1. Why?
A common reason why projects don't turn out as you expected is due to the purpose and goals not being clearly defined. Ask questions like:
- What is not working today?
- How will the new system help us?
Having a consensus on why you are doing this and what you want to achieve gives you a clear direction and a good foundation to stand on.
2. List features and requirements
If the users are allowed to think completely anew (which is not entirely easy), what features and desires for the new system do they have? Make it an early part of the project to collect all the ideas and thoughts about functionalities - high and low - that you want to have in the new system. Make sure you use your words and naming conventions and that the definition of these is clear. Of course, a requirements specification is also needed.
3. Technology
Don't let technical issues become surprises during the actual implementation; carefully research the prerequisites and ask questions about operation, management, security, performance, how to access data, etc., before you push the button. Here, as a client, you must place great demands on your implementation partner to contribute. Does the system have any special requirements on the existing platforms? Make sure to get all the details beforehand.
4. Press play
It is now time to, together with your implementation partner, start! At the start of the project, roles, responsibilities and channels for communication must of course be established. Make sure everyone involved gets their questions answered before everything gets started.
5. Data modeling
Now it starts to get hands-on. Together with your implementation partner, it is time to model the data. How should the data be structured so that it becomes logical and easy to work with? We recommend that you:
a) clean and filter the old data,
b) determine naming conventions,
c) structure and model a new data model.
A central concept is the organization tree - how you see the organization is usually a good way of how the model should look. Translate the tree into a model with nodes where each node becomes a point where data is aggregated and where there is a responsible person.
6. Verify the data
As the models are finished, it's time to review the data to make sure everything looks good. Does the data look as expected? Verifying the data is important not only to see that it is correct, but also for those involved to feel accepting of the new system.
7. Educate the users
When the data in the data models is approved by the system owner, you should educate the users as soon as possible. It is a success factor for the project that the users quickly get into the system and understand the basics.
8. Start working in the system
Now it is the time for the users to apply their knowledge directly in the system. If possible, arrange for there to be "power users" available for quick help at the beginning. The implementation partner should have preparedness for second line support if these also get stuck.
