Industry lingo as the preferred common language as basis for collaboration in data science

HBR came up with an article directed at people in the business: “Better questions to ask your data scientists”(1).But is this the right approach?

As an analogy, when you go to the see a doctor, do you think it makes more sense for the doctor to talk your language or for you to speak in doctor-language (medical terms)?  

By talking in plain language, the doctor can uncover symptoms that you might not feel worth mentioning, that you might not feel are important or related. Or do you know enough medicine to be able to describe your symptoms properly and completely in medical terms?

I know I don’t know enough about medicine, and even less about medical terms. Call me a chicken, but I’d rather not get paracetamol/panadol for a headache that’s the beginning of a brain tumour. ‘cluck!’

To me, a doctor consults with you, the diagnosis is a collaborative process between 2 parties, but who bridges the gap in terms of language makes a huge difference in the outcome.  

Would you rather be treated for what you really have, or what you think you have?

Now, how does this relate to “data science”?

I go back to the Drew Conway definition of data science(2) which looks like this, but here the "data scientist" is shown as a unicorn:

 

A “data scientist” is someone who, on top of hacking Skills/Computer Science and Mathematics/Statistics skills has substantive domain knowledge. Basically, the “data scientist” should be able to speak to the business in business language. 

One of the things that I have learnt while working with clients from various organisations and industries is that, very often, the clients don’t mention issues they don’t think “data science” can help them with. That’s not because these issues are not important to them, but simply because they do not know that “data science” can help solve them; they do not know what they do not know.

Similar to the case of getting an accurate medical diagnosis that will help cure the underlying medical issues, I believe that the data science process is in essence consultative. The best outcomes are always from collaboration between the business and the “data scientist”. 

It is the role of the “data scientist” to understand where the client is coming from, dig deeper, ask relevant questions, know the data that is required to tackle these issues, and maximise the benefits the client can get. For that, domain knowledge is critical. (And of course that’s just a small part of the “data science” process.(3))

Furthermore, the interactions and collaboration between the business and the “data science” are not limited to the “initial diagnostics stage”, but through-out the whole data science process. Therefore a common language is very important to facilitate collaboration, and in my opinion, it should be the “data scientist” who speaks business language rather than the business speaking “data science language”.

(1) https://hbr.org/2016/11/better-questions-to-ask-your-data-scientists?utm_content=buffer2d70e&utm_medium=social&utm_source=linkedin.com&utm_campaign=buffer

(2) http://drewconway.com/zia/2013/3/26/the-data-science-venn-diagram

(3) http://thegatesofbabylon.blogspot.sg/2016_04_01_archive.html