Intelligent Automation for Cloud Value Calculator
The customers I talk to know that deploying a private or hybrid cloud will both save them money on IT operations and make them more agile to respond to the business. There is a low grade euphoria over the cloud opportunity that gets the conversation going. The conversation drives development of both our solution and our customers’ sophistication in thinking about how and why they will use Intelligent Automation for Cloud (CIAC).
However, finance guys and IT management don’t get that feel-good feeling over the opportunity or even the coolness of the technology in the absence of dollar numbers to motivate them.
Nor should they.
We are in a part of high-tech that does not do technology for technology’s sake. We do it because it makes business sense.
We’ve been working on a couple of different methods to help quantify the value of CIAC. One is to walk through particular, heavy, existing IT procedures and identify and quantify the time savings for each step. Types of steps include discovery of available services, self-service requesting, project management, provisioning/configuring systems, and maintaining IT business system data. For many of these, CIAC saves time by either reducing/eliminating human effort required (person-time) or simply reducing the end-to-end process run time by doing things faster (clock-time). Both can be assigned a dollar value and a number-of-executions to arrive at a total annual dollar savings for the analyzed services. In some cases, one or two detailed examples can show justification for the acquisition and deployment costs (payback), and the potential savings on the rest of the processes are gravy (ROI). These are very good for showing justification to senior management and “the finance guys”. This method has been very well received by the early customers with whom we have done the analysis.
The second is to take a holistic look at IT operations, attach general savings percentages to various segments of IT spend, such as operations (keeping the lights on), new deployments and service requests, efficiency/standardization, optimization of resources, and IT ticket reduction. I had a chance to run through this with a customer last week and it, too, was well received. It’s more of a back-of-the-napkin grade of analysis, but it’s exactly the kind of thing to show around internally to build awareness and buy-in for this big new “cloud project.” We’ve vetted the broad savings percentages with multiple IT shops to ensure that we are not blowing smoke, and have erred on the side of less savings than we think we can achieve to ensure we’re not overselling before proceeding to undeliver and disappoint. I assure you, Cisco is very serious about customer satisfaction and pushes that into every level of the organization, including into our compensation.
For the customer with whom I tried this calculator last week, the savings was on the order of 10% of their annual IT budget, and based on their experience from a prior software standardization effort, the customer thought that was entirely reasonable. Payback time was less than 6 months.
Overprovisioning and Sprawl
Rodrigo covered some of the causes in his recent blog post but not quantitatively. We estimated as little as 3% reduction in sprawl by using CIAC. When you have really fast delivery time (clock time) for new machines, and can reclaim the resources automatically at some expiration time, we think we can do much better than this, but we left it at 3% of the hardware component of IT’s budget.
This is one of those areas that are excruciatingly hard to quantify, yet everyone knows that a lack of agility costs real money and real pain. In our ROI model, we used the idea of a “launch” to indicate those big, one-time items driven by the business that IT is supposed to be able to rapidly pivot to support. Most of the time, these require one-time custom infrastructure and/or software deployments, which obviously do not come out-of-the-box from IT’s vendors. They may not even fit in with existing operations. Yet in nearly all cases, these can take advantage of building blocks that an IT shop already has. Cisco Cloud Portal and Orchestrator already have extensive building-block concepts for rapid reuse. Standards, Active Form Components, reusable rules and atomic processes are what I call “complex primitives” that can be mixed and matched to quickly create new services out of existing service components. We like to work with partners (for example) who have similar concepts so that the building-block concept extends to our broader ecosystem. Of course, for a “launch” there will be new pieces that have to be built from scratch, but that’s what makes it new and innovative. For the ROI model, we simply considered the reduction in time of rolling out whatever this big new thing is by being able to use these building blocks. The model allows you to plug in the dollars-per-day saved by either being early, or being less late. It’s not at all unreasonable to have $10,000 or even $100,000 per day of savings and added value, so 10 launches per year, each of which save an average of 5 total project days by being able to use these blocks, means another $500,000 (or $5,000,000 if your operation is big enough) in real savings and benefit.
Of course, the Value Calculator has a switch that allows you to turn this off if it’s not something that the finance guys can believe in.
Operations Models, Policies, Processes
The big savings is in standardization. At my last employer, I worked with a large customer who was able to show quantitative savings of $100,000,000 per year in their IT budget simply by standardizing laptop and desktop hardware and operating systems, something like $500-$1000 savings per end-user. The customer I spoke to last week had shaved somewhere around 5% off of their IT budget just by standardizing their software installs.
CIAC extends that standardization to everything that the IT shop offers. It’s very difficult to get motivated to write procedures or build scripts for IT services that will have to be modified in three months or replaced in nine months. The short lifespan of many things in high tech means that standardization just doesn’t happen, or that the lifespan is artificially extended. For example, the large customer I mentioned who saved $100M/yr by standardizing their desktops was unable to move to either of the next two releases of that operating system because the standardization itself was so much work.
The promise of Cloud is that standardized services can be built much more cheaply and rapidly on top of hardware that can adapt to the changing requirements. IaaS on UCS underpins exactly this opportunity. Active Form Components, Atomic Processes, and Service Profile Policies can be quickly built, recombined, and reused to create new, yet “pre-standardized” services from IT. Individually, these components have much longer lifespans than the composite services that combine and use them, but now you only have to write a procedure or document around the new piece and not the whole services. While the document for this new service piece-part may have a short lifespan, it also took little time to write, so the balance is struck and the service actually gets documented, standardized, deployed, used, and retired without unreasonable effort.
For our ROI calculation, we estimated that:
- as few as 30% of service requests could be self-serviced (reducing project management and architecture time). Among standardized services, this number is around 90%, but we stayed with 30% for our calculation
- 5% of low-value or out-of-date services could be eliminated or outsourced. In the short term, this is an effect of evaluating which services to move to the CIAC cloud. In the long term, running reports on what is used and the value of those services continues this ongoing optimization.
- Only 50% of custom build requests could be eliminated. With standardized services, end-user visibility into cost, and orders of magnitude more rapid delivery for the standard options, there are intrinsic motivators for the end user to go with standardized options. This should get to over 80% in short order, but we stuck with a conservative value.
- 60-75% of provisioning person-time and lifecycle management could be reduced using CIAC automation and service management capabilities. When the end user has visibility into the service delivery, there are fewer calls to IT asking how things are coming along. With regard to human effort, automation can shave off more than 99% of the person-time involved in provisioning
Obviously, getting to the high end of these values will take more time to implement, but we think the low-ball numbers reasonably represent the near-term value of CIAC.
The low-grade euphoria is from the instinctive sense that this value is here. We’re helping quantify it to those who want numbers more than instinct.