For more than a year, Dagens Nyheter has been preparing and implementing a migration from one customer relationship management (CRM) system to another — from Adobe Campaign to Salesforce Marketing Cloud.

The migration also included setting up a subscription data model with data from our data lake. The model includes data points, most of which are needed for different kinds of segmentation. For example, the data model holds information about the subscribers’ subscription package, cancel reasons (if any), sales channels, opt-out preferences, start and stop dates, and much more. It holds both active subscribers and passive subscribers (up to two years).

The initial idea was to keep the data model as simple and clean as possible but still support the most important journeys in Marketing Cloud. But, the data model grew as more use cases were identified. We also added brand-specific functionalities when we adjusted the data model to fit the requirements of more Bonnier News newspaper brands.

Over time we also believe the subscription data model will have a wider usage than just supporting marketing automation. It will hold definitions and subscription logic that can be used when querying data from the data lake for various analyses. It’s also probable that subscription data model will be complemented by other data models such as a price model that holds logic and data for subscription prices, discounts, debiting frequency, upcoming debiting dates, etc.

Limited behavioural data is also a probable addition to the model in the near future. In that case, we will be able to act and communicate on decreasing engagement or special interests among our subscribers.

We are now live in Marketing Cloud, but we still have a lot of fine-tuning and ongoing improvement ahead. The new marketing automation setup has brought a lot of challenges and insights. This is a brief summary of some key choices we’ve made along the way.

1. Building a scalable solution

Dagens Nyheter is a part of Bonnier News. Massive efforts have and will be put into aligning and synchronising the data, systems, tools, and ways of working during the past years. This project is no exception. A scalable solution comes with both compromises and benefits, but we are certain it is the right way to go for us. The synergies overcome the potential downsides.

2. Gathering logic in the data model

Subscription logic, rules, and calculations could probably be done both in the MA system as well as in the data model. We choose to gather most of it in the data model. Hopefully, this will make us more flexible. Additionally, it will enable us to use the logic for purposes other than just marketing automation use cases.

3. Update frequency

The data in the model is updated once a day. This is because some of the underlying data sources update on a daily basis so there is no need to look for fresher data more often than that. The higher frequency, the more recent and relevant data, but also the higher cost related to scanning for updates. For now, we’ll start with daily updates and hope that it is enough.

4. Testing

Just because we are live in Marketing Cloud doesn’t mean our setup is done. We prepare for extensive A/B testing, fine tuning, and improvement of our journeys along the way.

5. Democratising the data

We need to make sure the data model is easy to understand and use. Well-defined and documented logic will be vital for the model to be handled appropriately and reused for other purposes.

Looking ahead, we see great potential in developing the data model further. Depending on what kinds of use cases we see in the marketing automation system, the subscription data model will develop and grow with the cases.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .