Are your growth rates real, or real misleading?
True Business Days: An experiment in growth normalization
When it comes to start-ups, if you are not growing, you are already dead. Growth is a guiding light for current and potential investors, a linchpin for executive performance evaluation, and the end goal of business strategy.
Growth metrics can be used to measure everything from sales revenue to average session duration to daily active users. Every business has a set of key metrics that speaks to them with their own justifications as to why it suits them best.
At Wave, we only make money when our entrepreneurs make money, so one of our KPIs is payments processing volume. This is the money that entrepreneurs all around the world trust us to move from their customer’s hand into their own bank accounts. Growing this processing volume is critical to Wave, and we monitor those payment metrics closely.
On the surface, measuring growth can seem straightforward. You compare one month to the month before (Month over Month, or MoM), or compare consecutive years (Year over Year, YoY) or a Compound Annual/Monthly Growth Rate (CAGR/CMGR).
But all of these growth calculations can be misleading.
An obvious example: March vs. February will always look good, because March has 31 days and February has 28 (10% more days). This is of no fault of February, but of the nature of our unbalanced calendar year.
But not all flawed comparisons are that obvious.
In 2017, July had 5 weekends. In 2018, it’s 4.5. In 2019, it’s 4. If your business metrics on weekends are naturally softer than they are on weekdays (true for many businesses, and definitely true for Wave), comparing July over July can lead to some flawed conclusions depending on how the calendar happens to fall.
Holidays can create challenges for some businesses, too, like Easter which moves around tremendously, appearing in April in most years but occasionally landing in March.
Introducing True Business Days
To normalize quirks of the calendar and really understand how much Wave’s business is growing, we started working on the idea of a true business day.
The concept is rooted in the fact that not all days are equal in terms of the opportunity they provide. We went deep into an analysis of everything from long term fluctuation down to weekly seasonality, to better understand the distribution over time, and uncover the underlying growth in Wave’s business performance.
We created a model whose foundation is a base value for each calendar day, dependent on the day of the week it falls on. This initial value is then modified by a series of rules that were developed from findings of the analysis.
Using True Business days as a normalization factor, we could uncover Wave’s true growth, confirm our understanding of seasonality trends, and identify a few trends that we hadn’t foreseen.
For example: Though it had already been obvious to us that Wave processes more payments on weekdays than weekends (a 60% reduction), the model showed us that Fridays see 10% lower payment volume compared to the rest of the business week.
We also learned that the first Monday of the month shows a 15% boost in payments volume, even when it’s not the first business day of the month.
While the focus of this project was an understanding of our normalized processing volume, the outcome is a tool that can be applied to almost any metric, including our most highly visible metrics such as Top of Funnel and Daily/Monthly Active Businesses.
There are also opportunities for using this model for other purposes including forecasting. With each month having a unique true business days total, we can layer this knowledge into a larger more complex forecasting model.
As Wave continues to grow rapidly, we will continue to monitor the model for performance to ensure that it remains relevant to where we are today and where we want to be tomorrow.
Drilling deeper: Geographic fluctuations
Wave serves customers around the world, and it dawned on us early in the project that there was a marked difference in how seasonality affects Wave’s key markets. Filtering for geography and applying true business days gave some powerful insights:
Canada is affected by holidays 50% more than the U.S.
We even looked at proximity to holidays: Does the weekend before a holiday Monday perform worse than an average weekend? Turns out only the Sunday saw a decrease compared to average.
With a plethora of new insights into our seasonality, we are able to consolidate it into a model to create a true business day value for each calendar day, for any key market.
The specifics of Wave’s True Business Days model will certainly not apply to all companies. You will have to identify the external factors that affect your business, that require a distinct solution to the normalization problem. But the idea of normalizing your key metrics to understand true growth has untold value and should be a priority for any business. For the start-up where growth is king, it remains an absolute imperative.
Daniel Marco is a Business and Insight Analyst at Wave. Daniel’s work spans the data and analytic lifecycle including domain areas of data integration, business intelligence and analytics. With a passion for problem solving and elevating data literacy across Wave, Daniel delivers actionable insights in support of product development and business strategy.
