What IS a Data Network Effect?
First, what's a data network effect? Well, it's based on network effects. A network effect is when a product or service gains more value as more people use it. Think Facebook becoming more valuable than MySpace. A DATA network effect is where something becomes smarter as it gets more data. This is particularly relevant to machine learning and other AI algorithms. Generally, the more data on which they have to work, the better the results.
Think about your talent data - in your ATS, HRMS, CRM, etc it can be quite "skinny" (lacking in flesh!). It's also usually in silos, and it can be quite stale. This presents a problem when you start to use AI because algorithms do better when they have more to chew on. Some people argue that the volume and quality of data for ML is more important than the algorithms themselves!
Data Network Effects for Talent Data
If you are going to get the best value out of the looming wave of AI tools for talent and recruiting, you're going to need good data. This challenge is in three parts:
The breadth and richness of available data
Curation of data into usable forms and formats
Separation of data from algorithms to reduce the "black box effect"
It is the first two of these that give you a data network effect. By connecting available data (public and private), and keeping it a format that machines and people can both use, you improve the results of the algorithms. In our experiments, improvements in perceived quality have been up to 27%.
I say "perceived quality" because the test we did was matching candidates to jobs. In the short run perception is only way we have to assess that match. In the long run there will be more objective measures of course.
Of course having comprehensive, fresh data available wherever and whenever you need it has a lot of benefits - but data network effects are one you'll be needing to get on board with.