The Enterprise: I’m Not Sexy And I Know It(techcrunch.com) |
The Enterprise: I’m Not Sexy And I Know It(techcrunch.com) |
The CEO talks about wanting scalable cloud solutions using hypervisor technology without block-based attached storage.
The people using the system are treating it as Excel with a slightly crappier UI.
In most large organizations, the problem isn't providing a way to let their large amount of data scale. The problem is getting people to even input that data into a useful format so that it can be analyzed at all. You will spend man-months designing a system that handles all edge cases only to find out that they can't or won't even enter any useful configuration and support data to make it usable.
Everyone wants to talk about how your resource allocation and planning algorithms work with mobile technician work forces. No one can answer whether these fancy systems actually save costs or increase efficiency. And they don't care either.
We sell software to enterprises. It is a kafka-esque nightmare.
Dear Businesses, it is not your fancy code & tech specs that make you special, it is your business data. Pay attention to your data and the people who create value with it.
I'm always disappointed by enterprise customers who can talk for days about their endlessly complex systems and their fantastic dreams of building and customizing exotic new & infinitely flexible applications, and yet these same people are completely unable to answer the most basic questions about how many, how often, and how value is created. When pressed, they call some poor intern to wrestle with SAP for a few weeks in attempt to answer what should be routine business questions.
Then come the horrors of talking to the people who are forced to use these systems and witnessing the heroics required to get them to work for the business. And that's just the folks who haven't given up entirely and instead spend their days with a list in Excel.
Some folks are more creative than that. Crappy UI? No problem, we have markers: http://javlaskitsystem.se/2012/02/whats-the-waiter-doing-wit...
This is what many startups miss when promoting technologies to Fortune 500 companies.
In the enterprises usually you don't change stuff for the sake of it, unless you're trying to keep the budget for the next year and need to avoid having it reduced.
As such it is easier to introduce incremental changes as radical new ideas.
If you're interested in enterprise startups you should start looking for a really specialised piece of workflow that is common among huge company's and try to solve that in its entirety better than the existing solution.
This breaks the pure SaaS model though, as they'll probably not want their pieces given to their competitors.
Everyone wants to be special, and it's a very compelling internal narrative, that your stupid/wasteful/annoying/irrelevant processes (and thus software) is actually a strength.
A fair few businesses took up the offer. Essentially they're paying to bump up functionality in the dev cycle that we'd eventually get round to, but maybe years away still.
SAP themselves are pushing HANA, which is seriously cool tech and will be the foundation of moving big ERP into the cloud. A bit interesting how in-memory architecture is not discussed at all here at HN. Started in pure BI/Analytics with tools like Qlikview, but now it is growing up.
See here: http://www12.sap.com/solutions/technology/in-memory-computin...
ORA is a bit on the slow side, they are pushing better bundles with hardware instead (exadata).
Any one who wants to both have his cake and eat it is destined for disappointment.
So let's consider the logic of the proposition instead of particular companies.
The benefits of multi-tenant come from not having to pay to dedicate a resource. We can both use it and we each only pay a fraction of the cost. However, there is trouble if we both try to use it at the same time.
The benefits of SLA or any kind of quality of service guarantee come from having a dedicated resource. You know that you can always use it because I (or anyone else) can never use it.
These two goals are not compatible.
Especially when you buy down the rates for 'reserved' instances (or even the FFP GovCloud stuff).
Second, the reason for this "cheapness" is not multi-tenancy but the complicated rules and structures.
The larger point is that Rodney Rogers' article asks for too much.
In fact, I believe most companies would go bust trying to fulfill his requirements. If I were to win his business I would turn it down.
I believe the best he can hope for realistically is that SAP add some more hypervisor support to their software.