Math for Marketers
Has it been a minute since you took algebra? Or have you had to refresh while teaching your teenager via distance learning? We know executives are looking to marketing teams for numbers, metrics, percent change, percent difference, and to figure out which numbers are significant to the company (and ultimately, how to replicate that kind of performance in the future). Whew, no pressure!
Good news, we took notes from our data scientist extraordinaire, and former high school Statistics teacher, Marjean Pobuda, during her Futr Form session and we’re here to share what we learned.
Math and Marketing?
Prior to technology and data driven marketing, marketers were mostly creative types tasked with coming up with the “next big thing.” However, as campaigns moved to digital, and we were able to better understand your future renters and residents, the skillset of a marketer must continue to adapt as technology changes. Marketing today lives at the right brain, left brain intersection. Yes, creatives still have a seat at the table, but so do mathematicians and data scientists. Today’s marketing is grounded in data, formulas, statistics, analytics, correlations, patterns, predictive modeling, and testing. The common denominator on that list? Math.
Marketers must brush up on their mathematical skills in order to deliver meaningful reports for each and every campaign your company launches. As Marjean pointed out, mathematics is one of the most powerful tools in a marketer’s tool belt. It enables us to make sense of data, and provides the confidence needed to make better decisions.
Looking for data-driven insights and measuring marketing performance to determine the next best steps for your property or community is an essential marketing task. In fact, just over 75% of marketers are reporting on how their campaigns are directly influencing revenue. In a year like this one, your marketing budget is under scrutiny, and proving that you’re getting a high return on investment from your team’s marketing campaigns is vital to the success of your property or community. Plus, understanding the impacts of your marketing efforts allows you to repeat your marketing team’s Oscar-worthy performance. Long story short, historical trends can’t give you the insights you need as a marketer to make timely decisions or to take your marketing to the next level. While this isn’t necessarily novel news, it’s important to understand why and how math matters for marketers. Let’s go!
The Myth of Correlation and Causation
When we are examining our dashboards and reports, it can be tempting to notice when a relationship exists between two numbers, and think of them as causing one another. Myth! Correlation does not equal causation.
In our daily lives as marketers we track campaign performance, monitor occupancy, and adjust advertising spend. We’re often aware that many metrics are changing, and as creative storytellers, we might be asked to make connections on what may have caused these changes. However, we need to be cautious about what else might be influencing these values.
Let’s break this down a bit more. Correlation measures the relationship between two variables. For example, you could measure the correlation between the time of the year and the increase in advertising conversion rates. Or, another example, you run faster mile splits on rainy days…Oop! Maybe that’s too anecdotal, but the point is that your speed correlates to the temperature outside. And unless you’re a sponsored athlete, you probably don’t have the bandwidth to do complicated calculations for your Strava stats.
It’s common to assume that because a relationship exists — or two variables are correlated — that two things cause one another. In other words, in our example case, advertising campaigns lead to more form fills and phone calls at the start of summer. Unfortunately, this is really hard to prove. There is a formula to calculate correlation between two numbers, but there is not a simple formula to calculate causation. In other words, it is really hard to know when one variable is truly responsible for causing another.
While your team is no stranger to the high and low leasing seasons for your metro, this year has opened our eyes to how factors outside of our control — such as COVID-19 or natural disasters — have influenced the usual patterns and trends we were expecting on January 1, 2020. Since proving causation is tricky, it means we must look at our dashboards a little differently.
The Myth of Meaningful Change
Each week, month, and quarter, your team is checking in on metrics. Again, you need to be careful. Just as our brains are hardwired to make up stories (causation), our brains are also hardwired to look for patterns and trends. We are likely to find a rhythm to the changes we notice, and develop reasoning that can inform and shape our understanding of how the campaigns are performing and why that is the case. A word of caution here: why. Think back to the correlation and causation myth we unpacked above. To understand if a change is meaningful, we need to use a couple of tools to help interpret the shifts.
Think about how frequently we look at averages. Ask yourself, how do you know when an increase or a decrease in a metric like this falls within our usual performance range, or when it is a notably larger increase or decrease? To answer this question, we must consider the metrics we are examining over a period of time.
Let’s start with standard deviation, which tells you how varied the data is from the mean. For our purposes, we can think of the mean as the average, which represents our ‘normal’. Mathematicians like to look at a large volume of data in order to quiet the noise from outliers, and to establish a sense of ‘normal’. Likewise, for marketers, observing a larger pool of data will dampen the effects of seasonality, which may influence what we observe in ways that are hard to measure or control. Now, with this large data set, we use the standard deviation calculation to determine how spread out the data is from the mean. Through this calculation, we are unpacking whether average is truly very ‘normal’ for a property or community. A large standard deviation says that most data is not close to the average. This means it is much harder to have confidence in what the average is telling you. Meanwhile, a small standard deviation tells you that most data is close to the mean, giving you higher confidence that your average is a good representative of ‘normal’.
Evaluating and understanding large quantities of data allows us to check for statistical significance. You have likely heard someone say that a finding is “statistically significant”. But, unless they have run a true statistical test, that is not just something to throw a guess at. Statistical significance helps quantify how likely an event is to happen, based on some given factors. For your day-to-day purposes, know that looking at a wider set of data is generally better than looking at less when attempting to make a prediction about the outcome of a campaign. But, looking for comparisons, benchmarks, standard deviations, and statistical significance can take your data-understanding of your campaign performance to the next level.
Need to calculate percent difference and percent change? Don’t worry! G5 data scientist and former teacher Marjean will help you with the basic equations you need to up your numbers game. Check out the on-demand “Math for Marketers” session from Futr Forum for the math refresher we all need.