When you finish this course, you'll be able to answer three essential questions.
How do I predict what will happen in markets?
How should I position my products? Which metric should I track? And what insight can I glean from them?
To answer these questions, we'll cover
How to size markets and identify their trends?
How to position products as part of the market segmentation process? How to work with strategic metrics to identify strengths and weaknesses within the organization?
Module 1: Introduction to Marketing Analytics
Chapter 1 - Marketing Analytics: Strategic Models and Metrics In this session we will give an introduction to marketing analytics. Upon completing this lecture you'll become familiar with definitions, drivers, and advantages of marketing analytics; definitions, styles, forms, and variables used in models; definitions, families, and dashboards of metrics. How do we define marketing analytics?
One broad definition states: marketing analytics as data analytics for
marketing purposes from data gathering to analysis to reporting. But that definition is quite vague, so we prefer the following definition instead. Marketing analytics is the state of techniques and tools that provide actionable insight. By techniques and tools we mean models and metrics. We'll define models and metrics later, but now, understand that models refer to decision tools such as spreadsheets, and metrics refer to key performance indicators to monitor the state of the business. In a way, models and metrics are analogous to common items found in an automobile. Metrics are similar to automotive gauges, and you see some on the left. They monitor the situation
and help diagnose problems. For
example, a driver could like down at her dashboard and notice that the temperature gauge was reading high.
Based on that information she could
diagnose that something was wrong with the cooling system. Models, on the other hand, are similar to global positioning systems, also known as GPS units. Models show a representation of reality to help you decide on a course of action; so does a GPS. Modern GPSs represent reality in that they show the roadways in your current location. They help you decide on a course of action by presenting relevant information so you can make decisions. For example, the GPS unit could notify the driver of snarled traffic ahead and she could decide to use an alternate route. As we'll show in the course, one should take care in how metrics are interpreted. For example, aircraft returning from battle in World War II showed significant damage from bullets, such as holes in the tail section. Military leaders interpreted this observation as a requirement to reinforce the damaged areas. Statistician Abraham Wald interpreted
the same observation differently,
recommending to reinforce the nondamaged areas. He stated that the observations included a selection bias. That is, it only included the aircraft that returned, not those that did not make it because they were damaged too badly. Therefore, be careful how you interpret data. Now we turn to several trends that are driving the adoption of marketing analytics in organizations across the globe. First we see the trend of increasing accountability for marketing performance. CEOs expect their marketing departments to improve productivity and reduce costs. Second, organizational emphasis on data- driven presentations is increasing the need for analytics to back up proposed new plans. Third, data is increasingly being stored online, also known as the cloud. With online data storage we get increased speed and convenience when accessing data. Fourth, many marketing departments have reduced resources, forcing them to do more with less,
making the most of every marketing
dollar. Fifth, thanks to initiatives to capture customer information, companies are awash with data. The big question is: What do we do with all that data? We analyze it of course. Marketing analytics offers us many advantages. First, it can help drive revenue, positioning the marketing department as a profit center instead of a cost center, second, it can save money. CEOs will no longer tolerate the old approach where marketing [ sent out 04:02] campaigns hoping for results. Now marketing analysts can predict the outcome of their efforts. Third, executives trust numbers. Analytics is thus a powerful tool to persuade executives. Fourth, we can even apply analytics tools to sidestep politics in some cases. Some CEOs do not appreciate marketing departments, especially if they themselves do not have a marketing and sales background. But virtually every CEO appreciates revenue, and analytics can deliver revenue when used
effectively. Last but not least,
analytics can encourage experimentation by allowing marketers to test multiple scenarios before proceeding. That way they can make their mistakes on paper where they have little to lose. We now turn our attention to models and metrics. We can define a model as a simplified representation of reality to solve problems. We simplify reality to make models easier to use. For example, we can examine an advertising effectiveness model. The graph shows that sales revenue increases as we increase advertising spending levels. That should come as no surprise, but we can see that the effectiveness begins to level off once we hit the spending level marked by the letter a. After that point we actually start to lose revenue. We can imagine that this level of spending would create a supersaturation effect where consumers are overexposed to ads and thus less likely to purchase from the company. The purpose of the model is to evaluate the
impact of its input variables. In our
case we can assess how advertising affect sales. Models help us make decisions. They provide guidance on our marketing actions. In our advertising effectiveness model we can see that we do not want to spend beyond the level represented by a, or we risk losing money. Models come in different styles. We start with styles. Styles tell us how the models are expressed. As we can see in the image, they can be expressed in three different ways. First, they can be expressed verbally. For example, in the advertising effectiveness model shown on the previous slide we could state the expression " Sales is influenced by advertising." Second, models can be expressed pictorially in graphs or charts. For example, we saw the plot of sales revenue vs. advertising spending level in the previous slide. Third, models can be expressed mathematically. In our advertising example we could state that sales equals a plus b multiplied by our
advertising level. We strive to express
models in the mathematical style whenever possible. Models also come in different forms. Forms tell us about the power of the model. At the lowest level is descriptive forms. Descriptive forms simply describe the marketing phenomenon. At the next level comes predictive forms. With predictive forms we can determine the likely outcomes given certain inputs. This is the classic what if spreadsheet exercise. For example, we might say " What if we increased advertising spending by 10%." With a model in a predictive form we could see the resulting sales level. At the highest level we find normative forms. Normative forms help us decide on the best course of action to maximize our objective given the limits on the input variables. Instead of what if, we can " Given x where x is a certain situation, what should I do?" In the course we'll discuss variables. Variables are defined as something that can change, such as advertising and
sales. Independent variables are a type
of variable whose value affects the dependent variable. Think of independent variables as the input to the model. In our advertising effectiveness model viewed earlier the independent variable would be advertising spending. Independent variables are classified as controllable or non- controllable. Controllable variables represent things we as marketers have some degree of control over. For example, marketers can control variables such as advertising spending levels, features of the product we sell, stores in which we sell the products, and so forth. Conversely, we have no control over non- controllable variables. Noncontrollable variables include areas such as customer age, economic conditions, and so forth. Dependent variables represent our marketing objective, which we can think of as output. In most for- profit companies our marketing objective would be
related to revenue, such as sales
revenue, units sold, profit, and so forth. In not- for- profit companies we could have related objectives such as donations generated, pledges made, and so forth. In this slide we review the terminology around the most basic of models, the linear response model. Here we have only one input variable, similar to the advertising effectiveness model we reviewed earlier. As usual, our output variable is revenue- related, in this case sales revenue. We express the linear response model as y equals a plus b multiplied by x. Where y is our dependent variable, here representing sales revenue, a is the perimeter called the y intercept, b is the coefficient to x which we can interpret as the slope of the line, and x is the independent variable which in our case is advertising. As shown in the graph, the y intercept is level of you when x equals zero. The graph also diagramically shows slope, defined as
the rise divided by the run of a line.
We'll be using this nomenclature throughout the course. We end this module by introducing metrics. We define metrics as business- oriented key performance indicators. Examples include sales per distribution channel, cost per sale, profit for sale, and so forth. Metrics help us monitor and improve marketing effectiveness, allowing us to take corrective action as necessary. For example, we could track marketing expense as a percentage of sales. If we're trying to maintain a constant spending level of 10% of sales, we could monitor it. If it creeped above 10% we could make a note of that event and reduce our spending levels to bring it back down to our target value. It is rare that companies only track a single metrics. Metrics families, which are groups of controlled metrics, are more common. These families of metrics offer diagnostic and predictive information beyond that provided by a single
metric. Examples include sales metrics
families such as sales per industry, sales per product, sales per month, and so forth. Metrics dashboards are groups of relevant metrics and metrics families displayed in graphical form. Marketing automation systems such as those offered by Eloqua, Marketo, and Pardot, as well as sales force automation systems such as those of Netsuite and Salesforce. com offers metrics dashboards as part of their core functionality. We close the module by looking at our check for understanding. Number one: Do you understand the definitions of marketing analytics, models, and metrics? Number two: Can you discuss the advantages and trends driving marketing analytics? Number three: Can you explain model variables, model styles, model forms, and the linear response model equation?