Managers nowadays are aware that data-driven decision making is key to successfully managing a team or business; however, with the sea of metrics to choose from, how do you know which metrics to focus on? Driver trees are a great tool to help managers determine this, especially if they are integrated into a management dashboard for easy access to the most recent data.
In this article, we’ll cover the following:
- A common problem with management dashboards and how driver trees can help
- What are driver trees?
- Structure of driver trees
- What KPIs can have driver trees?
- How to use driver trees for decision making?
- 6 steps for setting up a driver tree dashboard
A common problem with management dashboards and how driver trees can help
According to DataPine, a management dashboard is a tool used to present all important management KPIs in a single place, share insights with c-level executives efficiently, and empower management to make fast and data-driven decisions based on the latest information.
You’ll frequently find management dashboards that are built to show what happened, but not why it happened.
As mentioned in How to Learn Data Analytics, there are many types of charts and each has its own purpose, so it’s important to know the ideal type for what you want to display.
For example, bar charts show comparisons, histograms show distributions, pie charts show composition, and scatterplots show relationships between a couple or few variables; however, driver trees aim to show the cause of an outcome.
What are driver trees?
A driver tree is a graphical representation of the relationships between an overarching metric and the factors influencing its outcome. This helps analysts identify what specific factor or factors may have been the cause of an issue or improvement in a metric.
Structure of driver trees
The image below provides a general example of the structure of a driver tree, where an overarching metric has multiple lines branching out to its influencing factors:
Figure 1 – The structure of a driver tree
The area called “Level 1” shows the 2 factors that closely influence the main KPI, while the area called “Level 2” shows the factors that influence each of those factors. We could potentially keep adding levels indefinitely, but it’s important to be practical and focus on the few that are strategic.
Additionally, it’s important to note the arrow at the bottom of the graph. It indicates that on the left we’ll have mostly descriptive data, telling us what happened, while as we move towards the right we’ll see data that aims to show why it happened.
In the example below, we can see a driver tree that shows the factors that influence a company’s revenue:
Figure 2 – An example of a driver tree
In this example, we can see a 5% increase in revenue driven by similar increases in sales volume and avg. transactions per client. From the start, it is very intuitive to track what metrics influenced the increase. A data-driven manager will use this as a starting-off point to investigate further, for example, by sitting down with the team that owns the “transactions per client” metric.
What KPIs can have driver trees?
As shown in the four charts below, driver trees can be used for both quantitative metrics (e.g., time, value, rate) and qualitative metrics (e.g., customer satisfaction, employee satisfaction):
Figure 3 – Different types of driver trees
The above examples include a single level of factors, but rest assured that these driver trees can be extended much deeper for more granular analysis.
How can driver trees be used?
Driver trees can be used in many ways to create data-driven organizations, from getting everyone to understand how their KPIs will ultimately impact the organization’s goals to empowering operations teams with a framework for fast and effective problem solving.
1. Fostering a data-driven culture
Few things are as powerful for creating a data-driven culture as showing people how their metrics impact the organization as a whole.
For example, a CRM team with a KPI of increasing user purchase frequency by launching effective campaigns could use the driver tree in Figure 2 to see how an increase in their metric will in turn, increase the number of total transactions and, ultimately, the company’s revenue.
For an even stronger impact, this driver tree can be implemented into a dashboard so that the teams can see this with the most recent data to track their progress and make any necessary adjustments.
2. Driving focus of strategic initiatives
A strategy team aims to develop initiatives that will steer the company towards its goals. It is easy to come up with ideas, but can be challenging to know what to prioritize.
However, a data-driven strategy team designs and leverages driver trees to pinpoint where strategic initiatives need to be implemented for the highest impact, facilitating prioritization.
Also, a strategy team can set goals for teams based on a driver tree.
Going back to the example in Figure 2, if leadership asks that revenue be increased by x% in the following quarter, a strategy team should be able to use the driver tree to identify the best course of action. Should that growth come from price or volume? Identifying constraints will play a big role in defining this. It will be much easier to identify constraints as you move down the driver tree to more detailed metrics.
Below is an example of some strategic questions that can be asked to the owners of each metric:
Figure 4 – Using a driver tree for strategy
By asking these questions, a strategy team (or any leader) can assess where there are constraints and where there are “low-hanging” opportunities to prioritize.
3. Identifying bottlenecks in operations and other issues
By breaking down a KPI into its parts, it will be easier to identify where a problem is and fix it. This is a live operations team’s bread-and-butter and a driver tree live dashboard can be one of their best friends for ensuring a smooth-running operation.
6 steps for setting up a driver tree dashboard
Now that we’ve coved what driver trees are and how to use them, let’s see the general steps for implementing them into your management dashboard.
1. Select the desired KPIs, factors, and levels for your driver tree
Selecting the overarching KPI is mostly a strategic decision based on where you want to focus. This is made easier if you already have a north star metric that is your primary focus. If not, then ask yourself what is the KPI you get mostly asked about by your manager, investors or other stakeholders. As for the factors and levels, this will most likely be a tradeoff between what is desired and what is practical (technically feasible). Each additional layer will add exponential complexity to your dashboard development, so maybe start off with just a couple of layers, as seen in Figure 2.
2. Create an outline of your driver tree
An outline is basically a drawing of the driver tree you want to implement into your management dashboard, showing its KPI, factors and levels. Figure 2 is an example of an outline.
3. Determine priorities metrics
After creating your outline and sharing it with your analytics team for development, you may find that some factors are not technically possible to report on at the moment or to integrate into the dashboard. You may end up having to prioritize some over others.
4. Determine the time frame
Will you need to check this dashboard daily, weekly or monthly? Knowing this from the start will help you develop a dashboard that is aligned with your needs. Many times this can vary by team. For example, an operations team may need to see this daily or hourly, while a strategy team may need it weekly or monthly.
5. Select the BI tool that best fits your requirements
There are many tools you can use, but where should you start? If the data sources required to report on the KPI and factors selected in Step 3 are already integrated into your BI tool (Mixpanel, Looker, Tableau, etc.) then that could work and be pretty quick to set up, especially in the first two options. However, if some data sources have not yet been integrated, it may be worth building a first version in spreadsheets. This latter tool will also give you a lot more flexibility and control over how to visualize your driver tree.
6. Enhance your new driver tree dashboard with user or event properties
To make your new driver tree dashboard even more powerful, you can add functionality that allows users to filter by event properties (such as city, day-of-week, hour-of-day) or user properties (such as age group, language, activation month, subscription type).
Stay tuned!
In this article, it was assumed that you already had a handful of KPIs in mind for your driver trees; however, in startups (especially those in early stages) it is common to have to define KPIs from scratch.
In a next article, we’ll define KPIs from scratch for an example business, design a driver tree, and learn how to introduce it into your management workflow to put data at the heart of your team.
Subscribe below to stay in the loop!
