10 Tips on Effective Communication in Business Intelligence

Business intelligence (BI) is about leveraging data into actionable insights. This process involves data mining, processing, analyses, building the models, visualizing, and interpreting the results.

The process, of course, moves concurrently with efficient communication with stakeholders on whom we rely to clarify requirements. Eugene Klyuchnikov, senior data engineer, shares some tips to improve communication between BI and the different departments that work with them.

1. Remember that the customer comes first. This first tip aligns perfectly with the core value of any marketplace company. The only difference with GetYourGuide, is our customers are our business partners. If it's not an internal optimization, the only reason for creating a new model is to deliver more or higher quality data and hence empower the business with tools for better decisions.

Put it into practice: Each time you start something new, ask yourself — What is the potential business value? It’s good to remember that zero value is also acceptable (technical debt is real and has to be tackled from time to time), but a clear articulation of it is a must.

{{divider}}

Every project, request, and communication is an excellent opportunity to talk to different people, ask questions, and learn from their strengths and areas of interest.

2.  Always think 10x and beyond. The second tip goes in parallel with and is inspired by being customer-centric — and it would be a problem if it didn't. Even if there are no signs that the solution we are building is going to amplify. If it's a good solution, it should. I'm not a big fan of over engineering, but thinking in scale and leaving enough capacity for expansion is what makes an engineer a good one.

Put it into practice: Even if it's not a direct requirement, spend some time thinking about the system evolution under different circumstances: higher data volumes, more complex dependencies, and also cross-domain integration.

3. Think of every model as a simplified version of reality. Once we trim the frills and highlight what's important, the essential is all that remains. Hence, it's important to be specific about limitations. It's OK if the model doesn't work under certain circumstances, but it's way worse when it looks like it does.

Put it into practice: Think about your models' application area and communicate it to the business users as early as possible. Be specific about the questions that are beyond the model scope.

4. Ask yourself and others, So what? This question is an effective technique for collecting business requirements. While the original version may seem full of fluff: Hey, this new model is still missing some important metrics. — So what? Meaning: What's the point of adding them? I always recommend asking your business partner: Imagine, you already have the metrics or the model you're asking for. What's next?

First, this may encourage both sides to think about real implications and fine-tune the request. Second, it's always a great way of learning about real business processes. Finally, no question is a bad question — and this one often delivers some effective results.

Put it into practice: Well, start asking this or any other question that can help you to understand the implications of the models you build.

With that being said, every disagreement should stay within the original scope of the issue and never go beyond what you actually disagree on.

5. Stay on topic during disagreements. Most importantly, nothing can drive progress better than discourse in the form of thoughtful and productive disagreement. If two people always agree on everything when solving a problem, that clearly is an issue in effective collaboration. With that being said, every disagreement should stay within the original scope of the issue and never go beyond what you actually disagree on.

Put it into practice: Question the most critical ideas that matter. When in doubt, ask twice. First, listen, then talk.

6. Respect the experience of your peers. By default, unless we have evidence of the opposite, we should assume a high level of professionalism from our business partners and respect their experience. It's their call to decide what metrics to use and how to make the assumptions about their domain, while we take full ownership of the technology and data deliverability process. The overlapping grey area at the edge is where all the debates should go on.

Put it into practice: Clarify and demarcate zones of responsibility and respect the decisions of other people who control the area outside of it.

Your goal is not to demonstrate your knowledge and make the other party feel dumb, but to make sure that your partner understands the concepts.

7. Give context. Converting data into insights entails back and forth meetings and requests for clarifications from varied stakeholders. To ensure the effectiveness of them, it is pertinent to explain why we need this information. This is equivalent to tip 4 applied to ourselves: we do not request the data that doesn't add any value to what we already know.

Plus, clearly stating that lack of information on A is a blocker for executing B is a good motivation for productive collaboration.

Put it into practice: Accompany any information request with an explanation of why and when you need it and how you will use it in practice.

8. Speak in layman’s terms. We're often surrounded by people who don't know about data processing as much as we do — and they don't have to (otherwise, they wouldn’t need us). If we feel it's important to be on the same page, then educating all stakeholders on data should become part of our day.

It doesn't have to be full statistics theory, but explaining the problem in simple words or preparing a basic visualization of the pipeline execution often helps to align the views.

Put it into practice: Explain what you do, ideally using simple parallels from everyday life. Your goal is not to demonstrate your knowledge and make the other party feel dumb, but to make sure that your partner understands the concepts.

9. Communicate often and clearly. One of the most powerful ways to demonstrate respect is to keep in mind other people's needs. Being transparent, especially with shifting deadlines, is key. Providing business partners with regular status updates, being prompt in notifying them about unexpected difficulties, is a critical part of our day-to-day job. The rule of thumb: all our stakeholders should be aware of what's going on.

Put it into practice: Do not neglect the status updates, and don't hesitate to inform your business partners if you think you cannot deliver on time.

10. Learn every day. Finally, the core tip and the one guiding all others is to learn as much as we can. Every project, request, and communication is an excellent opportunity to talk to different people, ask questions, and learn from their strengths and areas of interest. Even if this won't come in handy, showing genuine interest in what our colleagues do is appealing.

Put it into practice: Don't be shy or afraid to ask questions that do not directly relate to your day job. How do they make decisions? What is their day-to-day job? How do they define efficiency? And lastly — How can you help?

If you are interested in business intelligence or data science, check out our open positions in engineering.

To hear more insights from Eugene, follow his updates on LinkedIn.

‍