About King Context.

In short: on context, (streaming) (big) data & information architecture.

More info soon.

Feel free to contact me with your questions or remarks!

Assuming a closed world

Hardly anyone will disagree with the statement that no-one posesses all information and knowledge that is out there. Yet many software systems are built as if they contain or control a complete set of information. They adhere - often unknowingly - the principle of the Closed world assumption.

New systems and apps today do it differently. Smart (streaming) (big) data applications unlock new sources of data. Open datasets are linked and used to enrich enterprise data. Analytics derive new insights. It becomes clear to many that there is a virtually unlimited, open world of datasets waiting to be explored.

Enterprises first had data lakes and now evolve to use streaming data platforms. But still there’s often no Open world assumed when designing the new solutions using all this data. Wrong by design, already. And why would this be so important, you ask?

Take for example a Consumer Profile, composed for the purpose of omni-channel marketing campaigning. The customer order history files are combined with a stream of online (web, app) visits, albeit identified. Some visitors can be linked to known customers, and so their Consumer Profile gets a true value in the field isCustomer.

A lot of the stream of visitors however cannot be linked to an order. Especially in the very beginning of this new data collection effort this occurs frequently. If at this point they get a false for the isCustomer field, it is based on a Closed world assumption.

This example of a wrong deduction should be very clear. The isCustomer should instead read unknown as long as no order can be linked to a visitor. It seems obvious, but I’ve seen this very example been implemented incorrectly already a number of times. And I don’t have to explain the consequences, of course.

A lot of incorrect insights are also implemented in other, less obvious use cases. And some are also blamed to poor data quality, but are in fact the result of a wrong - often implicit - assumption. Or it’s the result of a data integration (or migration) effort that should have - but didn’t - considered that even for such an effort, the logic of the Open World Assumption applies.

In short: it’s a good practice to always assume an open world, because after all, it is.

Context: Bad customer data will sabotage your CRM initiative

The marketing team unintentionally sent the campaign to a lot of deceased customers because the ‘deceased’ data field wasn’t properly migrated to the new CRM application.

Get in touch