From wish list to decision
Most organizations looking for a learning platform end up with a long wish list and no way to turn it into a decision. This article offers a narrow tool: four questions about licenses, hosting, data privacy and exit options, with a simple scoring method to compare candidates and a set of questions to ask each provider.
It is a decision aid, not a ranking of products, and it assumes you have already written down what your learners and teachers need. Those needs decide what “good” means for you. Nothing here is legal advice, and your own circumstances may call for expert or official guidance.
Question 1: What do the licenses allow?
Licenses operate on two levels, and both matter.
The software license. Is the platform proprietary, meaning you rent or buy the right to use it, or open-source software, meaning the code can be read, run and changed under stated conditions? Open-source licenses differ in what they require of anyone who modifies and redistributes the code, so read the actual terms. Open code is not automatically free of cost, and proprietary software is not automatically worse.
The content terms. Who owns what you upload? Can you publish your own materials under an open license, such as a Creative Commons license, and does the platform respect that? Can learners keep copies of their own work? Some services claim broad rights over uploaded material in their terms. If you intend to share openly, check that the platform does not restrict reuse.
Ask the provider what rights it takes over uploaded content, and whether courses can be made public.
Question 2: Where and how is it hosted?
Hosting determines who is responsible for keeping the system running, safe and fast.
- Provider-hosted. Less work for you, but you depend on their reliability and decisions.
- Self-hosted. More control, with obligations for updates, backups and security.
Also consider the physical location of data. Some countries have rules about where personal data can be stored or sent. Check what applies to you through official sources.
Ask the provider:
- Where are servers located, and can we choose?
- What are your backup and recovery arrangements?
- How are security updates handled, and how quickly?
- What was the most recent significant outage, and what changed afterward?
For self-hosting, ask internally: who will do the work in two years, and what happens if they leave?
Question 3: How is learner data handled?
This question should carry weight for any platform that holds the names, work and activity of real people.
- What data is collected, and is each item needed?
- Who can see it, including provider staff?
- Is any data shared with advertisers, analytics companies or other third parties?
- How long is data kept, and can it be deleted on request?
- Does the platform support secure sign-in features such as two-factor authentication?
- What does the privacy policy say in plain terms, and does it match what the sales material says?
Where learners are children, expect stricter obligations and consult the official guidance for your country. For adults, informed consent and clear notices remain important. Collecting less data is often the safest design. Learning analytics can be useful, but they should serve learners and teachers, not become a goal in themselves.
Question 4: How would we leave?
An exit plan is the part most often skipped. Ask for specifics.
- Which formats are available for exporting courses, grades and learner records?
- Is the export complete, or does it omit structure, comments or quiz banks?
- Does the platform support open standards so another system can read the files?
- What happens to our data if we end the contract or the company closes?
- Are there fees for exporting?
Then test it. During a pilot, export a sample course and import it elsewhere. If the result is unusable, treat that as a finding.
A simple weighted scoring method
Scoring does not make a decision objective, but it forces trade-offs into the open and gives you something to explain to colleagues.
- Choose the criteria. Here we use the four above.
- Give each criterion a weight, with weights adding up to 100, reflecting what matters most to you.
- Score each candidate from 1 (poor) to 5 (strong) on each criterion, based on evidence you gathered.
- Multiply each score by its weight, add the results and divide by 100.
Hypothetical example. Suppose the weights are licenses 30, data privacy 30, hosting 20 and exit options 20. Candidate A scores 4 for licenses, 3 for privacy, 5 for hosting and 2 for exit. Candidate B scores 3, 5, 3 and 4.
- Candidate A: (30 × 4) + (30 × 3) + (20 × 5) + (20 × 2) = 120 + 90 + 100 + 40 = 350, so 350 ÷ 100 = 3.5.
- Candidate B: (30 × 3) + (30 × 5) + (20 × 3) + (20 × 4) = 90 + 150 + 60 + 80 = 380, so 380 ÷ 100 = 3.8.
B comes out slightly ahead, mainly through stronger privacy and exit scores. The numbers are illustrative only. If you changed the weights, the order could change, and that is worth testing: if a small change in weights flips the result, the two options are close and other factors should decide.
Treat a very low score on any one criterion as a possible veto, whatever the total.
Red flags
Vague answers about where data goes, refusal to describe export options, broad provider rights over your content, and pressure to sign quickly all deserve a pause.
Keeping a record
Write down the answers you received, with dates and names, because providers change terms. Review the decision after a year, and again if the provider is sold or changes its policies.