In part one of this blog, we discussed embedded and platform learning analytics solutions, and how to choose the right type of system for your needs. If your organization is ready to move on to a robust analytics platform, the next step is to determine how best to implement it.
There are three main ways to implement your learning analytics platform:
1. Build Your Own Software
Pros
- Fully customized to your needs
- Complete control over code base
Cons
- Large investment of time and resources
- Maintenance and updates will be required
- May not be designed for extensibility or scalability
- Higher chance of security vulnerabilities
The main advantage is that you get a fully customized solution with complete control of the code base, allowing future tweaks and changes as needed.
On the other hand, creating your own solution can be very expensive and time-consuming. Once deployed, the system will continue to demand in-house resources for regular maintenance and updates. Another concern is that software built in-house can often be of lower design quality, which may result in a system that is not easily extensible or scalable over time. Or it may mean less secure software, since vulnerabilities are harder to spot without testing the system over a more diversified user base. If any of these problems occur, you’ll have to solve them internally, without the benefit of official vendor support.
2. Implement an Existing Solution
Pros
- Lower total cost of ownership
- Faster deployment
- Official vendor support
- Designed for ease of use
- High reliability and security
Cons
- May not be suitable for highly specific needs
- Less customizability over time
The second option is to implement an existing solution from an established software vendor. This path has the benefit of a much lower total cost of ownership, i.e., time and resources. Since you're not building from scratch, the deployment is smoother. Once the system is installed, official vendor support is typically available. The design quality of an established learning analytics platform is also likely to be higher, due to significant R&D investment by the vendor. The downside is that it may not provide the specific functionality you need. There is also less customizability over time since the vendor controls the code base. So if you pick a solution that doesn't properly meet your requirements, there is little you can do to change the software. So, be sure you select an analytics solution that has been designed specifically for L&D, and that’s based on a learning-specific analytics standard like xAPI.
3. Modify an Open-Source System
Pros
- No need to develop from scratch
- Complete control over code base
- Designed to be extensible over time
- Bugs typically fixed quickly
Cons
- Often not very user-friendly
- May not be designed for learning-specific analytics
- Lack of official vendor support
- Still requires some in-house expertise
The third option is to modify a freely available open-source analytics system. In a sense, this is a hybrid of the first two options, since it gives you full control over the code base without requiring you to develop the entire software from scratch.
However, open-source solutions are often not very user-friendly “out of the box”, which may increase user training requirements. They also tend to be designed for general use, and may not perfectly fit your specific needs. This could make the user learning curve steeper, and in some cases, might make certain applications impossible without customization.
Making the Right Choice
Deciding how to implement your learning analytics platform is a major decision, with long-term effects on customizability, operational costs, and usability. In addition, implementation paths vary in the time and expertise required to deploy and maintain them.