How will RapidMiner compete with free online platforms like Cognitive Class Labs?

mlg1988

Hello everyone,

I am exploring more data science platforms and tools.
I was using Python primarily so far, but I fell in love with RapidMiner now.

My question is in the title,
How will RapidMiner compete with free online platforms like Cognitive Class Labs?

The question is not saying anything negative about RapidMiner of course,
I think it is an amazing amazing tool.
I just wonder if free online platforms keep coming out,
how will RapidMiner stay in competition?
Which unique aspects does it offer?

Thank you

kypexin

Hi @mlg1988

I think company founder @IngoRM is the best person to explain this in much detail. But I found the question interesting, so just sharing my own thoughts here. 

I think that any free data science tool comes at the cost of NOT being capable of something at some point Think of things which make RapidMiner platform as great as it is:

• how mature it is, being developed by professionals for many years already
• how well it is maintained in terms of adding new features and support
• how many certain capabilities it has and how fast new ones are appearing
• how rich and complex the ecosystem it offers is, in terms of production use
• etc etc etc

I think in this sense RapidMiner and free platforms are just in different 'weight leagues' to be compared directly. As far as you are able to accomplish your tasks with free platforms and tools, it's fine. Once you need a bit higher level of complication or more capabilities, free platforms are left behind.

sgenzer

I agree with everything @kypexin says. I will add some more color if you're interested...

From my perspective (which has a serious bias ) RapidMiner continues to be ranked #1 year after year because it both innovates and executes. Every year. For six years. If you dig around, you will find many other free and freemium ML tools that have waxed and waned over the years. CCL is just one of many - you can check out Orange, Dataiku, Driverless ML (H2O), and many others including of course Watson Studio and other tools by our friends over at Big Blue. I had someone the other day bending my ear about DataRobot - saying that it was going to eat RapidMiner for lunch. It's possible of course...but my hunch is that their star will rise and fall like the rest of them.

I also think that RapidMiner has some key assets that people value both on the short and long-term. These are not "corporate secrets" - more like common sense to me:

• RapidMiner is transparent. You can build a model with operators or with Auto Model and peer inside as deep as you want. Its core is 100% open source Java, viewable by all on GitHub. Some people (who shall not be named) have taken umbrage with our #noblackboxes hashtag, but we're really serious about this. IMHO no company should ever deploy a mission-critical model unless they have complete confidence that they know what it's doing. Simply saying "trust us" is just not good enough.
  
• RapidMiner meets the needs of both 'coders' and 'clickers'. There is a clear bifurcation in folks using these tools these days, and you have to be able to meet both of their needs. At one end, you can click-click-click your way through AutoModel, and at the other you can insert Python, R, Java, shell scripts right into an operator. Or you can code your own operators. Sometimes I think RapidMiner is more like a layer on top of your operating system than software in how it accommodates the user where s/he is.
• RapidMiner is the #1 tool used at the university level to teach data science to the next generation of professionals. We're in over 4000 universities world wide. And we give it away for free to all students and professors. So guess what tool these students will want to use in 3-5 years when they are in a job? Yup.
• RapidMiner has been around for a while. People trust us to deploy a stable product. Our customers deploy these models in the field in mission-critical situations and know that there's a company behind it that has been there, done that. 

Don't get me wrong - I think it's fantastic to see other folks come out with new tools. I play with them when they come out just like many others do. And some do things much better than RapidMiner (image analysis immediately comes to mind). But after 15 min or so, I usually say "meh" and go back to RapidMiner. I'm sure this is as much about my own comfort, experience, and speed with the tool as it is a measure of the other tool. I know many really good, smart people who are just as emphatic about their tool of choice. This is a constantly evolving field and a wide open market that has room for lots of unique players.

So there you go...

Scott

IngoRM

I love it when that happens: as the founder, there is really nothing else to add from my side

To summarize all of this and add a different aspect to this at the same time:

The secret is that you build products people love.  And then act full of respect to those who use it.  Finally, you give people a reason to commercially engage with you.  Here are the main reasons:

• Need for more "power", i.e. rows or performance in our case.
• Need for productivity gains, i.e. Auto Model and Turbo Prep.
• Need for operationalization, i.e. RapidMiner Server.
• Need for collaboration and the support of complete teams, i.e. again RapidMiner Server.
  
• Need for advanced technology support, i.e. Radoop for pushing computations on the cluster while keeping the ease of use.
  
• Need for a deeper relationship with us, i.e. for support or having insights and the possibility to influence our roadmap.
  

It's that simple 

Hope this helps,
Ingo