Wednesday, 7 April 2010

ERP from the Cloud? Or just mature CRP?

I'm making a habit these days of writing huge comments to good blog posts, and then deciding to turn them into blog posts on my site
Lee Provoost's blog post on Spotify at Dachisgroup was what triggered me to write this lengthy comment, and I've just turned it into this post

Lee's gone back on a music trip to memory lane because he saw cassettes in the Callooh Callay bar's bathroom in London, and making a great comparison from CD
to iTune to Spotify - read the post at Dachis, there are some great points in it

Going back 20 years, the really bad thing was knowingly paying for numbers you didn't like - as of course you listened to the entire LP or CD before you bought it. Maybe you fastforwarded a few, but you listened to every single number before buying the record. Wow. Waste of time? No, actually having a good time and it simply was part of the music-buying experience

The thing I really like about MP3 is that I will just physically delete a song I don't like. Not because it saves space, but simply because I don't want to hear it again. Kick! Out of sight, out of heart

ERP from the Cloud, however? No. is proving my point, wrote a blog post about that a few months ago and rant about it on a regular basis on Twitter: it's simply not static enough to SaaS highly dynamic business, certainly not on that scale. is taking so very, very long for Chatter because they are trying to fix a moving target onto a static base. And I wouldn't be surprised at all if Marc Bernioff were to "redefine his ambitions" in the next few weeks. That on a sidenote
Cloud's certainly not the same as outsourcing, but when it comes to plugging in and out of the Cloud with great ease, it must all be similar stuff
You can't outsource manufacturing a car before you specified to death every single part of it, thereby freezing the design in its entirety and only allowing for a bit of bandwidth for the car parts to fluctuate within.

Back to my going back 20 years: paying for parts you don't like is what ERP comes with. You can't start to delete ERP-parts at will unless they've become absurdly cheap (compared to what we are accustomed to now) and completely detached from the monolith we're used to see them in.
They won't become that cheap unless they're "made in China". And they won't be "made in China" unless they've been specified to death and become boringly dull, static stuff. But when they have, we can greatly standardise them and UDDI will finally take off

So, answering Lee's question: primary processes are what make businesses tick. The businesses all support the same business, but just slightly differently so they can nick a few extra customers the competition can't get to. Ask a CEO what makes him or his company so special, and he'll start by saying: "what differentiates us from our competitors is..." - meaning that they all do the same business just slightly differently.
That's why ERP hardly pays off. And that's why Cloud will pay off for storage, virtual machines, networks and such: because the business just doesn't care about that at all. Like you don't care whether your car is manufactured in China, Poland, US or India. You just ordered that very car...
At a higher level, SaaS will service a standardised part of business applications. SaaS and service are the keywords there: it's small, fine-grained, and standardised. Unlike ERP, the big monolithic ball on Atlas' shoulders

I'm sure that System Integrators will loudly applaud the idea of ERP out of the Cloud as they now support SAP because there's great money to be made in implementing and maintaining it. To me that's totally defying the purpose of a package, but then again I haven't seen a package that wasn't customised to a great extend, so maybe that's just today's IT reality? Or business? Or are we all just spoiled rotten, and would we be equally happy when our options are limited to a wide few, like in the car industry?
So maybe we'll not get ERP from the Cloud, but just Cloud Resource Planning. And will that finally help to make the IT industry mature

3 reacties:

Patrick Brinksma said...

I do not have much insight on ERP in the cloud or not. But I do want to comment on the extensive amount of customizations when implementing packages:

The fact is that the 'core' or bulk processes are the same when looking at different companies. Even cross-industry. That is the reason why software companies can productize packaged software with some processes build-in.

BUT, it is rare that a customer is actually implementing a package, which requires a very different approach to building it yourself. It is amazing that after so many years organisations have not been able to do this. It has been done successfully on smaller scale, but when the implementation size increases, the likelihood of customization increases exponential.

And this has off course to do with the maturity of IT AND Business, but even more with the relationship between IT and Business. It is rare that they are throughout the project or program in a cooperative mode. And that has nothing to do with the software, but that's all about the Organisation and thus People!

Final note: Will IT ever mature? No. As much as I will never mature. Because you are never done! I hope people will have more fun with IT. Make it fun! It's fun to implement a new software product. It's fun to change. It's fun to make success!


Martijn Linssen said...

Thanks Patrick ;-)

You've had your fair share of package experience of course (...) so I highly appreciate your insights on this!

You are very right on the increase in customisation along with the scale of implementation: again, that's what EAI showed with the point-to-point connections decades ago:
All parties that come along will share a significant piece of business. But, they'll also share a significant non-piece of business

The first part is what we call a business rule. The second part is what we call a business exception. Yes, every party that joins the club introduces (a piece of) a business exception

Business rules we can standardise, outsource, Rightshore, Cloud and SaaS. Business exceptions we have to bespoke, tailor make, build ourselves

I think that's why the amount of customisation increases with the scale of implementation, and why Enterprise-wide package implementation takes years and millions

I like your take on maturity in general ;-) and you're right. Evolution means continued growth and permanent 'adolescence'. However, immaturity is a bad sign. Growing up too fast will make you immature (not making a comment about certain countries here), and there's a fine line there...

Patrick Brinksma said...

Well, there is also the way you implement business exceptions. But that requires proper thinking, common sense, but also creativity to a certain degree. Customizations are very often required, and not a bad thing. But in practice they become like small 'tumors' as they are not included in the overall solution architecture.

These tumors grow and grow and increase cost of maintenance and further releases (mainly impact analysis become expensive and very unreliable). And then after a couple of years organisations start to think about upgrades for many reasons (both Business and IT), and want to go back to 'out-of-the-box'.

I am not saying that all designs should be there with all detailed requirements included, as time does not permit it, and it's crazy. But, here is where you bring in senior people with experience and expertise who can design frameworks for implementing Business Exceptions and assigning in what domain certain functionality belongs.

Post a Comment

Thank you for sharing your thoughts! Copy your comment before signing in...