Time Travel for Metrics — Why Time Series Data is Key to Success
A Metric in Time Saves Nine.
Have you ever walked through a manufacturing plant or watched “How It Is Made?” We marvel at how pencil or French fry get made. I toured a McDonald’s fry factory, and over a decade later, it is still a career highlight. What was shocking was how few people there were in the fry factory. A guy was using the hose to spray the potatoes into the little trough that started the process, and at the end, there were people loading fries on a truck. It was a magical and elegant self-contained system. You couldn’t monkey with any part of the process without affecting the entire multi-million-dollar factory.
But what does this have to do with metrics? Metrics must serve the system, measure upstream and downstream services, and provide visibility. However, I have seen in business that we often chase nuanced analysis. We are human, and we are looking for the exceptions. Therefore, we confuse metrics with analysis. Let me say that again: we confuse metrics for analysis. I have done a ton of analysis in my day. It includes filters and assumptions influenced by the person asking for the data. I hate to tell you this, but even CEOs who are 65% are likely to make an acquisition higher when they are overconfident. They are not immune from cognitive bias hunting. Good metrics are defined and state what it is, like it or not. When metrics are defined and captured in a time series, they can drive business change. Time series data allows us to find the outliers in the data visually and not go cognitive bias hunting.
Examples of Time Series Data
Below is a list of common time series data points. Unfortunately, these will be too generic for most use cases, so ask yourself what you want to see and go with that.
- Sales by day by product
- Inventory by day by product
- Number of transactions by day
- Website Visits by Day
How do you get daily, weekly, or monthly time series data?
- If you would like to see it, most likely everyone else will.
- Define the metrics and consider how you would want to see them daily. NOTE: it is okay and often best to start with summary data like sales by day by product.
- Figure out who in your organization owns Business Intelligence or the standard dashboards.
- Ask them to take a daily snapshot of the data and have them put it into the dashboard. Don’t worry if the data set is small. I have been successful in asking for 30 rows of data a day. The key is usability and visibility.
- Do this with a couple of metrics, then show your boss.
What do you do with the time series data once you have it?
- Use your corporation’s standard delivery method of this data, such as email, a standard dashboard, etc. Add the time series data to this location. Your new data needs eyeballs, so go where the eyeballs are already looking.
- Include it in your QBRs and PowerPoint presentations.
- Reach out to other departments and share your time series data. You will inspire your colleagues, and they will add to your great work.
- Enjoy more time in your day because you are not wasting time on useless analytics requests.
Why is time series data so effective?
- Managers don’t have time or the brain space to understand what normal is. Time series simplifies this and normalizes conversations.
- It allows everyone to have a foundational view of what is happening.
- Time series outliers scream on a dashboard. These data points could be when the system went down. It could be when a supplier was out of stock. You must remember these, but it also shows how often these systematic outliers exist.
- Time series data provides a solid foundation to build on instead of the sand of analytical cognitive bias hunting.
Driving Change with Business Metrics
- Trends will emerge once you have tracked metrics for a couple of weeks (or preferably months).
- The patterns will take you on two paths. 1) Knowing the next steps or 2) Identifying the next missing metric
- Make a plan to drive the metrics through business change. Keep in mind that this will be a hypothesis and experiment.
- Expect a counter metric to go up. For example, if a company decreases inventory in a warehouse, the backorders will increase.
- Continue to experiment and drive change.
When it comes to metrics in business, they are undeniable when done correctly. Time-series metrics create a solid foundation and can help you become recognized by management and executives. These steps will empower you to find the people who can help you get this data into managers’ hands. Be persistent because a little time series can go a long way.
Berthet, V. (2021). The Impact of Cognitive Biases on Professionals’ Decision-Making: A Review of Four Occupational Areas. Frontiers in Psychology, 12. https://doi.org/10.3389/fpsyg.2021.802439