Simple framework for systematic approach to usage of analytical tools is proposed here.
So let me summarize the principles which help your team to learn quicklier than your competition.
The first example is from a commercial segment. Every right CSO, Sales Director or Account Manager wake up with the question "What Shall I Do to Increase My Sales?" The source of the data can be internal customer system combined with external data about potencial customers at the particular market - list of companies in B2B, list of segments with estimated size in population in B2C.
Bonus - Be ready to be sustituted by AI in data analysis
Well, this is slightly visionary.
- First, just summary. AI finds the patterns between sample inputs and classified outputs and apply the patterns on the new inputs to estimate the outputs. See more e.g.in my example of AI driven remastering of songs.
- Now imagine that you build the sample data to train AI by systematic capturing <business questions, source data question, business actions> triplets together with the reflextion, honest feedback "it was/wasn't useful tip". Year. Two. By all decision makers in your company. By thousands or millions of decision makers using some world-widely used cloud system like Salesforce.
- After a while, AI will suggest you the right actions to be done or at least hypotheses to be evaluated based on the your up-to_date source data from yesterday. Like recommend you,which particula customers shall be serviced first with what to gain the most revenue.
Can this one work? May be yes, may be no. I believe in it much more than "blind AI big data" approach where the huge amount of data are passed through AI algorithms and some magic is expected to happen. The quality learning data are available here. Like "Yes, this is s semaphore", "Yes, this is a bridge seen from the road" etc. as a learning data for AI in automonous vehicles business.
What do you think?
- Can maps work?
- Can AI based on maps work?
- Other ideas, how to improve this?