I had the opportunity to talk to Jon Anhold, an Associate Technology Partner at Rosetta, a leading Digital Agency that specializes in IBM WebSphere Commerce about how they consistently deliver Omnichannel success to their clients. His feedback and responses were really insightful and I just couldn’t help but share them in their entirety!
Matt: Rosetta describes its mission as ‘connecting rich data, engaging experiences and robust technologies to forge meaningful relationships with customers that drive business impact.’ What does that mean exactly?
Jon: Perhaps the simplest summary I can give you is that our objective is to help companies better interact and engage with their customers. It could be using data to determine what experience will most engage a specific e-commerce customer, or working with our clients to improve the in-store experience of their customers.
Matt: You mentioned e-commerce and one of the buzz words that we’re used to hearing is omni-channel. You must hear that a lot, what does it actually mean to you?
Jon: It used to be that we’d work on e-commerce platforms and mobile was a stated requirement, like an additional feature that a marketing team had thrown in. Today, customers expect that they’ll get the same experience on their connected device that they do on the desktop or in the store. For example, they want to be able to order online and pick up in the store seamlessly. Perhaps the most important aspect for brands is being able to create omni-channel experiences that are contextual, personalized and delivered in real-time in order to create meaningful engagement with their customers.
Matt: Can you describe what challenges you’re facing delivering these projects?
Jon: The main challenge is the pace at which brands can keep up with the ways consumers are accessing, interacting and digesting online services. There’s a very real melting pot where user expectations for performance – like the range of available devices/browsers/OS’s, the choice of fixed and mobile networks, and client initiatives to bring new services to market – all come together into a set of challenges that need to be driven by the business with a set of priorities that align to corporate goals. Brands must remain agile in order to the meet ever changing demands of their customers.
Matt: So how do you overcome these challenges and actually deliver?
Jon: End-to-end visibility is really important. The ability for Dynatrace to provide data that starts with the customer’s browser or mobile app and follows that interaction all the way back to the back-end systems, database queries and even the methods being called in code makes it much easier to pinpoint where a problem may be occurring. Understanding user behavior and capturing the atomic detail of the customer journey allows us to provide information back to the key stakeholders to make educated decisions based on hard facts.
Matt: Could you define who you see as the stakeholders and the kind of visibility you’re providing?
Jon: Most of our customers have three distinct sets of stakeholders that own the decision making in a project.
The Line of Business Owner – This individual is interested in the health of the overall application and infrastructure, but is primarily focused on revenue, conversions, cart abandonment rates, all the usual things you’d expect that define the critical data around the user experience. They’re increasingly interested in how users are accessing the app and want insight into whether there are trends and correlations across devices, browsers, bandwidth etc. – especially when they have a positive or negative impact on conversion.
The Operations Group – The misnomer is that the Ops team is there to keep the lights on. The reality is that Ops is directly responsible for revenue – if the site or app goes down, both the brand and the bottom line are impacted. They’re interested in lowering mean time to repair (MTTR) and increasing the mean time between failures (MTBF). Knowing where to invest their time and money is critical, as they must make decisions based on ROI.
One interesting aspect is that Ops teams are moving away from solely MTTR and focusing more MTBF – how can they prevent the issue in the first place. I think this is largely due to the culture of DevOps, where Ops are no longer just getting better and faster at fixing issues but they’re also feeding back root cause data to developers so they can fix bad code at the source. This improves MTBF and allows Ops to be more proactive instead of reactive. It’s a much more holistic effort.
The Development & Test teams – The Dev/Test teams are seeing the value of working more closely with the Ops team. If the Ops team is empowered to use a tool where they can identify issues at code level and then send an entire user transaction or set of transactions with atomic-level context back to Dev, the Dev and Test guys know exactly what code to fix, what the use case was and what to modify in their test scripts for future testing. Ultimately what they all want is visibility into what Users are doing, what the impact is, and what they can do to provide a better experience.
Matt: What kind of data do your end-clients specifically request?
Jon: Most of my clients are trying to answer one of two questions. Why did a problem occur that caused one of my customers a poor user experience? Or, will my site be able to handle the next flash sale, Black Friday or other peak load? In the past we’ve had to rely on analytics data from systems like Adobe Omniture, Google Analytics or IBM Digital Analytics. When we have Dynatrace in the mix, we can also start to react in real time to incidents and provide correlated capacity data from historical periods to determine if we need to add capacity in advance of a planned peak.
Matt: You mentioned DevOps earlier, walk me through a typical scenario.
Jon: Our Managed Services & DevOps teams used to look at things like server CPU utilization, memory utilization, network utilization, database query response times, as well as JVM health (garbage collection timings, heap utilization, and thread pool utilization). We realized, however, that measuring these independently and outside of the end user transaction times didn’t really matter – in fact 90% CPU and Disk utilization can actually be a good thing when done right.
If there is an issue that impacts the user where a performance baseline has been violated, an alert is raised and the Level 1 team will hop on and take a look to pinpoint where the problem is occurring. Using Dynatrace we can quickly determine not only where the issue is, but we also know the business impact it is causing. So if the issue is severe enough, we can make the decision to setup a conference bridge allowing the incident manager to track progress of the incident and keep the client and other stakeholders up to speed, including the Dev team if there is a code issue.
Matt: What is a typical way for Ops to communicate issues that need fixing by Dev?
Jon: Depending on how each project team is structured, it could be as simple as standing up and talking to someone a few desks away, walking across the floor, or getting on a group chat took like HipChat to work the issue. Off hours our teams are available for escalation to jump on and collaborate on a solution. An incident management system or ticketing system like JIRA is typically used, and we’ll pass details and updates in the ticket as well. One of the great things about Dynatrace is the ability to export a session file and share it with other team members so they can see the exact same thing as they work a problem. It really saves a lot of time trying to recreate the issue.
Matt: How many application code releases are you rolling out for your retailers using WebSphere Commerce?
Jon: There is a definite trend towards releasing faster – no one is ever telling us to release slower that’s for sure! I’d say a typical release cycle for our large retailers on their ecommerce channels tends to be roughly monthly. Rosetta’s Managed Services & DevOps team has automated the release process for IBM WebSphere Commerce for our clients to reduce the risk of deployments and also decrease time to market for new features.
Jon Anhold is an Associate Partner in Rosetta’s Technology practice and leads Rosetta’s Managed Services & DevOps practice. Jon has been involved in some of Rosetta’s largest ecommerce platform implementations in his 7 years at Rosetta. When not keeping the wheels of electronic commerce turning, Jon can be found chasing his four small children around his suburban Cleveland home.