Balancing Urgency vs. Sustainability as an Analytics Team
In the dynamic world of Data Analytics, the need for immediate responses to data queries is a common challenge. Whether it's a last-minute request from a manager, an impending board meeting, or a sudden, previously overlooked question from the executive team- analysts and ground-level practitioners often find themselves caught in the crossfire of urgent reporting demands. While this can be frustrating for those who actually need to provide the analysis, we can't always fault executive / mid-level management for not giving advance notice on business questions.
So how can you prepare for and minimize these scenarios? And are there any appropriate times to push back on these requests?
Well, just like every response in data science – it depends. These scenarios often place team leads in a true test of their leadership abilities, underscoring the importance of maintaining strong relationships with stakeholders and possessing a deep understanding of the business model to facilitate swift action. But through adhering to well-established development and organizational frameworks, you can position your team to be exceptionally well-prepared for these situations.
Here is what I have learned on how to best navigate through these scenarios in my career.
Stop and Reflect. Does this metric already exist in another place?
Stakeholders are likely not aware of all the reports that your team has generated. When an urgent request comes in their inbox, their first inclination might be that the analysis is something that needs to be built from scratch. This is an understandable perspective, as leaders aren't responsible for understanding the ins and outs of your BI environment.
Take a step back, evaluate the needs of your leadership, and identify what existing business logic can be leveraged for the request. Often times, leaders are asking the same questions about the business, just with different variations, filter criteria, or against different populations. It may be an exercise of further explaining business context behind an existing dashboard, metric, or model.
The 2 Time Rule
Let's assume that the required metric does not exist anywhere else, and it is crucial for the organization to answer this ad hoc question immediately. So you and your team move forward with the request to pull the required data to answer the question. However during the following week, your stakeholder comes back asking for an update on the same reporting request, with a refresh of the underlying data. In this case, it could be an appropriate time to push back to answer this immediately and take the required time to build a more scalable solution.
In short, "The 2 Time Rule" accounts for the first request to be truly ad hoc which analytics teams can provide the answer in "the quick and dirty way". But any subsequent request means that it is recurring for the business & should be treated as a scalable project. In this case, its detrimental to all parties involved to not build this question into a scalable dashboard or model moving forward. This is important for the following reasons:
- There is consistency in how the business question is answered & updated daily
- Leaders can get the immediate answers that they need on a reliable schedule
- Analytics teams can use their skillsets to focus on more impactful work.
As a manager / leading contributor to any analytics team, it is important to scope work accordingly. Stakeholders may not know what the best tangible solution is to their problem (ie. an Excel Workbook, a Dashboard, a Jupyter Notebook saved as a pdf), nor is it their responsibility to. For many department leads who handle smaller reporting requests independently, Excel is the dominant tool – likely in an environment where underlying data is stagnant and needs to be exported from an underlying source system & manually manipulated to summarize a pivot table. Because domain focused stakeholders are often not working with advanced tooling, it can also be the role of the analyst / team lead to scope the appropriate tool to answer a business question.
Use This experience to Prepare for Future Reports
Once the smoke has cleared, find time to meet with your stakeholder after the report is delivered & discuss what prompted the immediate need. Through this, you should look to understand how you and your team can build reporting in the future to better prepare for these scenarios.
Think of related questions that might also need to be answered
Lets say that you have been tasked with understanding how many new customers have been acquired since the start of a new sales initiative. Your initial reporting request might have been to categorize each customer into their respective regions, or show a timeseries chart of cumulative daily customer growth. But whenever you provide an effective analysis, it sparks additional conversation amongst leadership (this is how you can validate that you're outputting quality work). But this also means that you should be prepared to dig into the "why" and greater impact of your initial analysis.
In other words- based on the outcome you just provided, what else might you think is relevant to its impact on the organization? And Im not just referring to adding additional dimensional fields to re-categorize the same metrics. But rather, what other perspectives are also important to providing full context around the business question your answering?
From this example, another thoughtful idea might also be to compare the new sales initiative against customer churn. You might show that new customers are being acquired from the new initiative, but how is this impacting the net growth of your customer base? You might also want to consider conversion rate. In which case, what is the average conversion rate for different customer segments? And finally, for customers that failed to convert, what are the reasons as to why they said they are not interested?
Of course, you cant flood your dashboards / reports with an endless amount of KPIs since you will not be able to effectively get a message across. However, you need to pick and choose how to frame the full context of a business problem as it relates to your stakeholders.
Get in the Mind of your Stakeholders
The better you can understand the concerns, roles, and perspectives of your stakeholders, the better you will be able to future proof dashboards or models for them. Its beneficial for all analysts to work closely with stakeholder teams in understanding specific nuances / challenges in their domain. For example, how does your operations team log orders, and what are the challenges in accurately calculating time to fulfillment for various SKU numbers?
When you can better understand your stakeholders role in the organization, you are more likely to ask the same questions as them about the business. I've worked with many great analytics managers who were great at scoping out what stakeholders would find most useful in a dashboard before they even asked for it. In one scenario, my team was preparing a product revenue report for a board meeting, where we were given very specific directions by Management on what visuals we were required to build. I asked one of my managers:
"How do you conceptualize / truly predict what the board is going to want to see on our general product update analysis?"
And their response (paraphrased, in short):
"Put yourself in their shoes and think about if it was your own money invested into our company. Our board spent $XXX to fund this new product initiative, so they want to understand the related success metrics to ensure that they will see a return."
The response may seem overly simplified, but I can assure you this perspective can easily be overlooked as an individual contributor. As an analyst, it can be easy to get more focused on the code behind the solution, answering the questions provided granularly and at face value. When we find ourselves in this perspective, its best to take a step back and holistically understand why we are answering a particular business question in the first place.
Get in front of "Company News"
Analytics managers should also be aware of new company initiatives and evolving competitive challenges. This can help teams identify easy opportunities to provide effective tools in Reporting solutions that might relate to "whats hot" right now.
Let's refer to a real world example – You might recall the implications of Apple's IOS 14.5 release from April 26, 2021. During this update, they implemented a new privacy feature on IPhones where a pop up message prompted the user with approval to track their personal data. Effectively, when IPhone users opted out of giving away personal information, this posed big threats to advertising revenue for tech giants like Meta (then Facebook), as it meant they couldn't monetize ads as effectively to each user with this blocker.
Now, imagine working in Meta'a data science team when this was first announced from Apple. You could imagine that Zuckerberg would have wanted some immediate answers on how this was going to effect Facebook's ad revenue.
Lets also assume there was an existing dashboard that showed average advertising revenue per user in different segments. After hearing this news, it would be beneficial to add a toggle / benchmark to that dashboard which shows the impact of how this IOS feature would be affecting those existing metrics. This way, you can scope out how to properly build this dashboard (hopefully before) it becomes an urgent request from management.
Conclusion
Mastering the art of balancing urgency and sustainability in analytics is not just about responding to immediate reporting requests but also about building a robust and forward-thinking data analytics strategy. By understanding your stakeholders, drawing boundaries to management when appropriate, and proactively preparing for future reporting needs, you can set your team up to drive impactful analytics. Embracing the dynamic nature of your organization and staying ahead of industry trends will also ensure that you are always prepared for whatever challenges may come your way, making your team an indispensable resource for data-driven decision-making.
