While I’m no expert on the topic, I’ve spent a lot of my last years doing just that within our company – which is not a pure player in the space. Here are some ponderings which should be useful both for providers and for customers.
If you’ve not been living under a rock for the last years, you probably noticed the hype called “cloud” (not to mention “big data”, let;’s just leave it to that for a moment). As many pointed out, “cloud” has many meanings and while some are new, a lot of work had already been done under different not so cool names like PaaS, SaaS (why not AaaS?), IaaS.. etc. Infinite elasticity, scalability and “always on” are subliminal promises vendors lure customers with. Some deliver this better than other. Some customers need it, some not. Well… it’s a free market after all.
Say you are a solution provider (integrator, services, whatever) and want to jump on this bandwagon. You call somebody in the company and say “let’s do this!”. If you’re thorough enough, you’ll make a business plan (at least). Please consider this:
1. You can’t just take on-premise software solutions and make them run in “cloud”.
Especially ECM stuff. If that solution is a good onw, it’s probably very complex. Complexity doesn’t play nice with reusability. Cloud is about reusing as much as possible so you can make economies of scale. If you don’t care about that (eg: you expect less than 20 customers) then consider the people administering it. They will be your staff? If so, do they need to learn a lot of specific characteristics for each of your customers? Hmm….
I’m sure all your enteprise customers will use the same software application, just configured differently. Aren’t they? You’ll educate them? Or you will be the next Force.com? Consider it. Of course, you can isolate them and basically have an individual instance with all the peculiar characteristics customized for each customer. I sure hope they will be paying you the big buck, because it will be fun to keep that updated in time.
Or you will take the enterprise solutions you have and scale them down to fit a more “general”/small business need. It might work. Remeber: keep it veeeery simple to use. Very. For example; Metadata? You should innovate here. Checkout/checkin? Find another way. Your 20+ action menu? Trim it down to 5. Can’t?
2. Infrastructure is not scalable to the infinity
Put a virtualization layer on top, it eases things up. But you thought of that already. Reality check #1: virtual stuff runs slower than barebone hardware. Reality check #2: top grade infrastructure is expensive. Of course you can try the Google approach and build a lot using commodity hw. But you need special management tools and procedures fo do it. And maybe even rethink your software solutions. Goto no. 1.
Some ECM solution components might not even be compatible with virtualization. Hehehe. Are they yours or do you purchase/oem/integrate from other vendors? What;s their plan?
3. Business model si a different beast
How do you sell it? With the same sales people? How do you incetivize them? I’m sure you are aware that selling cloud stuff brings a tiny-tiny fraction of the initial revenue vs. selling one-off licenses & implementation services. When does your salesperson get bonused? Will this work for them or they will continue to push for the classic sale even if you have cloud to offer?
On the good news side, if you sell in the ECM space you should have already been thinking about this for at least 1-2 years now… why? The days of multi-million solution sale are gone. The majority of customers are now planning for smaller initial costs and to have services and extensions spread over multiple years. With the ocassional exceptions, but if you’re counting on those, then you’re reading the wrong post.
What about licenses ? Your ECM solution will most likely sum up an impressiove number of third party licensing. Are those vendors providing a clear and adequate licensing model for the cloud? Do you change your user per GB stored? If so, you’re storage vendor is charging you based on the monthly reported usage? Eh?
4. Your internal processes and resources
Once you go live, you need to run it. If you have only a few customers, it’s probably not very much different than what you did until now. Especially if you already provided some sort of SaaS/PaaS before. But this is not the idea… “Cloud” should mean many more customers for you as it means more resources for your customers. Cheaper, for both. So, you will need to redesign internal processes (and tools) to be more effective. When you sell at 5 USD/user/month you need to take inspiration from Henry Ford. Considering you will still keep you current business model, beware how you mix people and resources around.
Support for an ECM solution is complicated. Will you take calls from all the end-users? Are your people capable of understanding this for all those many “cloud” customers you will hopefully have? Or they will just create frustration which will generate a lot of comments in the area of “why did we externalized this? it was much better when our (then thought lousy) IT handled it”. Remember, you’re changing a way of working, you’re not providing a solution to a greenfield area. What works for a telecom company providing commodity services and products will not apply in the ECM space.
The list can go on…
Is it all bad?
No way! Just look around. Although I called cloud “hype”, it passed that stage. Vendors are started to mature and many have realized the above items and are at avrious stages to address them.
This post is for you, our customers! Take care when choosing your cloud provider(s). I really do think you need one (the chances are you probably need cloud vs. you don’t). Do not evaluate in the same way Amazon with Box with EMC or with a specialized solution provider.They are different, at least for the reasons stated above. Work with them in understanding how did they solve the challenges, see their actual experiences and be cautious of “i know how to do it but we have no customers yet” offerings..