Implementation of usage analytics 6 central criteria for Software Usage Analytics
Those who analyze the use of their software products can make data-based decisions and strategically expand their product roadmap. If you employ developers yourself, you can implement such an analysis tool yourself. But there are some aspects to consider.
Companies on the topic
Developing and implementing your own software usage analysis solution poses several challenges.
Meaningful usage data is absolutely necessary for software providers to optimize products and to retain customers in the long term. The analysis reveals how customers use applications, which features are used most often or on which platform the software runs. UI / UX problems or platform-specific bugs can be identified and fixed faster.
The developers also gain a better understanding of what users actually need for their daily work. This is the only way to achieve precise coordination between the product portfolio and the practical requirements on the customer side. However, when it comes to implementing software usage analytics and implementing a corresponding system, there are some important aspects to consider.
1. Planning in advance
Like any complex software project, building an analysis system for software use requires a high level of cooperation. A methodical selection of metrics and metadata guarantees that all teams will receive meaningful reports later. This includes computer information, installation and runtime statistics, as well as general and feature-related usage data.
Those who define questions together in advance do not have to search for answers for long afterwards. Who needs access to what information? How does the system support ad hoc analyses (Exploration & Discovery)? Telemetry is also an important question. Thus, the time of data transmission determines which data is collected at all. The setup of protocol formats, communication protocols, processes, synchronization schedules as well as caching and handling for offline use also helps to configure telemetry in detail.
If the systems are to be hosted internally, a scalable hardware infrastructure, software, backup, firewalls, anti-malware and other security measures must be planned. Even if the solution runs in the cloud, there are questions to be clarified and the cost-benefit ratio to be evaluated. This includes the number of instances required, the expected costs, scalability and security.
2. Client-instrumentation and telemetry
Another important task is the uninterrupted instrumentation of clients and the establishment of reliable, secure telemetry between client and server. To create unique user profiles, track usage trends for each installation, and link installation profiles to download sources or marketing campaigns, the system must be able to automatically generate user installation IDs and a so-called machine fingerprint.
The to-do list regarding telemetry is long:
Build a secure and streamlined client-server communication protocol with integrated security and encryption of sensitive data
Erweiterung Extension of the client-server protocol for handling proxies, firewalls, web filter gateways and other network configurations
Develop a method for generating unique customer IDs and tracking user/installation usage, even in anonymous mode
Development of an application logic for aggregating, compressing and optimizing data transmission
Set up an application logic to fix communication errors (e.g. ” Network not available”)
3. Data collection and management
Software Usage Analytics solutions typically collect and process terabytes of data. With relational databases and SQL queries, developers quickly reach their limits with this amount of data. Therefore, it is recommended to check the use of alternative databases and corresponding visualization frameworks, which are specifically designed for storing, querying and linking large and heterogeneous data.
The development of functions for collecting application-specific data is also complex. Web analytics systems such as Google Analytics or Microsoft App Insights often lack important features. In general, it is important to check exactly how much flexibility commercial solutions offer in adapting to company-specific requirements. In many systems, the selection of the elements to be tracked and recorded is severely limited. If data is then collected that does not create any real added value, costs can rise unnecessarily.
In-house systems are often relatively static. If the requirements change, it is difficult to expand the data collection afterwards without changing the client code and distributing a new build. Therefore, it makes more sense to choose from a remote control solution, where users can quickly and easily start or stop tracking certain metrics via a dashboard.
4. Data visualization and reporting
Once collected, the data must be evaluated and translated into interactive reports and meaningful visualizations. This task should not be underestimated. If there is no visual processing, manageability and context, even the largest data pool makes little use of it. The transfer of data into usable reports takes time and foresight in order to be able to answer current and future BI questions.
Central questions concerning visualization:
Implementing a modern dashboard that can be adapted to user needs
The implementation of a visualization frameworks, including API
The Design of high-performance back-end queries
The implementation of Reports, including Drill-Down functionality
The construction of a data aggregation logic for recognizing trends and patterns
The provision of functions for data export
5. Privacy and Privacy
Data protection plays an important role in software usage analysis. Failure to do so can cause significant legal and image damage to companies. Data security and granular access controls at all levels are therefore standard. In addition, there are comprehensive opt-in/opt-out options that allow users to restrict the tracking of their software usage or block notifications (e.g. in-app messaging).
As more and more countries tighten their data protection regulations, developers need to closely track and implement regional differences in the design of usage analytics systems. In some cases, data must be stored locally, ensuring that personal data does not leave a particular region (e.g. GDPR).
6. Robust and powerful environment
There are many aspects that affect the time-to-value, reliability, manageability and overall ROI of a Usage Analytics solution. Usage analysis promises real added value if the solution can be integrated into third-party applications such as download and auto – update systems, CRM, BI solutions and licensing tools.
However, integration is only possible if the data can either be exported in a processable format or synchronized via clearly defined APIs. In the case of in-house developments, the necessary interfaces may first have to be established. In addition, comprehensive testing is required to ensure compatibility, safety, reliability and performance.
These include the telemetry system, client-side SDKs on various client platforms, the infrastructure (using stress test frameworks and tools) and the application (including telemetry user interface, reporting and visualizations). And last but not least, as with all mission-critical systems, backup and recovery, automatic failover, 24×7 server and service monitoring, as well as precautions for peak usage and uneven patterns in network traffic are necessary.
Nicole Segerer (Image: Revenera)
Whether companies rely on in-house development for usage analytics is a question of time, resources, expertise and corporate strategy. However, it is clear that software providers need comprehensive, accurate and actionable information about how users use their software products.
Developer teams are under pressure to present new products in ever faster cycles without losing sight of security issues, budget constraints, costs, technical debt and data protection. Without a precise understanding of the requirements on the part of the users, it will be difficult to define a roadmap that will bring companies safely to their destination in the future.
* Nicole Segerer is Vice President of Product Management and Marketing at Revenera.