[Plugin development] Crafting a data model with your editorial team

This step-by-step is for developers who want to create a data model with their content management colleagues. There are both the technical and content angle covered on how the process of finding the best data structure looks like. If you include both aspects the resulting data model will be better suited to be utilized for content automation.

Content perspective

Get a colleague that has nothing to do with the data whatsoever and has decent editorial skills. Let that colleague write a piece of content that should be automated, without insight in the uploaded data. He may use any other sources of information, to write the content to the best of his abilities.

Technical perspective

Meanwhile, you acquire insight and knowledge about the current data structure and which goals it was built for and what components of your infrastructure are reliant on the status quo. From this information collect options and restrictions of the current data model. Where it can change and where it has to stay the same.


When both of you are finished solidifying their individual perspectives, continue to reconcile those two by matching the existing data structures with the required structure for the pursued content.

  • Conflicts are going to happen during that process. Each one needs to be reviewed from both sides, to find the easiest solution.
  • If you are stuck contact us. We can provide some experience on how to accommodate advanced data structures on our platform.
  • Keep transferring the different pieces of content and different iterations of your data model onto the NLG Cloud and start working on the automation. That should give you a good understanding where your data model is not matching the desired results.
  • This is a cyclical process. Repeat until both perspectives are completely congruent with each other.

The next step is to finalize your import to match the necessary data structure and composition for your use case.

Some hints on the implementation

Please note that even slight alterations in the desired content can have quite a big impact on structure and content of the underlying data model. Also content requirements and thus requirements in regard to the data structure are changing from customer to customer.

  • You have a great deal of freedom while adding fields to an existing data structure. For quick results, this is the most effective way to go. On the AX NLG Cloud, there is no downside of adding more fields to an existing data structure, so feel free to do so.
  • It is most likely necessary to switch between several predefined data models to create content from different perspectives.
  • That makes it also necessary to have some form of access management for the different import patterns.

When the data model is finished for the first time it should meet the following criteria:

  • The data model should be highly flexible, because future changes are almost certain.
  • The data structure should be easy to process on the AX NLG Cloud. As a rule of thumb, when you are struggling with the ruleset, your customer with lesser insight and experience regarding data structures will be struggling as well.
  • As stated earlier, a range of data models for different use cases are a necessity to assure a broader range of use cases, that can be implemented, utilizing your plugin.