How to Use Data to Inform Important Business Decisions
We’ve all heard it a thousand times before… The amount of data being collected is growing at an enormous rate. In fact, it’s growing so fast that we’re now at a point where we are using quintillions of bytes every single day (1 quintillion has eighteen zeros).
Nearly every aspect of business is being affected by this trend. An increasing amount of data is being collected on customers, internal processes, employees, markets, products, etc. And not only is this data becoming more abundant, but companies are also shifting their culture to be more transparent with data. New forms of communication and advancements in data analytics are fostering greater transparency and allowing for greater accessibility to uncover and share insights.
So what’s the point? Where’s the power in having so much data readily available with tools to easily analyze and share it? The power is in decision making.
The power is in decision making.
People at every level of business are making decisions throughout their day, big and small. Should I use marketing copy A or copy B? Should we target customers with <100 employees or <1,000 employees? Is this feature worth implementing in the product? Should I respond to this customer first or the other one? What time is best to send our daily tweet?
When decisions such as these are informed by historical data, the results can dramatically impact the company’s value and bottom line.
In the era of data, it’s no longer sufficient to make decisions based purely on gut feel. Employees across the company need to be armed with adequate insights to compliment their intuition, so that they can use both to make confident and fruitful decisions.
Why should you use data to influence decisions?
No matter your role, there are inevitable applications in which data can be used to inform decisions that you’re making or influence the decisions of others. Using historical data to inform future results helps prepare for and predict different outcomes, which in turn minimizes risk. So the goal is to do exactly that. Use data when applicable to increase preparedness, minimize risk, and maximize the potential value in making a decision.
This will go a long way in improving your credibility and will help you to become recognized as a thought leader who understands the business and can be trusted to inform important decisions.
How can you use data effectively in driving decisions?
Let’s get practical. How can you put the mass amounts of data to use in making better decisions? Let’s walk through concrete steps with one of the examples mentioned earlier, “Is this feature worth implementing in the product?”
Step 1: Define the Objective
Before doing any work, take the time to define the objective you’re seeking. Be specific in what you’re looking to learn, why you’re taking the time to investigate, and how you intend to use the insights. This can end up being a more formal document, but in our experience we typically keep it lightweight so that we can move quickly.
Example: The purpose of this assessment is to determine if it’s a worthwhile investment for us to build Feature A into the product. We believe Feature A may produce significant value to our customer base as it will help them save time when managing their users and user properties. To make an informed decision, we need to understand both the costs to implement the feature and the value generated so that we can calculate a projected ROI and compare that against our other initiatives.
Step 2: Assemble the Data
We recommend first making a list of all the different inputs you’ll need to make the decision, and then you can go back and identify the sources where you’ll obtain the data.
Costs: historical time spent on product stories and features, engineering salaries, surveyed input from developers on projected effort, product architecture expenses, customer support costs, surveyed input from support reps on time savings
Value: revenue from customer base, surveyed input from account managers on projected value and customer risk, market & competitor pricing data based on similar features, CRM data on customer feature requests
Step 3: Analyze the Data
It’s time for the fun part. After you’ve done the work to gather the necessary data, you can begin analyzing the data to identify trends, spot outliers, piece together models, and work towards conclusions.
In our example, we may start by analyzing the development time spent on historical product stories and features. The goal would be to determine the engineering team's current level of productivity so that we can forecast how long it would take to build the new feature based on its complexity. And then from there we can apply salary and architecture expenses to obtain the base layer of cost estimates.
Our example has a lot of different moving parts; it’s more of a model that needs to be pieced together. In this case, our recommendation is to break everything down into small actionable parts that can be analyzed on their own. This will help to avoid overwhelm and keep your momentum moving towards an end result.
Another pro tip is to document all of your work and thinking along the way. You’ll thank yourself later as this will help with building trust in your modeling, answering questions, and serving as a reference later when you need to revisit or adjust the model. It can also help other users better understand your thinking to build trust in making a decision off the data.
Step 4: Determine Key Insights
As mentioned in step 1 with the objective, the goal of the analysis should be to draw conclusions that are helpful in making a decision. After you’ve analyzed the data, perhaps the most critical step is telling a story from the data. If you aren’t able to communicate the relevant insights from the data, then all of your work may fall short of influencing others.
In our example, the conclusions we are trying to draw relate to the cost, value, and ROI of implementing Feature A. To tell our story effectively, our goal would be to communicate those concepts as clearly and succinctly as possible. So as we wrap our analysis, we need to ensure we have a clear answer to those specific components as well as the methodology that led to our calculations.
In most instances, it’s also important for the person analyzing the data to form their own opinion on what decision to make. If you’re the leader responsible for making the decision, then this is obvious. But you may be an analyst asked to perform an analysis for a leader. In this case, you should expect that leader to not only ask you what the insights are, but also what your opinion is. So be confident in the data, in your analysis, and in your formed opinion!
Step 5: Communicate the Story
At this point, you have analyzed the data, extracted the key insights that you wish to communicate, and formed an opinion on the matter. Now it’s time to package everything up into a clear and cohesive story. Keep in mind that the goal is to influence a key decision, so an effective outline will typically hit on a few key themes: objective of the analysis, methodology, key findings, and recommended course of action. This outline will ensure that you keep the story succinct and to the point.
Depending on your role in the decision, you may or may not be presenting the material. If you are presenting, be confident to deliver a story that showcases the work you’ve put in. Here’s a helpful tip for how to set the appropriate context for the meeting.
If you’re passing off the material to another presenter, make sure both of you are adequately prepared to cover the information and answer any questions about the underlying data or modeling techniques. Here’s a tip on how to use speaker notes to come prepared to answer questions.
Step 6: Identify Iterative Improvements
In our experience, an analysis is almost never performed just once. It should continue to evolve and be adjusted over time to generate additional value to the business. To help answer future questions that cover the same domain, it’s helpful to identify some areas of improvement to make the analysis faster and richer the next time around.
Improvements may need to be made on the collection of the data to save time, make it more accessible, and improve the data quality. Improvements can also be made to the model so that it's more reusable, more well documented, and more user friendly for others to adjust. You may also receive feedback on the presentation of the data, in which case you can identify improvements to better package the story, which insights are most important to share, and which questions you may have been unable to answer that prompt further analysis.
We can’t overstate the importance of combining data with intuition to make business decisions. We may not always listen to what the data says, but it’s helpful to use it as a guide to help us think through risk, value, and potential outcomes.
For these reasons, we make it a point to create a company culture that encourages curiosity and data-driven thought leadership. Every employee across the company should be armed with the data and tools necessary to thrive within this culture.
This requires a self-service model where stakeholders can access the data they need, have tools that make it easy to perform analysis, and have the means to package it up into a compelling data story. If you’re interested in learning more about how you can use Superdeck to create this culture within your organization, check out superdeck.io