Sunday, April 15, 2018

Bringing Continual Service Improvement to Life

If you are thinking about implementing ITIL® processes and you ask the question, ‘where do I start?’, congratulations – you have started down the Continual Service Improvement (CSI) path. Likewise, if you are looking at improving your services, applications etc., then you have also started a Continual Service Improvement activity.

Continual Service Improvement: Organizations talk about it and think about it, but in reality, often do not plan for it, schedule it, allocate resources to it or monitor it. Improvement initiatives are often reactionary in nature to a specific event and are not proactive in nature. Implementing a CSI practice requires management commitment and participation to move from a reactive to a proactive practice. Whether improving services, service management processes or the service lifecycle itself, there will be a cost to implementing a CSI practice; however, there is a much greater cost to not implementing the practice. Organizations will spend tens of thousands of dollars, hundreds of thousands of dollars and even millions of dollars to develop and implement service management processes; yet, they don’t have any plan on how to continually improve the processes. This paper will discuss the scope of CSI and where to start improvement initiatives.

Implementing CSI can be done in different ways, and the correct way is dependent on exactly what your organization wishes to accomplish in the short term and the long term and then aligning your improvement initiative to business goals and strategies. The scope of CSI in IT address three primary areas of IT Service Management:
 • IT Service Management Processes
 • IT Services
 • IT Function and Life Cycle Management

My recommendation is to start with improving IT Service Management Processes first, as improving the processes will lead to improving your IT Services. As an example, for many organizations, if they review their Incident Management data they will discover that around 70% of major incidents are change related. This percentage is too high and ultimately has a negative effect on the availability of many IT Services. Even though ITIL V3 is made up of five core books, the reality is that when starting on any CSI initiative, most organizations need to address pain points first in order to show value and gain the support of the business and functional groups within IT. There are often some quick wins with some low hanging fruit that will provide immediate improvement without having to develop and implement a full process.

Processes Improvement – Where Do I Start?
For many organizations, process pain points are usually found in Incident Management, Problem Management and Change Management. The lack of mature documented processes often drives organizations to consistent firefighting, reacting to events that are often self-inflicted, such as a high number of failed changes; incidents that are escalated to the wrong support groups; or a total lack of Problem Management to identify and remove errors from the infrastructure that often cause a high number of recurring incidents.

Change Management is a control process and thus it is important to obtain a level of maturity that provides the IT organization with the efficiency and effectiveness of managing changes in order to protect the production environment. There are often some quick wins with low hanging fruit that can provide some immediate improvement without having to develop and implement a full process. This could be creating a Request for Change (RFC) form if one does not exist.

Incident management is a data gathering process, so it is important to ensure that all incidents are logged into the appropriate tool and that there is a consistent priority model used for all incidents so that you can find improvement opportunities. An organization may discover that they handle priority 1, 2 and 4 really well; however, they have a tendency to breach more on priority 3. Developing assignment and escalation procedures can provide a quick win for Incident Management.

For Incident, Change and Problem Management it is important to start providing some level of management reporting to understand the health of the process. It is also important to realize that Change and Incident Management are customer-facing processes that provide a lot of IT visibility to the business and thus have the ability to create a positive or negative perception of IT.

Another quick win is the documentation and agreement of Operating Level Agreements (OLAs). OLAs need to be in place to support any existing Service Level Agreements (SLAs), Service Level Targets or Service Level Objectives. An OLA between the Service Desk and the rest of the IT functional groups such as the desktop, database, or application groups is often a good first step to ensure that there is a consistent handling of incidents through the Incident Lifecycle in order to meet any agreed to response and/or repair times.

You can use the below diagram to start your continual Improvement journey in a simple framework. You identify the Improvement opportunity, plan the tasks that need to happen, Implement the tasks/changes and review the progress and goals & strategic objective achieved.

How can SaaS companies use AI and Machine Learning moving forward?

Artificial Intelligence (AI), powered by machine learning principles has long been a driver of fiction. But with recent breakthrough in computing power and availability of massive data have made some of those fiction ideas into a reality today.

AI and Machine Learning are one of the big trends and it should be no surprise that SaaS is very much a part of this “big trend.” 

The SaaS Market today:

report from IDC shows that the SaaS segment, which makes up 68.7 percent of overall cloud market share, was the slowest-growing segment of the cloud market with a 22.9 percent year-over-year growth rate last year.

Venture capital funding, usually an indicator of how “hot” investors view a market to be, has been found to be on the decline when it comes to SaaS startups. TechCrunch attributed this in large part to market saturation and the fact that those newbies seeking funding are often trying to compete with large, established players.

While the SaaS market is still expected to show some growth, it’s expected to be at a lower rate than we’ve seen previously. SaaS markets are maturing and it seems that those who are going to come out as winners will need to be onto the next “big thing.”

AI and machine learning can help SaaS create a more strategic position

All big player like Amazon, Google and Microsoft in the market are announcing offerings that integrate AI. Oracle, another big player in SaaS market said it is placing a big bet on AI and Machine learning to overtake salesforce in SaaS. These are a solid indicator that that AI and machine learning could be the next step in differentiating a SaaS and helping it to carve out space in the market.

How SaaS use AI and machine learning

SaaS is picking up on the trend for AI and machine learning and investment in this area is on a continuous rise. Below are some of the SaaS solutions where machine learning is playing a strategic role.

1. Personalization

AI can bring the option of hyper-personalization to SaaS, something which we have already seen in mobile apps in particular (looks at Starbucks’ “My Starbucks Barista”). In SaaS, natural language processing and the ability of AI to learn from user’s previous interactions can help configure user interfaces so that they cater to the individual.

For example, if you think about any SaaS without AI capability, adding more functions or features tends to cram the user interface and add complexity for the user. AI can help not only with personalization but with the easier adoption of features.

 2. Automation

Automation is demonstrated in many different ways in SaaS with AI baked in. It can take over where previously manual functions were required, for example, in the case of chatbots that help to provide users with answers to basic questions.

Automation reduces costs because they eliminate the need to hire additional people to handle more work. A bot replies to login reset questions with an automated response in a link to a knowledge base, freeing customer support reps to focus on more challenging questions.
One of the challenges for SaaS is always keeping an engaged customer base from a remotely operated perspective. It can be difficult staying on top of customer service requests and ensuring that every customer gets a good experience. AI can help with that by reducing that remoteness and stepping in to supplement human effort.

For example, there are already several examples of apps (Verizon, banking apps etc.) where chatbots answer questions up to a point, but refer users to human operators where necessary.

3. Predictive analytics

There are many potential ways in which AI built into SaaS can leverage predictive analytics to create a better user experience and/or help to arrest churn for SaaS. For example, machine learning can help to predict user preferences or behavior, then perhaps trigger alerts or actions when it appears the user is disengaging.

5. Product Search

When the user searches for a product, how do we find the best results for the user? One factor used in product ranking is user click-through rates or product sell-through rates. In addition, user behavioural data gives the link from a query, to a product page view, all the way to the purchase event. Through large-scale data analysis of query logs, we can create graphs between queries and products, and between different products.

We can also mine data to understand user query intent. When a user search for “Toyota Prius”, are they searching for a new car, or just repair parts of the car? Query intent detection comes from understanding the user, other users’ searches, and the semantics of query terms.

5. Deploying Code

The consequences for a SaaS rushing through the code and deploying early, only to have a crash or bug that affects all users can be very costly. Reputation and potential liability issues abound, yet being able to deploy quickly can be a distinct advantage. If you’re in a competitive market, the difference between leading or lagging can be if you are first to reach people.

AI is a game-changer for SaaS developers because it can augment their own coding abilities by providing the necessary checks that the coding is good. Deployment can be cut down from months to a very short time when AI can verify that the SaaS is built to scale to thousands of users.
Docker is an example of this, checking and testing code for quick deployment. Look out for further developments in this area – Microsoft and the University of Cambridge are working on teaching AI to write the code itself.
7. Enhanced security

Cloud security issues are always a hot topic among SaaS, and traditional security measures tend to be static, perimeter devices which require human input to update for new threats. AI gives SaaS the possibility of security services that can replicate and learn from new security threats automatically. Oracle has recently added machine learning and AI to their cloud security services, facilitating automated threat detection.

Big Opportunity Ahead

AI represents a new generation of SaaS products and the opportunity to embrace new ways to gain a market edge. We are seeing many of the bigger players move into this space already and industry experts predict it will continue to grow.

By and large, the most frequent applications of machine learning in SaaS today are efficiency applications - automating the high-volume manual processes and reducing costs. Consequently, if you looking to build a machine learning based SaaS company, find a really expensive internal process and automate it.

Are AI and machine learning considerations for your SaaS? The market trends show that perhaps they should be