Presenting Open Source to the Enterprise - Rules of Engagement

Posted 11 Sep 2000 at 06:53 UTC by mechanix Share This

I thought I'd share 'Rules of Engagement' with everyone. Based on my experiences 'selling' Open Source the last few years, it's a simple bullet list of do's & dont's that might be useful to people who have to present Open Source Solutions to customers or management.

Over the course of the last few years, I've been had a number of jobs that involved 'selling' Open Source Software (O.S.S.) to enterprise customers. Along the way, I've compiled the following short list of tips for effective presentation:

Rules of Engagement for Open Source in the Enterprise

1. Provide a solution.
Not just a code base.

2. Open Source MUST compete on the basis of stability & feature set.
Companies generally want the best solution.

3. Initial Cost may not be a deciding factor.
Companies look at the T.C.O - not just purchase price.

4. Having Source code available is not necessarily a benefit.
Code is only a benefit when used.

5. Open Standards can provide more safety then Open Source.
Open Standards allow the integration and migration of applications.

6. Not all problems are currently solvable by Open Source.
Don't create false expectations



1.) Provide a solution.

Most companies are not in the Technology Industry. To them, M.I.S. is an expense - not profit center. They continually weigh the benefits M.I.S. provides against the cost of supporting it. Program development/maintenance is a complex task with many uncertainties, and it is notoriously difficult to schedule and budget. On the other hand, shrink-wrapped solutions provide fixed cost and (relatively) fixed implementation schedules. As such, companies generally want solutions provided for them, not produced by them.

To be successful, you must provide a full solution. Remember that installation of a product is just the beginning of the lifecycle. Make sure that all applicable areas are addressed: training, documentation, support, etc. Explain how your solution will fit within the companies existing M.I.S. infrastructure. Remember that the goal of most M.I.S. systems is to automate workflow and make a company more efficient. Show how your solution will help accomplish this.



2.) Open Source MUST compete on the basis of stability & feature set.

In order for O.S.S. to gain acceptance in the Enterprise, it must provide at least the same level of stability and feature set of existing applications. Enterprises have a lot riding on their M.I.S. systems, and generally want the best solution available. While 'good enough' may work in informal settings, Enterprises won't be guinea pigs or accept jury-rigged solutions.

Only present O.S.S. projects that you know are stable and provide the needed functionality. Test the scalability of the components. Companies typically perform testing with a limited number of users, and expect to be able to 'throw the switch' and scale up to the full organization. Become familiar with each component, and what benefits it provides. Make sure the solution presented adheres to any relevant standards, allowing it to be easily replaced when the time comes.



3.) Initial Cost may not be a deciding factor.

Companies generally treat M.I.S. as a long-term investment. They weigh out the Total Cost of Ownership (T.C.O.) for a project over its lifetime (hardware, implementation cost, maintenance cost, etc.) vs. the perceived benefits (increased efficiency, additional revenue, etc.) Additionally, the initial purchase price of software is typically a fairly small portion of the T.C.O. This means that the other costs involved in O.S.S. solutions must keep the T.C.O. of 'free' software competitive.



4.) Having Source code available is not necessarily a benefit.

Code is only a benefit if it is used! Many companies would prefer to have someone else do the development/maintenance of their programs. This is why standard tools and shrink-wrapped software became popular in the first place. Indeed, one of the biggest complaints heard is the cost of maintaining legacy applications. If a particular vendor stops supporting a product, these firms are more likely to simply choose another product rather then attempt to maintain the current one. They would prefer to pay any conversion costs, rather than assume the open-ended expense of maintenance. Given that these firms are unlikely to use (or even look at) the source code, it is essentially of no value to them.

Even firms that keep a programming staff on hand to support applications (such as E.R.P. systems) would prefer NOT to. They are generally looking to lower the amount of staff M.I.S. requires. Be very careful not to make O.S.S. sound like a do-it-yourself project! Remember, hacking the code should be an option not a requirement!



5.) Open Standards can provide more safety than Open Source.

Let's face it...We live in a proprietary world, and it's not going to change anytime soon. With that in mind, Open Standards can potentially be more important to companies than O.S.S. Open Standards apply to proprietary solutions, as well as O.S.S solutions. This means that companies can potentially convert to any of a number of solutions, without losing their investment. Many companies are already familiar with Open Standards of one form or another, and realize the value they provide.

When presenting a solution, make sure you know what standards are relevant and how your solution supports them. This may include industry standards, ISO standards, R.F.C.'s, or, if all else fails, de facto standards. Keep current on standards that may affect your solution and provide a migration path for your customers.



6.) Not all problems are currently solvable by Open Source.

The O.S.S. market is still maturing, and there are some (sizable) gaps in the applications it provides. This is not a shortcoming, just a fact. Never attempt to shoehorn a customer into an ill-fitting solution simply because it is a pure O.S.S. solution. Always remember that providing the best solution to the customer is the best way to promote O.S.S.

In order to accomplish this, you must be familiar with as many packages in your problem domain as possible (besides, keeping an eye on the competition is just good business sense). If a particular feature keeps an O.S.S. project from being competitive, allocate resources to implement it. In short, provide the client the best solution possible, and leave the religious arguments to /.




Re: Presenting Open Source to the Enterprise - Rules of Engagement , posted 11 Sep 2000 at 07:10 UTC by uweo » (Journeyer)

<person> mechanix</person> wrote:

4. Having Source code available is not necessarily a benefit. Code is only a benefit when used.
Not quite. Source code is a benefit when it is needed. (Write) Access to the source code means that you can fix problems in it. This doesn't mean that you want or like to fix problems, it means that you can deal with them if case you need to. Source code provides an option. Not more, not less.

5.) Open Standards can provide more safety than Open Source.
Sure. And together they provide even more safety.

Open Standards not substitute for Open Source, posted 11 Sep 2000 at 08:05 UTC by atai » (Journeyer)

People often compare Open Standards with Free Software. This comparsion does not make sense. They are not substitutes for one another. Open Standards are good. But that's not the reason for not going Open Source.

If you are trying to sell oranges, whether bananas are good is not the issue.

Most custom systems are OS for the customer, posted 11 Sep 2000 at 12:37 UTC by Dacta » (Journeyer)

In my previous job(s), I worked as a Windows (Delphi) programmer, writing big systems (the kind of things often called "Enterprise Systems") for utlity companies and shipping firms. The customer always got the source to them, and in most cases owned it. While we were using non-OS products (Windows, Solaris & Oracle for instance), there was no way the company was going to buy a custom solution if they didn't get the source.

I'd be surprised if this was different anywhere else - in most of these cases, the system was partly designed and/or written in house. The infrastructure type of compnents (Database, etc) depended more of support & reputation, though.

Benefit of source code, posted 11 Sep 2000 at 13:37 UTC by PaulJohnson » (Journeyer)

4.) Having Source code available is not necessarily a benefit.

The point to emphasise here is that the source code is a fall-back position. Suppliers can go bust or stop supporting a product, and what you get out of a "source code escrow" agreement may be an unmaintainable pile of spagetti. With open source you have maintainable source from the word go.

Now you don't want to have to touch this, but with OSS you have an option to do so. This puts an upper limit on the possible costs which is very reasuring for enterprise IT types: it means never having to go to the board and say "they want $1e6 to keep it running".

Also, it means multiple sources for support and enhancements. If you want support and enhancement for closed software then there is exactly one supplier qualified to bid, and they will charge accordingly. If you want an enhancement for open source then you can ask for quotes on the open market.

In addition I would add another point:

7.) No stiffing.

Stiffing is the practice of locking in a customer and then demanding huge fees for upgrades. You want to move from a 2-CPU license to a 4-CPU license? No problem Mr Customer, that will be twice what you originally paid for the license. And you have to get our latest version because we don't support version 3 anymore. Pity about all your legacy applications that it will break.

Daily dosage of FUD..., posted 11 Sep 2000 at 16:56 UTC by dneighbors » (Master)

I am sure I will get flamed for the following response, but I feel this article needs commenting that isnt held back.

At first this article popped out to me and said finally someone talking about a Free Enterprise and taking the time to write about it. While I agree on the surface with some points the details are extremely distrubing to me.

In the spirit of GNU\Linux one fights for freedom, in this case it seems one has rolled over for the FUD of corporate America.

Most companies are not in the Technology Industry. To them, M.I.S. is an expense - not profit center.

While I whole-heartedly agree a solution needs to be provided not just buzzwords and hype of the open source community. This line scares me. Most companies are under the delusion IT should cost a FORTUNE. Because IBM, SUN and Microsoft have TAUGHT them that. You seem to be saying one should just concede this fact.

FREE software is not only about providing solutions its also about educating those that use it. Educate that them good software doesnt have to be expensive. Your IT department doesnt have to be the cesspool that eats company revenue!

In order for O.S.S. to gain acceptance in the Enterprise, it must provide at least the same level of stability and feature set of existing applications.

Enterprises have a lot riding on their M.I.S. systems, and generally want the best solution available. While 'good enough' may work in informal settings, Enterprises won't be guinea pigs or accept jury-rigged solutions.

Your kidding right? I mean its all about PERCEPTION. Can you honestly look me straight in the face and say Windows2000 isnt a kludge that should be really classified as beta warez. At least until service pack 3.

Just because you paid a lot for it and has 600 pages of features doesnt make it the best solution. I think every piece of software needs to handled on a case by case basis with honest and open review. Once again your IT department deserves to hear the whole story. Maybe free project A has 30% less features, but non of them are ones you cant live with out. Also, its 50% greater uptime and additional security features and no entry cost are worth the loss of un used features. You seem to make this cut and dry. Something closed source loves to do. You dont have feature A go home, and enterprises buy it. UNTIL you educate them.

Companies generally treat M.I.S. as a long-term investment. They weigh out the Total Cost of Ownership (T.C.O.) for a project over its lifetime (hardware, implementation cost, maintenance cost, etc.) vs. the perceived benefits (increased efficiency, additional revenue, etc.) Additionally, the initial purchase price of software is typically a fairly small portion of the T.C.O. This means that the other costs involved in O.S.S. solutions must keep the T.C.O. of 'free' software competitive.

Come on. TCO is a black voodoo art of PHB. Anyone can turn ANYTHING into a negative or positive on TCO. I believe this to be more FUD. I have yet to find open source stuff that has higher TCO than closed. Simply killing license costs gives a great advantage. As open source support is usually cheaper and maintainability a LOT more friendly.

Finally point being TCO should be easy to help you get OSS in Enterprises not keep you out of it!

Code is only a benefit if it is used!

As others have said this is not true. One main reason is contingency. Most large Enterprises pay unGodly sums of money for rights to source code in the event a vendor goes under etc.. The legal costs ramifications behind this are astronomical. Open source makes this a NON issue. No lawyers involved no worries about someone going under.

You think people owning Baan software that paid millions arent sweating that the company is in serious financial woe? I bet they wish they had the source about now.

et's face it...We live in a proprietary world, and it's not going to change anytime soon.

This has to be the worst lie of them all. I mean why bother using GNU\Linux after all Microsoft rules the world and its not changing anytime soon. I think this type of limited thinking is the only thing keeping OSS out of the Enterprise currently.

Open Standards are rarely that, usually they are controlled by vendors. As we all know no set of vendors can all get along so that usually means several sets of standards. Ever play with a standard called EDI. HAHAHAHA. Every business has their own flavor. New B2B exchanges will be NO different.

The thing OSS brings to the table is no matter what the "standard" of the day is you can adapt to it. Being locked into a standard is no worse than being locked into a vendor. Vendors know this and thats why they push for their own standards. : )

The O.S.S. market is still maturing, and there are some (sizable) gaps in the applications it provides. This is not a shortcoming, just a fact. Never attempt to shoehorn a customer into an ill-fitting solution simply because it is a pure O.S.S. solution. Always remember that providing the best solution to the customer is the best way to promote O.S.S.

Here I agree. Though I think all problems are solvable by OSS there are just some that havent been yet. : ) I often find it best when a solution is not available to get an Enterprise to start and OSS solution. It often proves cheaper still than buying out of box and they get the joy of controling the projects direction.

All in all I think you bring up some good points, but trying to adapt OSS to an Enterprise is not what our goal should be. We should be educating corporations on the new face of software development and getting them to embrace it instead of trying to mold it into a bastard child of development for them.

I have found in my own experience when I quit trying to pitch OSS software and started educating the suits on what FREE software brings to them. I had much more success as I was no longer fighting their misconceptions, but educating them on what freedom is. There was much skepticism at first, but once you come through or can openly dispell things its amazing what effect it has. Once someone has a change in perception there is no going back.

My advice is get your boss to trust you by doing good work. Then educate them on what they are missing out by not understanding they are in bondage in closed packages. When given the chance make use of the community and let them know the effects.

Bottom line is all businesses want to be successful prove to them that OpenSource/Free software can do that and do it cheaper and you will have a friend for life. :)

Linuxmanship, posted 11 Sep 2000 at 17:47 UTC by dmarti » (Master)

Here's my 1998 guide to selling Linux, much of which is still useful.

My experiences, posted 11 Sep 2000 at 18:28 UTC by emk » (Master)

1. Provide a solution.
Not just a code base.

Very few businesses want to buy software. On the other hand, many businesses want to buy a complete solution: software, support and training. They'll pay money to make things work.

As one of my clients told me, "That was an easy sell. I'm happy to pay for anything that looks like insurance."

4. Having Source code available is not necessarily a benefit.
Code is only a benefit when used.

Here are some rules I give my clients:

  1. Don't modify the source code unless you have a compelling business reason.
  2. Document your changes thoroughly, and keep them in CVS. Learn how to use vendor branches.
  3. Submit all bugfixes and enhancements to the package's maintainer. Track the status of your bug reports, and document all bug IDs. Any business benefit you may get from in-house features will probably be outweighed by ongoing integration costs.

Here's a good open source sales pitch:

By the way, we'll be providing source code for all our software. If you're ever disastisfied with our work, you have the legal right to change vendors. If you decide to work with someone else, you won't have to replace all your systems.

People like to hear this. ;-)

Flamebait - Save it for /., posted 11 Sep 2000 at 21:13 UTC by mechanix » (Journeyer)

I must admit I was a bit saddened by the tone of dneighbors reply to my post. Seems like flame wars follow wherever geeks go :-P While I found most of his technical faults to be poorly reasoned, each has his own view. However, characterizing a document as FUD simply because you disagree is childish.

Below, I attempt to clarify my stance and address the specific concerns raised. All further flames will go to /dev/null

dneighbors said:

While I whole-heartedly agree a solution needs to be provided not just buzzwords and hype of the open source community. This line scares me. Most companies are under the delusion IT should cost a FORTUNE. Because IBM, SUN and Microsoft have TAUGHT them that. You seem to be saying one should just concede this fact.

FREE software is not only about providing solutions its also about educating those that use it. Educate that them good software doesnt have to be expensive. Your IT department doesnt have to be the cesspool that eats company revenue!


expense was used as an accounting term, as for most companies, I.T. does not directly generate revenue. I have since checked with the biz. people here and found I should have used the term 'cost'..as in 'cost center'. At no point did I indicate how expensive I.T. should be. I.T. budgets need to be figured on a company by company basis. Grossly characterizing IT as a 'cesspool that eats revenue' does no one any good.



dneighbors said:

Your kidding right? I mean its all about PERCEPTION. Can you honestly look me straight in the face and say Windows2000 isnt a kludge that should be really classified as beta warez. At least until service pack 3.

No I can't. Can you honestly look me straight in the face and say that companies RUN on windows? This might be true of smaller, newer companines (like .com's), but in my experience, its not true of older, larger companies. Most large companies I've seen (brokerage, retail, clinical data, chip manufacture) generally do their backend processing on mini-computers or higher. They generally have a hodge podge of equipment from smaller machines like the S/36, S/38/, and AS/400 to full blown 3090's and Amdahl's. From what I've seen, the most common usage for windows machines is desktops and inter(intra)net servers.



dneighbors said:

Just because you paid a lot for it and has 600 pages of features doesnt make it the best solution. I think every piece of software needs to handled on a case by case basis with honest and open review. Once again your IT department deserves to hear the whole story. Maybe free project A has 30% less features, but non of them are ones you cant live with out. Also, its 50% greater uptime and additional security features and no entry cost are worth the loss of un used features. You seem to make this cut and dry. Something closed source loves to do. You dont have feature A go home, and enterprises buy it. UNTIL you educate them.



I never said being expensive made something the best solution. The point I was trying to make was to be sure that the solutions you present are solid and up to the task. I clarified this in the paragraph that followed the one that prompted your above response. Had you bothered to read the entire section before replying, this misunderstanding could probably have been avoided.



dneighbors said:

Come on. TCO is a black voodoo art of PHB. Anyone can turn ANYTHING into a negative or positive on TCO. I believe this to be more FUD. I have yet to find open source stuff that has higher TCO than closed. Simply killing license costs gives a great advantage. As open source support is usually cheaper and maintainability a LOT more friendly.

Finally point being TCO should be easy to help you get OSS in Enterprises not keep you out of it!



Regardless of whether you feel T.C.O. is FUD, those numbers are treated seriously by management. Furthermore, I never said the T.C.O. was higher for O.S.S. I simply tried to point out that while O.S.S. is 'free' initially, its T.C.O. must be competative with other solutions. The largest contributor to T.C.O. is generally services after installation. As such, the 'free' nature of O.S.S. do not give it the giant headstart you seem to indicate.

Finally, T.C.O. could be a big help in getting O.S.S. in the enterprise if there were more published figures. I'd welcome any serious studies & figures you care to forward regarding the T.C.O. of O.S.S. So far the only data we've found was from vendors with a vested interest in the outcome :-)

dneighbors said:

This has to be the worst lie of them all. I mean why bother using GNU\Linux after all Microsoft rules the world and its not changing anytime soon. I think this type of limited thinking is the only thing keeping OSS out of the Enterprise currently.



Excuse me ? This is not a lie...its a fact. If you feel otherwise, I suggest you try working in I.T. for a while. The attitude people seem to have lately that gloried pc's drive the enterprise is just soooo narcissistic. Big Iron (& medium sized iron) is still generally the norm. Hell, there is still more COBOL code in the world then anything else. Not to mention, Big Iron is STILL the most cost effective way to due data warehousing for extremely large data sets. As such, neither O.S.S. nor MS are going to rule anything other then the desktop for a long time to come.

As for limited thinking, your statement that MS rules the world is the best example of that I've seen lately.

all this OSS, posted 12 Sep 2000 at 01:16 UTC by jmg » (Master)

And the REAL problem in M.I.S. is the fact that most people working in M.I.S. have no real understanding the the problems that they are trying to solve. The people that do have a good understanding and are able to provide a solution are both to hard to find and to expensive to keep on salary.

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!

X
Share this page