FME Freemium Edition - To Free or Not to Free?
One of the ideas that we have been throwing around at Safe these days is the concept of releasing a “free” version of FME. As you can imagine this idea creates a very passionate discussion from both those for and those against such a product offering and I invite you to take part in this discussion by posting your comments below.
What got us talking about this concept at Safe is the book called “Free: The Future of a Radical Price” by Chris Anderson. This book is an interesting and thought provoking read and I highly recommend it. What is also brilliant about this book is that Anderson illustrates the very point that he made in the book by allowing people to read it online for a period of time (the MP3 version is still available for free from Wired.com). If you had wanted the “value added” ability to download it or to have your own copy of the book (physical or for the Kindle) you must pay for it. Ingenious!
Throughout his book, Anderson talks about the freemium model which is attributed to venture capitalist Fred Wilson with the name “freemium” being coined by Jarid Lukin. Looking around the landscape we see examples of freemium all around us. Many of the most popular sites and tools that people use everyday are based on the freemium model. Examples include Gmail, AVG, Hotmail, flickr, Skype, and one of my recent favorites Dropbox.
Available products using the freemium model are also more and more common. QuickTime is one that immediately comes to mind for me as you can upgrade from QuickTime Player to QuickTime Pro. In the GIS space the most recognizable would be Google Earth. One look at AppStores for the iPhone/iPod and other devices and you will see that there are literally thousands of examples of the freemium business model in action.
Another interesting thing about the freemium model is that it cuts across both open source and traditional software business models. The idea with freemium is that you are giving something away in order to sell value added services or enhanced products or services. If you just build things and give them away with no goal of selling future services or products then that is not the freemium model, that is just giving stuff away for free.
Freemium is largely a digital phenomenon as the unit cost of software and services is very close to zero making it possible to give stuff away in the hopes of getting revenue from users wanting more (check out the Economics of Abundance video on the freemium blog for a good explanation).
As we explore the possibility of an FME Freemium Edition we of course need to be careful. How much do we give away? If too much is given away, then users will not see the value of enhanced versions and we hurt our ability to fund further product development. If too little is given away, then there is no value in the freemium model since few people will use it. Hitting it just right means your product usage spreads quickly and the number of users “upgrading” to the purchased version outnumbers those that no longer need to buy the existing versions.
I think Lakhvir who is our Director of Marketing said it best when she said “This is interesting but we better make sure we do our homework and get our ducks in a row if we do this!”
My question for you is, should we or should we not have an FME Freemium Edition? I’d love to hear your thoughts.
It’s an interesting thought.
Do you have any idea how you would approach the feature limitations? Number of features transformed? Formats? Transformers?
I’ve often thought that it would be cool if you could provide a zero- or low-cost FME for Open Source which only writes to SQLite, PostGIS, MySQL, etc
You’d probably have to add SHP too though, and maybe enforce a feature count limit.
I’m guessing that you’d have to have a special build for this too to avoid having to pay licensing costs for flexlm and some third party libs?
Definitely a lot of cost/benefit thought required…
The FME Fan-girl in me initially thought “that would be awesome!”
After mulling it over I still think it would be awesome, though the question does become how do you limit it. Jason’s FME for Open Source is a method I like. I would take it further and say look at what do people working with all the increasingly available open data need. Formats such as SQLite/PostGIS/MySQL as an end point, but what data are they importing into those formats? I would suggest SHP and KML as a starting point. FME would then become an option for a lot of the open data hacking currently happening.
An additional format I would tack on would be OSM. This would allow more widespread usage of OpenStreetMap data in some projects as well as make it easier for people to put data into OpenStreetMap. Exposing the community to FME could open additional markets as well for Safe.
Don:
I think a Freemium edition would be of great benefit for a couple of reasons.
First, you’re not just selling software, you’re selling a workflow–transforming data using Workbench. Giving users a subset of functionality for free gives them an incentive to become familiar with doing transformations the FME way. In particular, being able to reach junior GISers with these powerful tools can open their eyes to what can be done without being an OOP coding guru.
Second, a freemium version could be a powerful gateway drug for FME Server in the Cloud. People have a hard enough time understanding the distinction between “desktop” and “server”, and “the cloud” is proving to be even more conceptually elusive. Freemium FME will allow you guys to demonstrate that the leap from modest jobs on the desktop to big jobs in the cloud is not as scary as many would suppose.
Finally, I think for SAFE the likelihood of cannibalizing existing sales is slim. Your customer base seems to have significant enthusiasm for your products. But there is a much larger market of folks who could benefit substantially from FME if they had a firmer grasp on how the pieces and parts fit together.
I see only upside.
Brian
Don- Great book… I listened to it last summer.
I believe there are probably several ‘free things’ Safe could provide. I would suggest you consider providing the ‘free’ items in beta (more like alpha for Safe) web services (FME Server). Remember, bits and bytes are much more susceptible to being ‘free’.
There are probably some other models such as pay as you use, and making the cost marginal. Keep in mind, it doesn’t actually have to be free, the marginal cost just has to be low enough it “feels like free”. Itunes and the App store are proving this everyday with .99 items for purchase. It’s not a dollar, its .99
thanks for thinking about this…
Thanks for all the comments. We have discussed internally and in fact have already extended our fixed licensing technology to support many ways of restricting the FME usage. We can control precisely the formats and whether it is reading or writing. We can also control as Jason eluded to controlling the number of features that we translate in a single run. Transformers are also things we can control as is locale. We have many ways switches. Maybe even too many as it makes the decision harder.
The question as Kate says is really “how we limit it” so that it is usable and generates excitement so more get to work and play with workbench and FME while at the same time positively impacting our ability to fund and push the technology further, faster.
Learon mentions “pay as you go” and that is also something that we have been playing with in Desktop. Amazon’s DevPay (http://aws.amazon.com/devpay) makes it really easy to create an AMI with FME Desktop in it which users can use to author workbench scripts thereby only paying for what they use. Our work with WeoGeo is also something that is going to come out on their site once FME 2010 Server is released.
Exciting times. What is the minimal set of things that an FME Freemium edition should have in terms of formats and transformers? Should it have a feature count limit?
Brian,
I like the point you make of having an FME Freemium edition be able to play the role of a stepping stone from Desktop to Web/Cloud with a Freemium Edition giving users the ability to run smallish things on the desktop. If a Freemium edition also gave users the ability to push larger jobs to the Server that could illustrate nicely the power of FME Server. Heck if the Server was in the cloud using some “pay as you go” then this fits nicely into what Learon mentions.
To be honest we had not thought of that connection, but with the transition of technology to the Web or Cloud this is another reason to explore an Freemium Edition.
Well, one thing I’d like to see is a free way to read esri file geodatabase and convert it to something else. Your product would then become well known and likely downloaded by anybody who needs to read a file geodatabase and does not have ArcMap or ArcCatalog. I’d vote free read and write opensource formats, in addition to reading the gdb, and that perhaps is an incentive to purchase the product for enterprise usage as the users (not the enterprise) use it some, get familiar, recommend to the people in management, etc.
Hi all
Good job and nice that you think about alternative ways to bring FME service.
I think FME can have more benefit by introducing FME Freemium Edition.
For example: FME Freemium can read and write to all open source formats, but users needs to pay additionally to get special writer or feature for Enterprise Commercial Formats e.g. for Oracle or Phtoshops PSD.
It will be user friendly licensing.
You don’t buy Photoshop just for resizing JPG’s.
Of course it depends of your Clients spectrum, but sometimes it is easier to sell LEGO-parts instead of whole big LEGO-bundle.
A have another Question/Idea what about Hosted FME Services for Premium features?
For example like this:
FME Freemium edition is limited to some feature to use Premium features users need to buy an account for Premium features and they can publish/deploy their transformation on shared service and make transformation on FME Server and download results from server.
Greets
First of all I’ve rarely seen a product that is worth its price as FME is worth it!
Anyway I like the idea of thinking about a Freemium Edition. But I agree that the major problem would be to find the right tailoring for a Freemium Edition and the risk is high to give too much for free. Remember that most user only use commen formats and 5-10% of the functionality. A second thought I have is that because of the wide variety of functionality and the heterogeneous user community everybody will prefer different functions and formats in a Freemium Edition.
Conclusion is that I have to read that book first to form a final opinion!
Gev,
The idea of somehow using a Freemium edition as a cross to services is a good one and one. There will definitely be some sort of announcement coming in the new year about FME Server and the cloud. See previous information on the Safe/WeoGeo announcement earlier this year. We are also working at Safe to improve our product delivery mechanism. Our fixed licensing technology definitely would enable us to deliver “premium” FME features to the desktop user of Freemium very quickly without a reinstall or things like that. You will also see from FME 2010 Desktop that we have worked to make integration with FMEPedia much better as FMEPedia will be running an FME Server repository so that Desktop users will be able to download samples and take advantage of our user community contributions much more easily. Ultimately we want to make FME Desktop be more of an internet application (not browser application in the short term) and leverage in many different ways the benefits of a connected world. FME 2010 starts that trip but there is more work to be done.
Christian and Mike hit on the big question. While with FME 2010 we ship about 254 formats, with additional available from partners; the question of which to have in any FME Freemium edition is the krux of the issue as the vast majority of users of FME care only about a small subset of them.
Any freemium release would need to include popular formats but entice users to want a full version. With regards to the book it is definitely worth a read. Heck, it is definitely work the “value added” version which enables you to fold pages, tear them out, and have a field day with your highlighter. All things you can’t do with the free one.
With a product such as FME that is utilised by such a broad range of users; pitching ‘what’ is to be offered as free is indeed the challenge.
Like Kate, I too initially gave it the thumbs up - but on reflection where would you start? From my perspective I completely agree with Christian, it is rare to stumble across a software product that is worthy of the investment that it commands. FME is more than worthy, reflected by the frequent comments of surprise that I receive when conducting FME Pre Sales and we get on to discussing the bottom-line…
Safe supported by their strong reseller channel have internationally carved a niche with FME, so that in many arenas it is now used as the defacto Spatial ETL tool. In my opinion a Freemium Edition should help to expand the vertical markets that FME is exposed in and enhance the brand (Lakhvir will be happy!). For example the Universal Translator has been available in a cut down form to MapInfo Professional users for years. Most MapInfo training delegates I meet know of it and use it primarily to move data between ESRI and CAD formats, though virtually none know of its provenance. Whilst a Freemium FME Desktop would be an interesting proposition, offering a cut down version (however this is done) of FME Workbench in my opinion is giving away too much value. I see Workbench as a users ‘Graduation’ into data manipulation and perhaps considering Workbench and its plethora of capability as being the up sell from a Freemium Universal Translator and/or Universal Viewer (Data Inspector) might be an alternative approach? After all many other vendors have gone down this route with products such as MapInfo Professional ProViewer and Autodesk DWG TrueView that become light and ‘well known’ utilities, demonstrating the art of the possible.
However it is done though a Freemium version of FME will likely gain ground in communities where software installation is not constrained. In the corporate environment ‘Free’ does not mean ‘Free’. It is regularly the case that I come across corporates that are charged not inconsiderable amounts by their outsourced IT to package, test and install so called Free software which ultimately means they cannot be adopted, better then to go through these political hoops for a full version. That said a move to the cloud as you suggest Don would circumnavigate some of these irritating problems, as does the ‘FME as a Service’ culture that we are fostering at Dotted Eyes.
An interesting debate but we certainly need to get the ducks in a row!
I’m a bit late to the party here, but a ‘Freemium’ edition strikes me as a great idea. I would however take a slightly different approach.
I think Brian’s comments above pointing out that a theoretical ‘Freemium’ edition of FME would be an excellent gateway to cloud based services are the real key. To that end why not simply make the free edition of FME the same the the base edition, but force all results to the visualizer. Basically the only output would be an FFS file. You might put a feature limit on it as well just to keep people honest. Then if a user wanted to test the workbench to the output of his/her choice they would publish it to a cloud based (WeoGeo hosted maybe?) FME Server and pay a nominal fee (based on the CPU time consumed by workbench?). For a very small fee users would be able to get the data they need in the format they want. The process would give the users the ability to see and use the product, understand how it works with their source data, try different transformers and test out the entire process via the visualizer… all for free. If the process was seamless and easy (think ITunes) it would be a impulse purchase and it would open up a whole new market of users who simply need a ‘one off’ solution. Not to mention driving a few people to the more advanced desktop editions.
I’ve been thinking about this post a while now and I might say it’s pretty hard to come to a conclusion. I’m using FME Pro right now and I can’t see any use for me in person to switch to a FME Freemium so I have to think in “third person” in some way. So what does the third person want?
Perhaps a FME Freemium version would have, for an example, support for reading and writing 5 formats and that the user could choose from a list of commonly used format when downloading the FME Freemium installation. The support for Open Source formats it’s a good idea and the support of SHP might be necessary. So why not have a totalt of 10-15 format to choose 5 from and then download “your own styled” FME Freemium? The same manner could perhaps be used for transformers?
Another thought is to release FME continously as the developers put together builds of a new FME release (as in the Beta case right now) and then once in a while release a stable FME release. The stable release costs money and you can get support on it from Safe. But the continious releases of newer and newer builds would be free to use. Then you can skip the Beta-releases and go straight to a FME Freemium build-release. And if you are a “safe player” and want a real version of FME when you’ll have to wait until the next stable official FME release.
The responses to this have been great and have given us much to think about and been very helpful. Cam’s comment above about the Freemium Edition only being used as a bridge to FME Server installations that are available on the Web (via AWS Cloud or via traditional server deployment) is a very interesting concept that is repeated in several of the posts. One could imagine a data delivery service such as WeoGeo where the data providers post not only their data to be shared but also the workspace(s) that defines how the data is to be shared.
Doing this the author could also provide “freemium” data that may be generalized but also provide more “value added” views of the data where clients then pay for this. Adding workspaces to the mix means that data authors have more options on how the data is delivered. Indeed an author could build a whole library of workspaces that all use the same uploaded data. Some providing “freemium” views and others not. Is this called “freemium squared?”
The strength of broadening the user base would be to gain development by the brilliant masses.
Imagine all the custom transformers that could be created by skillful developers who focus on PostGIS, OSM-data, SQLite etc. but currently do not have access to FME.
These custom transformers could, with the right distribution method, be added to all versions of FME (Freemium or not).
These transformers could be free, open-source or paid for in any way!
And what if “FME Freemium” were something similar to Adobe’s Acrobat Reader ?
It could look like a simplified version of the Universal Translator, allowing the user to run workspaces prepared by others.
There would be no limitations of formats and transformers, making this new tool really powerful, yet simple and making the user want more and more !
Wow, great discussion.
I think, Cam brings it to the point. Give the users full access to FME, but they can only “route to visualizer” and not actually output the data without having access to an FME Server or a paid version of FME Desktop.
Great. This could make a lot of sense. And as Ulf says, this is some kind of crowd-sourcing for custom transformers etc spreading the fme know-how amongst the hordes of enthusiasts out there and thus growing a critical mass of FME fans that would ask for more and more (paid) FME Desktop licenses or FME Server Services.
Now, this is exciting stuff!
Jeff
Thanks Ulf and jean-Luc. Again, this gives us a whole new set of things to think about. You two touch on the benefits that a larger community would provide to all users. In some sense we are talking about the community hitting some form of “critical mass” quicker resulting in a large enough eco-system that it grows much more quickly than it is now.
Currently we see users creating “Custom Transformers” on a day to day basis with some being posted to “fmepedia”. The majority of custom transformers are however not posted. While there are I am sure many reasons for this the current Custom Transformer model doesn’t support any form of commerce(freemium or otherwise). It only supports the “free” model. The “freemium” model would make it easier for authors to make available custom transformers which they feel are really valuable. Workbench is a Visual Development Environment as Dale described in an earlier post (http://blog.safe.com/2009/09/graphical-programming-environments-the-challenge-of-hitting-a-home-run/). The crux is we have users creating great stuff with no easy way for developer’s of custom transformers to do anything except give them away. Wouldn’t it be great if we created the tools to support a Custom Transformer “Freemium?” marketplace.
We do enable authors to password protect their custom transformers so that only the author is able to edit it. Of course anyone can run it with a valid FME license. This makes it hard for authors to promote their “value add” as there is no way for them to promote or increase visibility of their “value add” service.
What if we created the necessary tools so that users could provide their Custom Transformers using the “Freemium” model? Doing this would encourage people to create their own Custom Transformers that have a free version with a “value added” version for a fee. Of course they could also develop some that don’t have a free version.
To make this possible all we would need is create a “LicenseCheck” Transformer and a method for users to create their own fixed license files. The good news here is that all the building blocks are already in place and so there isn’t much we would need to do to make this happen. In fact, I have logged an enhancement request for this for FME 2011(which is under development now). Anyone out there think that they would use this new functionality to market Custom Transformers, and Custom Formats?
Jeff,
Thanks. Indeed it is exciting. This discussion has gone in several exciting directions. Initially we were just thinking of a simple “freemium” model and now we see how coming up with a good “freemium” model can act as the gateway to either desktop or server. The ability for users to build and control FME software components is also an exciting development.
Any other ideas out there that people want to throw our way?
This is a great discussion with many innovative ideas. It reminds me a bit of the project management trade-off story where you can have your project delivered with a quality product, on-time, or cheaply … pick two out of three. In this case, you can have formats, transformers, or output. The choice depends on your goals for the freemium product.
If you are interested in training users and looking for product ‘buy-in’, I would go for all the formats and transformers, while severely limited the output. For training and demonstration purposes, a user would want access to the full functionality. I would not be a fan of ‘redirect to visualizer’ only, as you walk away with nothing from the transformation. However, I would very much limit the output. In this scenario, you may be able to visualize everything, but only export the first X records. Access to the output is important because users need to test the results with other applications.
If you are interested in providing a complete, but restricted capability, I would go for a limited number of formats and transformers, with unlimited output. This would let some users fully solve a narrow set of problems using the freemium edition. The previous discussion on selecting and limiting formats has been very useful. This approach would provide a useful service to the larger community. The draw here would be that use of the product with additional data sources or functionality would require purchase of the product.
Good luck as you sort through this issue … I think it will be a valuable contribution to the broader geospatial community.