Maneesh Joshi

[This article was written by Maneesh Joshi] Back in 2005, the concept Business Activity Monitoring (BAM) was coined by Gartner to capture the requirement of visibility into business operations. BAM delivered this visibility through aggregation and summarization of information around business activities. The BAM solution would capture event data, aggregate them into a single store, apply some context on top of the data, and then deliver dashboard-based visibility to business operations managers and sometimes executives. A few questions that a typical BAM solution would answer for the business operations manager at a mortgage lender were: How long does the typical loan application take from submission to approval? Where is the most time spent in the loan application workflow? How many calls does the average call center rep answer per day? Unfortunately, BAM never delivered on it’s promise. I can think of four challenges that made the BAM solution fall short. 1. Vendor-Focused Solutions The first challenge with this approach to BAM was that the solutions were typically built by the large enterprise software vendors that offered it as a complement to their own big and wide technology stacks. This approach made the offerings very vendor-centric and often didn’t cover the heterogeneous set of technologies that most businesses owned. The visibility hence was rather restricted to the large vendor stacks and created information silos and narrow visibility. 2. Too Much Customization The second challenge arose as a result of the fact that no two business processes are alike. There were very few commonalities between business processes for the vendors to build repeatable and scalable pre-packaged dashboards around. In order to build a solution that worked for these processes, heavy customization and services were necessary. For the solution to be meaningful and provide the right kind of visibility, the customization often included intrusive code changes to the application in order to raise business events, whether it was Java code in the Java tier, or the SQL packages in the database tier. 3. No Real-Time Visibility The third challenge was from the lack of real-time nature of the visibility that the BAM technology delivered. By the time the operational data was collected from the sources, aggregated into a database, applied context on, analyzed, and presented into a dashboard, the visibility provided was already stale. In order to address this staleness problem, the notion of complex event processing (CEP) came into being. CEP technologies were purpose built to correlate events and detect patterns across event streams in real-time. 4. Lack of Business Context Despite CEP attempting to solve the staleness problem, it introduced a problem of its own. The events processed by CEP are typically lightweight events and low on context richness. When CEP would attempt to correlate events to detect patterns across disparate streams, it lacked the business context. This lack of context made the solution less effective as no meaningful pattern matching or correlation could be done without the context. The business context needs to be captured at the source when the events are being captured in order to make the correlation relevant. 5. Premature Technology Stack The last and main blocker was that the underlying technology stacks and the surrounding applications were not mature enough to support the vision. For instance, the relational databases that couldn’t capture unstructured events, or the inability to crunch massive volumes of data in real-time at low latencies while applying the context severely restricted the vendors from delivering on the promise. The packaged applications were very monolithic with very little extensibility, which resulted into implementing intrusive code changes to raise business events. To be fair, the vision for BAM and CEP was spot on then and is spot on today. The vision was clearly ahead of its time and the problems it promised to solve very real. Who doesn’t want visibility into their business operations?! In fact BAM and CEP become all the more important is today’s day and age of software-defined businesses. In my next blog post, I will discuss a new approach to BAM. I will also review as to how the technology and application stacks have evolved, and why you should consider revisiting your opinions about BAM and CEP use cases.

if you’re just tuning in, please check out my previous post where i dispel myths 1 & 2 , bad app ratings are uncontrollable and it’s impossible to understand your backend services. typically, mobile app developers accept some things they feel they can’t change — ratings, end-to-end visibility, user experience, and so on. however, these can all be avoided and under your control with the right mapm solution to give you the proper insights to give your users a seamless experience. now, on to the myth busting… myth #3: users are an enigma that i can never really understand you cannot nail down the mobile end-user experience unless you know your audience. you need to understand where the end user is spending time while using your application. are they spending time scrolling down search results to find what they want? in other words, are you presenting the most relevant information at the top? are they abandoning the shopping cart at any specific points in the checkout process? is there funnel friction that you need to optimize your app against? the modern mobile apm tools have some great capabilities to understand your end user and their behavior. you can inject timers across any two arbitrary points and measure times taken for a collection of any number of steps. for example, you can measure how long it takes your user from conducting the first search to purchasing a product or a service. this can be done at an individual user or at aggregate levels. you can measure how much time users spend on which screen. this will give you great insights into who your typical user is and what interactions do they indulge in with your application. you can then optimize the app experience for those common patterns. myth #4: i’m going to spend the rest of my life certifying my mobile app on the infinite permutations and combinations of device types, os types, and network carriers/types this is where you need concrete data to understand your user demographics. a good mapm solution will give you detailed breakdown of who your core audience is. what device types they prefer, what os’s (ios vs android) they run, and which networks they mostly originate from. a good apm solution will also allow you to correlate this information with revenue or engagement information to determine your highest-value audience. with all this valuable information, you can prioritize development, testing, and certification of your mobile app. you can even optimize your app experience and test for performance bottlenecks for the high-value audience. and lastly, you can focus on retaining them by delivering on their roadmap demands over the lesser engaging ones. myth #5: there’s no way to know the business impact of the performance issues of my mobile app most mapm tools in the market today are too developer-centric. they deliver crash analytics and performance delays caused by delayed response from backend services but little else. often times, the mobile channel is an enabler of some business goals such as better customer engagement, additional revenue streams, cost savings from productivity or efficiency gains. plus, it’s the broader context that feeds investments into the mobile channel. ignoring the business context is like missing out on half the picture. the right tool needs to deliver full context on the mobile application. the full context should include what impact the app has on business metrics such as revenues, cost savings, customer engagement kpis, etc. a comparative chart that shows performance impact of mobile app on these business metrics can be incredibly powerful to raise awareness among the organization. with these myths dispelled, i hope you have gotten a different perspective on your mobile app initiatives and are rethinking your approach to mobile apm. feel free to leave comments and share your thoughts. also, check out my previous post to learn about myths 1 and 2 . from interested in trying appdynamics mobile rum? check out our free trial ! for a introduction to appdynamics mobile real-user monitoring, watch our on-demand webinar now.