If they don’t respect your work, your impact to the organization is going to be severely restrained because they don’t trust that your work will actually benefit the organization. Managers and executives have to mitigate risk when deciding on implementing new projects. If they don’t understand your work, then the risk of your work just went up. If they don’t trust your work, then they are not going to use it or the scope of your future projects will be constrained to what they do understand, which likely won’t be the kind of work you thought you would be doing.
Take ownership of bridging the communication gap. Sure, you can keep hoping that your boss will improve their technical knowledge or you can keep searching for a boss or organization that does understand your work. But, before you decide that you have to leave, try to solve the problem. The part that is in your control is becoming an effective communicator between management and the data team. By honing your communication skills, you’re going to become so much more valuable because you’ll become a person who can translate business problems into data problems, solve these difficult problems, and then convey the results to all stakeholders.
To increase your impact, you’re going to have to become good at presenting the same analysis or data project to different audiences who have varying degrees of technical or statistical knowledge.
- Be sure you understand the business problem
Before you start an analysis or a project, make sure you and your boss understand what business problem you are trying to solve. Is the problem a symptom of a deeper problem? You may need to do some digging to get to the root cause, using a method like the “5 Whys” technique. Which problem in the chain makes sense to tackle?
- Set expectations
Make sure you set expectations at the start. Are the expectations realistic? If your boss wants something done sooner, let your boss know if the solution or scope of the problem requires more time or resources.
- Define what success looks like
How will your boss know that you’ve accomplished your goal? Determine whether there are metrics to check or monitor to evaluate the performance of your model.
- Check in periodically
Don’t just have an initial meeting, do the work, and finally report back at the end of the project. Give periodic updates to make sure that everyone still agrees with the goal and plan of the project. Communicating more frequently also allows you to clear up any initial misunderstandings and make adjustments as things in the business change that impact your project.
- Start with quick wins
Go for quick wins (low-risk, low-reward) or low-hanging fruits that are big wins (low-risk, high-reward) to build trust and credibility. As you get a track-record for delivering value, you’ll gain more leeway for future projects, especially bigger or more complex ones.
- Use interpretable models
Based on the problem, consider using a more interpretable model that lends itself well to explaining its results or predictions. Interpretable models give you great insight into what you’re analyzing and require fewer resources, which makes them a great starting point. Sometimes, it’s all you’ll need, but if you want to see if you can get greater accuracy with deep learning methods, your simple model will give you a validated solution that you can measure against.
Originally published on Quora.