My second webinar report today features a free Gartner seminar from Jeffrey Mann, who I knew well in his META Group days as a great application analyst. The topic isn’t “what’s available” but “how do you make decisions”: potentially much more useful.
First, he’s talking about Cloud for absorbing capacity demand peaks: the right definition. But, as he points out: the high-end integration requirements of a portal don’t necessarily suit well to Cloud infrastructure. Security and confidentiality play as issues too.
Compared to my post last week, Gartner’s definition of Cloud matches in most elements but I included easy sign up without long term commitments. This matches the use case for absorbing capacity peaks, but for longer-term critical business functions (running your sales on salesforce.com, for example) most consumers will want some longer term assurance.
Gartner also add to the established model of System (or Infrastructure), Platform and Application as a Service: two further levels. Information (e.g. a Reuter’s feed), and Business Services. The shift in provision focus is from “capacity” to “capability”, and evaluation is outcomes based. I like that.
Jeff “gets nervous” if cost saving is the only reason clients are moving to cloud services; cost reduction may be part of the outcome, but there are hidden costs (e.g. increased network capacity) and many disappointments. “Disconnect price from cost … reconnect price to value.”
And, perhaps closer to the meat of the theme: “Portals … will follow … The greater portal opportunity [for Cloud] lies largely with B2B” – strikingly close to Mark Benioff’s Cloudforce message I was listening to earlier. More on that later.
Early Cloud deployment: look for something that will work with the vanilla service (“out of the box” requirement). And it’s easier to start greenfield than migrating from on-premises services. Complexity (e.g. customisations) mitigate against migration.
Jeff showed a self-assessment chart for issues such as data, compliance, policies and failure remediation – how complex is getting going again after a stop? Even with due diligence, it comes down to trust – usually lower for a pure-cloud solution. Users often prefer to be in control even of functions and processes they are not so good at.
What about best practices? Half a dozen use cases, for example capacity on demand (such as hiring lots of extra staff for a short time, I guess like Christmas postal deliveries) – Jeff calls this “Cloudbursting”. I’ve heard a presentation of this being done around a massive weekly sales promotion that, on its first outing, unexpectedly and grossly overloaded the company’s normal web servers. Cloud to the rescue!
Other use cases include: providing lower-end capabilities to segments of the staff population, such as floor staff in a store; secure extranets in an isolated environment (e.g. in M&A or restructuring when information needs to be kept confidential to a subset of staff but in more than one enterprise); or “splitting the stack”. Jeff proposes a small handful of hard-headed questions to help evaluate whether cost will really be saved (bearing in mind that he asserts that cost saving shouldn’t be the only target for a Cloud move). You must be able to identify where cost savings will come from, they’re not automatic!
And do, first, hold a mirror to your entire company and ask if – culturally – the enterprise is ready to make this kind of change (it’s trust, again, but legal issues such as e-discovery may be highly relevant here). Then be sure you understand why: examples might include flexibility, cost, being more easily current to latest software versions, reduced internal resource requirements, and so on. And you must have defined measures for a pilot to judge whether to move on.
In this, recognise the many constituencies in the business with different needs and expectations: not a new idea, but a useful categorisation of business interests on the chart. Think how to get them involved.
Gartner does expect that in the next few years all organisations will have some level of Cloud service: complete (few), or mixed (most).
In the brief Q&A, the issue of recent high profile outages (e.g. on Office 365 for some customers) was raised. Jeff’s view is to keep it in perspective: compare with internal capability, not with the ideal of zero outage.
I raised a question based on listening to Mark Benioff earlier. That presentation was put out via Facebook, and Benioff was strongly promoting “social technology” as the communication and collaboration platform of the future. As an analyst, Jeff sees very few innovations which totally replace what came before, and believes this will be a case in point. So social technologies are, indeed, very important; but the older platforms won’t go away.
Congratulations for Jeff for going beyond the technology of Cloud, and the (perhaps hyped) potential. This was a good counterbalance to Mark Benioff’s evangelistic case, and confirms that Jeff has lost none of his edge since last time I had the chance to interact with him!
• Cloud Strategies for Portals, Content, & Collaboration Projects, Gartner webinar, 14 Sep 2011, replay (link to be added when available) – in the meantime see Gartner Webinars
• Mark Benioff at Cloudforce London, ITasITis, 14 Sep 2011