible thanks to the use of well tested open source components which continue to evolve and give sophisticated ways of building wonderful software out of them:

Here I must mention the Django framework which includes an amazing user interface - smart people had already thought of everything - ready to serve whichever application needed. Django web site describes itself as “High-level Python framework that encourages rapid development and clean, pragmatic design”. DriverMatrix interface is based on this framework, and I just can’t agree more.

A minor issue that appeared due to the specific structure of the customers network was taken care of by modifing the virtual machine to use a local squid proxy server. This will probably also be useful in the future with bizarre network setups.

New “last minute” features requests which we got from the customer were implemented during a short 2 weeks cycle from beta to release candidate. The flexible architecture just made it possible also at this critical stage.

An urgent request of smooth upgrade between unsupported version to the release, which was not included in the initial requests was supplied within hours. We are very proud that our fix cycle including QA can be sometimes as short as 48 hours.

The customer was very pleased, and as a way of gratitude, actually delivered the payment one month ahead of time. I be you don’t see much of that with proprietary software beta.

So the bottom line is, Open Source code base proved itself. We also were probably lucky

Of course, Open Source alone is not enough - many thanks are due to our reseller team for translating and bridging the language as well as the cultural difference, making this installation successful, supporting the customer with routine work, and of course giving me good time in Japan.

• Adobe PDF format (supportmatrix.pdf)
• OpenOffice.org ODP format (supportmatrix.odp)
• Macromedia Flash format (supportmatrix.swf)
• Microsoft PPT format (supportmatrix.ppt)

have a great day.

The world awaits with held breath and clenched teeth to the arrival of this new magnificent wonder: help desk for the 21st century.