In this course we started by looking at what feedback is, and the types of feedback that can be useful for refining your application once it’s been deployed. Starting with exception or error feedback we saw how a loop should be designed to integrate with development workflows. We then learned how diagnostic feedback could be incorporated into mobile and desktop apps using App Center. Then we analyzed those exceptions and saw how to use those exception events to send notifications and create work items that were integrated with Azure DevOps and GitHub. For web applications we saw how exception data could be pulled from Application Insights telemetry metrics. We set up alert rules that would send notifications via email or integrate directly with third party systems. Again, we learned how work items created in Application Insights could be fed through into Azure DevOps and then appear in Visual Studio Team Explorer for actioning.

We need a way to easily assess the state of our development workflow and with Azure DevOps we can use the dashboard facility to create a custom dashboard with a variety of graphical widgets. 

Feature and usage is the next feedback element we examined. Using Application Insights, we can monitor the performance of an application not only in system performance terms, but also how users interact with it. The live metrics feature let us see real time usage. This is valuable information from a UI perspective. It allows to see whether the app is fulfilling its objectives in terms of conversion, and also is the application intuitive to use. This is valuable information for assessing efficacy of features, or whether those features need to be made more prominent within the application, or perhaps removed as they are redundant. Application Insights tells us where in the world and what operating system and browser is running our app.

Next, we went back to the App Center to see how we could implement event tracking and analytics in non-browser applications. Using the TrackEvent statement sprinkled liberally through an application you can see events being logged, and the log flow lets you see those events being triggered in real time.

User feedback is the final element in the feedback trio. This ranges from help desk contacts which are easy to categorize and integrate with workflows, through in app ratings and questionnaires, to forums and social media. Each of types of feedback have their pros and cons which more or less lie along the dimension of information accessibility versus information richness. Ratings are easy to deal with but are one dimensional. Questionnaires and surveys have more information, but the participants are self-selecting and normally have an axe to grind. Social media is where the unvarnished truth lies, but that is the most difficult to extract and integrate.

To make really good use of all this feedback data we need to know how it relates to our baseline state. We can use the feedback tools to arrive at baselines by analyzing historical data and monitoring changes over time. In the context of alerts the data coming in is dense and it will contain a fair amount of noise, so we saw how set up filtering to be notified of real events.

Using a combination of App Center, Application Insight and Azure DevOps you can build feedback collection and workflow integration solutions that will help you implement continuous feedback with relative ease.