AI Flavours

Things I wish I knew earlier in tech ▾

3 ways how to communicate better as a Data Scientist

Table of Contents

At some point I really started to believe that we are somewhat wired to miscommunicate with others and just couldn’t communicate better. Primarily, it was because of our various cognitive biases that force us to substitute conclusions when we don’t know straightforward answers. Or maybe because we often tend to grow up in an environment which educates us to be people pleasers and never someone who causes problems or is a bearer of bad news. 

Or maybe because miscommunication is easy (if it wasn’t some unnamed companies wouldn’t treat miscommunication as their business model).

But on the verge of the technical and non-technical worlds, there is a particular hell of miscommunication. On one hand, we have managers that can’t break the technical language barrier with a developer. On the other hand, we have developers turned into salespeople, frustrated that no one understands what they are doing. Of course I am exaggerating a bit on purpose here, but whoever didn’t at least feel a familiar discomfort in your stomach remembering some situations from your professional life, may cast the first stone (me included).

 

But not those ones, they’re too pretty.

 

So does this mean we are absolutely hopeless? Not necessarily. In this article, I show you how I try to be a better communicator by retrospectively contemplating on the work related situations in order to improve my reactions and actions for the future.

 

Pursuing different goals

 

This is in my opinion one of the major problems in tech. In such a dynamic world, as a Data Scientist I was always tempted by the new shiny tools and solutions and just desired to use them in any new project. And because we knew how to measure performance of those tools, we were always able to give multiple critical metrics to measure that could remain our key metrics for the whole project.

Wrong.

While we track a Data Science metric, we tend to forget that our client or partner may have different metrics in mind. We don’t measure how our metrics and theirs correlate. Even more, there are a lot of situations where after a discussion process we realise that the project itself is not even what we thought about. Moreover, the system error that may be just another mistake for us, for example this 1 incorrectly classified example, may cause a major stir in our client’s work and PR (someone heard about algorithms that discriminate?).

This situation happens, because we tend to think we understand what is expected from us instead of actually understanding. Instead of listening and asking questions, we focus on the first information that is given to us and right off the bat try to compress it into our frame of thinking and fit into our belief of what is right.

 

Questions to contemplate to communicate better

 

  • What will change between now and the future for the company when the project is successful? (see I am not asking about “what to do” or “what is the problem” – I want to understand the view of the world before and after, even if it is the most imaginary and unrealistic one)
  • Do I understand the full process of how this solution will be used?
  • Do I understand the needs of each person that will use it?
  • What would happen if the solution would fail? How would it influence the users?

 

Hiding behind complex terms

 

This is actually a two-fold problem, so let’s start with a question. Why do we like complex terms? I like them because I can appear smart, because they interest me and perhaps because I put time and effort at one point in my life to learn them. And because I often tend to surround myself with similar people that also like those complex terms and share my interests.

But this could give me a biased view that all other people care about exactly the same things that I care about.

I participated in many business presentations that included a lot of technicalities, solution details that were presented without a context, technical inside jokes. I also might have created a few in the beginning of my professional work, don’t judge me. But the problem wasn’t my learning process and doing crappy presentations so I can do better ones in the future – the problem was that instead of contemplating which parts could be misunderstood, I was simply proud of myself to be able to talk about such complex things.

 

This is kind of presentation slide I’m talking about.

 

Another side of the coin is that we often use complex terms as fillers. Didn’t make significant progress last week? Talk about simple things in a complex way. Stuck on a problem and don’t want to appear as unqualified? Fill the talk with complex terms. This unfortunately may lead us to an obstruction of reality, and give us a wrong impression of the progress. Don’t get me wrong – this doesn’t mean we should absolutely abolish using complicated notions, but it simply means we should contemplate more what, why and how we serve them.

 

Questions to contemplate to communicate better

 

  • What background does my conversation partner have? Create exact personas, with (sometimes even imaginary) background and experiences. 
  • Contemplate each information – what is the aim of conveying the information? (for example we can show the system architecture to show where it can cause problems in the future)
  • At which point in my life I obtained this knowledge? Is it a widely known one or very specialised?

 

Giving answers right away

 

No one wants to be seen as unqualified. Therefore when someone asks us about a specific piece of knowledge, especially when we are experts, it’s very hard to say “I don’t know. Let me come back to you with information once I check it out”. I had a period in my life when saying “I don’t know” wasn’t even an option – and I still struggle with it, as we all do.

But such behaviour may lead us to very poor decisions. First of all we think that by confirming our lack of knowledge we would lose a client or a conversation partner. This pressure may lead to accepting many terms or easy shortcuts that we wouldn’t normally accept, that may not even comply with our values. Moreover, I saw people sawing multiple algorithms together into a solution with a high risk of failure and saying it’s easy and straightforward just to sweet talk someone into entering a deal with them or to be seen as someone capable of doing a project.

Don’t get me wrong – I don’t think giving a raw, unprocessed thought in a conversation is bad. The difference is our attachment to this thought. Do we think this is an answer to someone’s problems and think it is inherently ours and thus important? Or do we simply let it out to be molded by us or someone else in the process? The more we attach ourselves to it, the bigger problem we have, because we are not open anymore to new perspectives.

There is however another side of the coin – what to do when confirming lack of knowledge is not an option due to cultural reasons, like losing face? This will be covered in another article specifically about this problem.

 

Questions to contemplate to communicate better

 

  • What is the worst case scenario that could happen if I didn’t know something?
  • What problems may be caused by giving the answer straight away? Track this answer. Can it significantly influence someone’s life (maybe even in an imaginary scenario)?
  • How can I differentiate between importance and ownership of my thoughts?

 

Conclusions

 

Those are just a few examples that come to my mind when I think about miscommunication in Data Science. Why do I think it’s important to talk about this? Because in my opinion communication, a clear and concise message, is a valuable skill that, just maybe, in a more and more automated world will be a stand-out skill preparing us for the future to come.

One Response

Leave a Reply

%d bloggers like this: