Defining the distance per dollar
In Tina's solution she defines the distance per dollar as
SUM(distance_to_travel) / SUM(monetary_cost)My first instinct was to instead do
AVG(distance_to_travel / monetary_cost)I've found that the two methods give significantly different anwers, which makes sense because my solution is calculating the distance per dollar for each trip, and taking the average of that per month, while Tina's solution is taking the total distance per month and dividing by the total monetary cost per month. My metric is looking at rides, Tina's is looking at totals across all rides.
I feel like these are two equally valid metrics, but I wanted to check and see if one of them is just incorrect. Any thoughts?
Hi Woodrow!
Thanks for reaching out!
I believe both metrics you described have their merits, and their appropriateness often depends on the specific context or the question you're trying to answer.
The total monthly distance gives a holistic view of the 'distance per dollar' for a given month, making it consistent for forecasting on a month-by-month basis. This approach is suitable when you're looking at overall efficiency or cost-effectiveness on a broader scale.
Your method gives insight into the average efficiency or cost-effectiveness of an individual trip. It will tell you, on average, what you can expect the distance per dollar to be for a typical trip.
Tina's method might be more consistent for monthly forecasting since it's based on aggregated totals. If the business operates with a focus on monthly or periodical figures (e.g., monthly budgeting or monthly planning), then this approach is likely more beneficial. If there aren't extreme variations in trip nature month-to-month, Tina's approach might be a better starting point for a naive forecast.
Nevertheless, I believe neither of the metrics is "incorrect." The best approach always depends on the specific analysis goals.
Thank you for sharing your method with the Community!
Best,
Ivan