As Boomi CEO Chris McNabb noted last year at Boomi’s Out of This World Conference, customers today are looking for integrated experiences: digital interactions that give them everything they need, anywhere, anytime.
What’s so special about the integration in integrated experiences? Traditionally, individual business units built integrations and workflows using just the data they controlled. When these disparate workflows were combined in an application or website, the resulting lack of coordination often created unsatisfactory experiences for customers.
Today, companies realize that customer experience is king. They know they need to integrate and orchestrate data across business units into a cohesive experience that gives customers exactly what they want, when and where they want it. Even if data in an experience comes from seven different business units, the experience itself needs to be straightforward and rewarding for the customer. That is, the experience needs to be integrated, coherent, and seamless.
The integrated experiences companies are building today take advantage of:
- Data readiness, which means finding, cleaning, and transforming data for use in customer experiences; it also includes establish data lineage controls, so that good or bad data can be tracked back to its source
- Pervasive connectivity, which means linking apps, databases, and data repositories to devices, processes, and people to support customer experiences
- User engagement, which means creating workflows and user interfaces that provide frictionless, coherent experiences
Integrated experiences are transforming customer interactions in industries such as financial services, healthcare, and transportation.
Observability can help IT organizations ensure these experiences deliver as promised. Observability means setting up controls so that if something goes wrong with any aspect of an integrated experience, the source of the problem can be quickly identified and the problem corrected.
Observability Helps Ensure Integrated Experiences Deliver as Promised
Observability sounds straightforward, but there’s a lot involved in getting it right. Imagine a ride-sharing mobile app that delivers a fast, frictionless experience for people hailing a car for a local trip. The app probably calls on all sorts of data sources: GPS data, weather data, location data, dynamic pricing models, mapping software, and so on. Ensuring that IT engineers can quickly pinpoint a problem with any of those data sources, workflows, or user interfaces requires visibility into all sorts of different components built with all sorts of different technologies.
Those application components might be the responsibilities of different business units. Some components might be in on-premises data centers. Others might be running in the cloud. Still others might be provided by third parties. But they’re all essential for the integrated experiences customers are depending on.
Think of observability as getting answers to questions that matter about customer experiences. In the case of the ride-sharing app, those questions might include:
- Why is the app running slowly?
- Why is the app not accurately estimating trip times?
- Why is the app not accurately showing the location of the driver’s car?
- Why is the app unable to find the customer’s account?
- Why is the app unable to process a financial transaction?
To get answers to questions like these, IT organizations need to build observability into their IT architectures. That observability should cover several types of monitoring, which I’ll discuss next.
General Integration Execution Monitoring
General integration execution monitoring traces the execution of integration processes in an experience. It helps answers questions like:
- Which integration processes are running?
- Which process in a workflow needs to run next?
- If processes are triggering alerts and error notifications, where are those alerts and notifications being collected?
- How are processes configured?
- Which processes have been scheduled?
- How can we measure the latency between processes?
- Are there third-party tools that can help us monitor process execution?
As you’re building integrated experiences, think about how to instrument your integrations and processes to report data that could be helpful in answering questions like this.
Instrumentation can take several forms, including messages sent to a log file, alerts sent to a ticketing system, and status information that could be queried. Think about what information should be provided by a workflow or application, so someone investigating a problem could quickly discover the problem’s source.
Business Transaction Monitoring
Business transaction monitoring means tracking the processes and integrations that make up a transaction, such as an ecommerce order being placed or received. This type of observability helps you answer questions such as:
- What is the status of an order?
- Has the order been received and properly processed?
- If an order was expected but hasn’t been processed, how do I determine its status?
- How do I determine which part of the process is delaying or losing orders?
- Can I track a business document, such as an order in the form of an EDI document or PDF, through all the stages of the order and fulfillment process?
- Can I track any business identifier, such as an order number, through all the integrations and workflows it’s associated with?
It’s a good idea to build this type of monitoring into all your business-critical workflows.
Think in terms of each stakeholder. If a customer doesn’t receive an order and they can call to ask about it, how would a customer service agent discover the order’s status? How would the finance department confirm that an order has been delivered? How would the shipping department answer the same question? Identify the stakeholders involved in a transaction, and build monitoring into workflows so all these stakeholders’ potential questions can be answered.
End-to-End Business Process Monitoring and Troubleshooting
Observability also needs to include the monitoring and troubleshooting of end-to-end business processes—the overarching sequence of activities automated through one or more integration processes.
This type of observability helps answer questions such as:
- Which processes, including non-integration processes, are running in support of an overarching business process? (For example, if a business workflow involves a currency conversion, can you track the state of the conversion along with the integration processes that make up the workflow overall?)
- What is the known state of a process?
- If there’s a problem, is it tied to a design problem, a problem with a Boomi Atom, a problem with other systems, or a problem with data itself?
IT organizations typically turn to process logs and log analysis tools such as Splunk to answer many of these questions.
It’s a good idea to set up some system for answering these questions. It’s also a good idea to make sure that system works when processes are running at scale. If an integrated workflow reads, transforms, and writes 100,000 records, and three of those records fail, can you easily track down what caused them to fail? Make sure that whatever monitoring system you put in place can help you answer questions like that one.
Performance Monitoring and Troubleshooting
Observability also includes monitoring the performance of integrations and workflows. It helps answers questions such as:
- Are processes running as quickly as expected?
- If a process is underperforming, what’s the root cause?
- Can we tell if the system has been designed to deliver high performance at scale? Or are we detecting performance problems that indicate we should allocate resources differently or redesign the system being monitored?
It’s a good idea to benchmark business-critical processes and workflows, so you can quickly determine whether or not performance is meeting expectations.
Historical Trending
Historical trending builds on the practice of benchmarking. How are processes and workflows performing over time? Is performance improving or worsening? Are data volumes increasing?
Trend analysis can help with capacity planning. For example, if data volumes or transaction volumes are increasing, it might be prudent to add computing capacity or storage.
Historical trending can also help executive teams and centers of excellence assess the use and performance of a new product or service. Find out which trends business and IT stakeholders would be interested in tracking. Then instrument your integrations and workflows accordingly so you can report regularly on those trends.
Assessing Observability with User Journeys
Ultimately, the purpose of observability is to make it easier to deliver great customer experiences. It helps IT teams and business units quickly determine how to solve problems when a customer experience falls short. Without it, problems can compound, leading to lost business and tarnished brands.
It’s a good idea to evaluate observability from the point of view of user journeys — the progression of steps and decisions a specific user makes as part of a workflow or application.
For example, imagine a customer service agent getting a call about a product not having been delivered. What systems would the agent check to determine the status? Can the agent determine if the product has been shipped, is waiting on payment, or has been lost?
Pick some plausible scenarios for important user journeys associated with a workflow or application. Then make sure you’ve put the observability controls in place you need for resolving questions or problems that are likely to arise with that workflow or application.
In every industry, companies are delivering innovative integrated experiences that transform traditional business. By building observability into those experiences, they can ensure that business-critical operations meet goals for performance, reliability, and customer satisfaction.
Read about Boomi’s vision for Integrated Experiences, and contact our team of experts to learn more.