NetSuite and SAP: a tale of three development modalities

Vinnie Mirchandani has been busting my balls the last year or so, complaining that SAP and other large enterprise players are no longer the real power houses they were. His book, The New Technology Elite makes the point that:

...if you analyze Apple, Google, Amazon, Facebook, Twitter and eBay, you marvel at their data centers, retail stores, application ecosystems, global supply chains, design shops. They are considered "consumer" tech but have better technology at larger scale than most enterprises. The old delineation of technology buyer and vendor is obsolete. There is a new definition for the technology elite - and you find them across industries and geographies.

In his latest post, Vinnie says:

For years, SapphireNow has been about new product announcements – HANA, ByD, SUP etc etc. Years later, SAP can barely show a smattering of customers or partner applications on them.  I had vigorous discussions in Orlando with my friends David Dobrin and Dennis Howlett who think I am unfair to SAP and its partners. I wish they could have seen some of the things at SuiteWorld.

Vigorous doesn't come close, more like verbal fisticuffs.

What 'things' does Vinnie mean? There were no NetSuite press releases talking about product but Frank Scavo says the company is working on a manufacturing module while Phil Wainewright mentions multi-book accounting. Are these supposed to get me excited? I guess if you're a NetSuite fanboy then sure but if you live in the enterprise customer space the fact NetSuite has not had these mods would exclude them from many deals. So now I am thinking this is a net good because NetSuite gets to play in more markets. In back channels I questioned Frank's qualified enthusiasm for a manufacturing module. He says:

NetSuite is making the right decision and good progress to build out manufacturing functionality as part of its core system, but there is still much work to be done to achieve functional parity with other cloud and on-premises solutions in the marketplace. Nevertheless, the rapid development capabilities of NetSuite's platform offer hope that it will get there quickly.

Plex Systems is pretty much the only other cloud manufacturing suite player in town so having NetSuite dip its toe in the water is good for the market generally. But Plex is barely a $50million business. As I asked Frank: what does NetSuite think it can do that Plex can't to build a much larger cloud business for the manufacturing industry?

To Phil's point, multi-book accounting has been around for many years in the on premise world. It's pretty much a pre-requisite for international organisations.

I'm sure Zach Nelson will agree that suite plays always win. He's said so on many occasions. When you view the world through that lens then other things come into rapid focus. As Frank notes:

As explained by Zach Nelson in a small group briefing, the customer order is central entity in NetSuite, and NetSuite wants to “own” anything that is input into, or output from, the customer order. In the manufacturing sector, this would include production work orders. It makes no sense, therefore, for NetSuite to hand off these business processes to partners.

I'm sure he'll also agree that these things take time. It's a point SAP knows only too well. Yet Lars Dalgaard seemed to be saying the exact opposite last week when he talked about functional plays coming out of Business ByDesign. And then that moment of clarity struck me.

We get excited by what's coming next in the cloud world and love that customers are so enthusiastic about what vendors are doing. Why is it then that Vinnie can chide established players? A good clue comes in his argument that SAP should be less about vision:

For  a change I would love to see SAP NOT announce something, and instead go sign up 500-1000 customers to early test a product, and quietly strive for some significant metric like 15,000 to 20,000 partner apps and then release them at SapphireNow.

I hesitate to say but Vinnie's application of these kinds of metric are unfair without qualification as to why these numbers are significant. On the other hand, cloud players rarely talk about roadmaps in the public domain. If you look carefully at Vinnie's post illustration, you might notice that one request is for a roadmap portlet (bottom right.) It is a point I often find frustrating.

Instead, vendors release new functionality on a two/three/four times a year cycle with hints about what's coming next at the time of a release. The multi-tenant model means they must have battle hardened code running at the point of release. So on the one hand we can go 'wow' x-times a year but on the other hand, cloud players have to be very careful not to release too much code in case they code in multiple break points that become unmanageable. For example Workday knows it needs to do a lot more to make its reporting functionality up to snuff. That won't happen in a New York minute. Last fall, Stan Swete, Workday's CTO reckoned there's a year's work in front of the company on this topic but crucially, the company will not name a date for release until they know they've solved the problems.

The on-premise world is very different. In the on-prem world customers have gotten used to vision and roadmap as a way of planning where they will be in a year or so's time. Large enterprise needs that breathing room because otherwise their landscapes becomes chaotic and unmanageable. For example, Business Warehouse on SAP HANA has been announced but those I speak with believe it will be 2013/14 before companies are ready to deploy in large numbers. This would be unthinkable in the current cloud world. Yet in the on-prem world you have a degree of flexibility the cloud world does not provide unless you are prepared to do what the on-premise buyer does: code out functionality you need. In Frank's piece:

...the ability to customize and extend the solution should not be taken as an excuse for not offering expected features/functions in the standard product. When prospects need full-blown work center capacity planning, for example, the last thing they want to hear is, “Oh, we can use SuiteCloud to build whatever you need.” For customer-unique requirements, SuiteCloud is a powerful attraction. For what should be standard functionality, no

SuiteCloud can be likened to ABAP where anything is possible if you want to code it. But in the multi-tenant cloud world, the cloud player ends up having to support all the custom code. In the case of Salesforce.com, I'm led to believe there are some 200 million lines of custom code kicking around. A nightmare in waiting if you need to refactor the platform?

And then we have HANA. While Vinnie's point about handfuls of applications is well taken, the fact is that the HANA of today is very different from the in-memory column store database that Hasso Plattner wowed us all with three plus years ago.

The HANA of today is a development environment where anything is possible but also where the complexity of application development is largely removed. That is because SAP is designing HANA as a next generation application and database server that has many of the administrative functions you normally need, baked directly into the platform. In turn that means developers can concentrate on two things: UX and business logic. That's it. As Plattner said during the day three Q&A: "Developers are coming to us with product they can build in days and weeks." Anyone heard that before in the SAP world? Is that something to get excited about? On the evidence I saw from startups that SAP is supporting, I'd say 'yes' and go further.

I'd argue this puts business process invention back into the hands of customers in a (nearly) unconstrained way. If that also means fast time to market then surely many of Vinnie's concerns go away. Perhaps we really can look forward to a day when SAPPHIRE Now is not just a flogathon but a parade of customers talking about the extraordinary applications they built. Add into the mix SAP's move to make HANA development licensing free in perpetuity and you really do have to be a died in the wool curmudgeon to put that down as a 'so what?' In my case it is more like 'thank goodness.'

Is HANA mature enough to deliver on that lofty vision? No. It is a work in progress with functions like disaster recovery a 'must have' piece that's missing from the equation in analytic applications. Is the business model for partners in place? Sort of. Negotiations around what shape the model needs to take are ongoing but my understanding is that it will be worked out very shortly in a way that puts smiles on everyone's face. Does this mean customers will have an 'on par' development platform that competes well with cloud approaches? We don't know and wont know until we see many more of the kinds of HANA application I saw from IBM that made me go 'wow.' Or when we see HANA running in the cloud.

Does HANA provide the bridge between the on-premise world and cloud such that the Vinnie's of this world are satisfied? I don't know. On the other hand it is a testament to Vinnie's carping about pricing and business models that SAP is at last flexing what seemed to be inflexible internal business rules. That alone should be enough to make those worldly wise folk think again.

In the meantime I want Vinnie to keep pushing back. It gives all of us reason to think and rethink what the heck we mean in the world of enterprise applications. It makes the SAP's of this world adjust to a different cadence in development for advanced applications. That's all to the good.