Intelligent Data-Driven Marketing: When Physicists Start Thinking about Marketing E-Book

Mathias Elsässer

• Herausgeber: ibidem
• Kategorie: Fachliteratur
• Sprache: Englisch

Seitenzahl: 315

Veröffentlichungsjahr: 2022

Beschreibung

Following Einstein’s sentence: “Everything should be made as simple as possible, but not simpler. If you can’t explain it simply, you don’t understand it well enough,” this book puts a spotlight on the complex marketing ecosystem from a physicist’s point of view. Today’s marketing world is overcomplex; CMOs face the challenge to transform their current target operating models towards a 100% customer-centric and data-driven way of working. A journey from good old mad-men toward math-men marketing. This book consists of three parts: The first part strips down the complexity of the marketing universe to the leanest frame of reference and then brings back the complexity, step by step, in single dimensions. Part two and three just follow these thoughts and provide a detailed description of 56 small atoms that can be used in a maturity assessment of your marketing. How to use them in a broader transformation concludes the book. In summary: An end-2-end guideline how to pursue and master the transformation from mad-men towards a math-men marketing operating model.

Leseprobe

ibidem-Press, Stuttgart

“Everything should be made as simple as possible, but not simpler. If you can’t explain it simply, you don’t understand it well enough.”

Albert Einstein

Contents

Introduction — #intelligentmarketing

Applying physics to marketing

Applying business engineering to marketing

The imaginary marketing lab, playing with data and thoughts

Gedankenexperiment 1.0

The Marketing Universe — our inertial system of reference

Starting with Audiences and Campaigns

Plan the activations via different impressions

Gaining a return on marketing invest

Be more precise — take advantage of segmentation

Close the loop — the advantage of data analysts

From measures to KPIs

Data-Driven Marketing: Measures, KPIs, and Dimensions

Share of …

Marketing Pressure

Cost per …

Conversion rate

Trends

Aggregated KPIs of unique

Audience centric KPI systems

Integrated Reach — a bit of both

Smart KPI — Reporting, Dashboarding and Data Science

Pillar 1: Standard Reporting

Pillar 2: Explorative Data Analytics

Pillar 3: Data Science

Gedankenexperiment 2.0

A house is not enough — strategic marketing planning

Different Balls — creative management & optimization

Smaller Segments — hyper-accurate targeting and look-a-likes

Multiple Players — paid, owned, earned

Throwing tactics — the media mix and attribution models

Multiple actions — goals, funnels and customer journeys

Mat Robo — trigger-based campaign automation

My bag is full — frequency caps and campaign maintenance

The Perpetuum mobile — sharing, communities, and influencing

Summary — Gedankenexperiment 2.0

Using more physics on marketing

Fluid dynamics — “reaching the perfect flow”

Harmonic oscillation — “everything is in motion”

The law of thermodynamics — “flow of change”

Solid State Physics — “building out complex systems”

Einstein’s Theory of Relativity — “There is one last thing”

The ma.tomics — Intelligent Data-Driven Marketing Framework

GW1 From counting actions to guiding audiences

GW2 Real-time and zero latency

GW3 Closed-Loop — Harmonic Oscillation

GW4 Mastering the Tech-Stack-Orchestration

GW5 Run Agile — Start Fast and Fail Fast

GW6 Enable the organization

M1 THE MARKETING UNIVERSE

M1.1 Get Your Fundamentals Right: Strategy, Process, and Systems

M1.2 Understand and Engage with your Customer: Audience and Content Management

M1.3 Take the right budget choices: Strategic Investment Management

M2 MARKETING TAXONOMY

M2.1 Align on a corporate taxonomy: Hidden Data Asset

M2.2 Embrace Change: Master Data Governance “Customer, Campaigns, Content, Asset, Price & Finance”

M3 STRATEGIC PORTFOLIO & PROGRAM MANAGEMENT

M3.1 Program, Audience, Content, and Time Management

M3.2 Marketing Budget Management

M3.3 Plan Alignment, Innovation, and Collaboration

M4 CREATIVE MANAGEMENT

M4.1 Creative Ideation — Design Thinking

M4.2 Creative Development and Delivery

M4.3 Creative Optimization

M5 AGILE CAMPAIGNS MANAGEMENT

M5.1 Be Fast and Flexible: Campaign Preparation

M5.2 Run Fast, Fail Fast: Campaign Execution

M5.3 Manage Trigger-Based Campaigns: Campaign Automation

M6 PROFILING — KNOWN & UNKNOWN AUDIENCES

M6.1 The golden record of your customers

M6.2 The “Known-Unknown” Asset of Audience Data

M6.3 Fundamental base work — Content Tagging

M6.4 Consent Management — Customer Buy-In

M6.5 Data Protection — Legal and Legitimate

M7 HYPER ACCURATE TARGETING and LOOK-A-LIKE

M7.1 Plan & Design — Strategic Audiences

M7.2 Build & Activate — Tactical Segments

M7.3 Reach is King — Accurate Targets and Higher Reach

M8 INNOVATIVE & OPTIMIZED PEO ACTIVATION MIX

M8.1 Paid (Media) Activations

M8.2 Earned (Media) Attention

M8.3 Owned Touchpoint orchestration

M8.4 Social and Voice — two special capabilities

M8.5 Planning Pricing & Trade Promotions

M8.6 Media Mix Planning and Optimization

M9 GET VIRAL — SHARING ECONOMY OF SCALES

M9.1 Tweets, Likes, and Shares

M9.2 AI Image Processing and Sentiment Analysis

M10 THE MOST EFFECTIVE FLOW OF ACTIONS

M10.1 Multi-Touch Attribution Models

M10.2 Large Scale A/B testing & 1:1 journeys

M10.3 Machine Learning — AI-Driven Optimizations

M10.4 Goal and Funnel management

M10.5 The Law of Action and Reaction — High Performing Machines

M11 CUSTOMER INTIMACY — DEMAND WINDOWS AND PRICING

M11.1 Find the Right Moment in Time — Demand Windows

M11.2 Find the best context — Demand Windows

M11.3 Find the Best Price — Data-Driven Pricing Windows

M12 CROSS DEVICE AND PLATFORMS

M12.1 Cross-Device Identification — ID Strategies

M12.2 Cross-Platform Profiles

M13 MEASURE REACH, SUCCESS & AWARENESS

M13.1 Integrated Reach

M13.2 Brand Awareness

M13.3 eCom — Direct to customer models

M13.4 Offline — Retail — Brick and Mortar Models

M13.5 General — Market Intelligence and Surveys

M14 CONNECT SALES

M14.1 Marketing Mix Modelling — Incremental Sales Uplifts

M14.2 Trade Promotion Effects and Pricing

M15 STAY LEAN AND SHARE — SMART KPIs

M15.1 Stay Lean: Measures, KPIs, and Dimensions

M15.2 Share Your Data Assets: The New Imperative

M16 EXPLORATIVE REPORTING AND DATA SCIENCE

M16.1 Get the Basics: Standard Reporting

M16.2 Explorative Insights

M16.3 Working Data Science

M17 PREDICT THE FUTURE

M17.1 Programmatic Program and Campaign Planning

M17.2 Next Best Action

M17.3 Text, Image, and Voice Recognition

M18 THERE IS ONE LAST THING — CREATIVITY

M18.1 Engage With Your Audience — Creativity is King

M18.2 Create Emotions — Storytelling and Purposeful Marketing

From Mad-Men to Math-Men Marketing

Works Cited

Introduction — #intelligentmarketing

It’s been a while since I embarked on this journey to write this book on marketing. Many people were directing the same question over and over again to me: why are you pursuing this? This was followed by the statement, “You are a physicist and not a marketeer.” So, why should someone read this book?

For me, the answer is crystal clear; it’s also the reason I started this project, which is based on sessions I’ve hosted in the past few years as a business consultant for data-driven marketing. It’s the feedback I’ve received — this was the reason! Over the years, I had a lot of positive feedback, and they all said similar things, such as: “Thank you, Mathias, for this fresh view on the marketing universe and the fundamental work on reducing the complexity to the pure and really necessary facts without blowing this all up with endless digital bullshit bingo.”

For this reason alone, I hope this book will help you gain fresh insight on marketing and the core principles that are still valid in this fast-moving, digital, and data-driven marketing world. This book follows my thoughts on how to reduce complexity to define the inner core of data-driven marketing. Later, it adds more advice and dimensions step-by-step, to express the real world in a smart framework.

Keep in mind while you’re reading each chapter, that people in a static position, fast-moving watches passing them run slower. This is Einstein’s relativity theory! We should build “run fast fail fast” data-driven marketing machines, allowing us to harvest a marketing universe in lightspeed and real-time.

So please, follow me on this journey of applying physics to marketing, and I hope that you enjoy it.

Mathias Elsässer

January, 2021

Applying physics to marketing

Let me take you back to the time I was studying physics at KIT (Karlsruhe Institute of Technology). It was a little while ago now, but I still vividly recall this wonderful, air-conditioned lab that was completely dark. It was on the upper floors of the “Physics-Skyscraper,” and was the only way to escape this unnatural “tropical” summers in the lower valley of river Rhine city of Karlsruhe. There was a rumor, that long ago, a former King of Germany sent his soldiers to fight for new territories in Africa to Karlsruhe since rain with temperatures above 23 °C is quite common there.

Long story short, it’s easy to imagine that I’ve spent a lot of my time sitting in this wonderful, cool laboratory that was void of sunshine at the end of my physics study. To be honest, the total darkness was also a good way to escape my professor’s ongoing “let’s have a Gedankenexperiment” lesson. He was young, had just relocated from MIT, and he loved to play with his scholars. I’m not sure whether he did those lessons to challenge us, encouraging us to learn, or it was so that he could feel far more powerful and smarter than us.

My diploma thesis was planned in two phases: the first one included setting up the laboratory and the experiment, so I could measure the unpredicted switch-off-on effects of semiconductor micro-lasers. I won’t go too in-depth on the subject matter since this is a book about marketing after all, but the effect I wanted to achieve (that I’d spent nearly a year of my life on) was the tiny lasers completely switching off in case you pump them up with a push of extra energy. The most logical solution — and the behavior that is expected — is to increase the illumination, measured easily as a more powerful output laser beam. With the huge laser pumped up through dozens of mirrors, we produced a tiny microlaser which finally ended up in an infrared camera to measure the output. It took me two whole months to set this up and calibrated. I had to do this before I could measure the first laser beams — including the picosecond short switch-off-on effect.

Once this was done, I started the second phase. The goal was to replicate the curves I’d measured exactly, and this was using the Maxwell equations. It was the advent of PCs and the institute I was working for had ten of the 80286 Hewlett Packard PCs — what unbelievable computing power back in the day! We decided to veer away from the standard to buy computer time on a big Grey mainframe for all the calculations. We started what we would call hyper parallel processing nowadays. Every evening, I occupied all ten PCs and started a joined calculation of my Maxwell equation simulation C++ program. The results were very impressive, and we found an answer to the switch-off-on effect. Are you curious? Here is the full solution (M. Elsässer, 1997): “Subpicosecond switch-off and switch-on of a semiconductor laser due to transient hot carrier effects.”

But I’m sure you are more curious about the reason why this marketing book begins with a historic account about a lab and the laser experiments that were conducted. Well, that’s an easy thing to answer. When I looked back (years later, might I add), it was these 12 months that included the most important information you need to know when you’re contemplating data-driven marketing.

Let’s start with the most important lesson we can take out of this case:

It is possible to measure the real world and to simulate it via predictive IT models

The word ‘measure’ in this context means focusing on the core measures and then start thinking about the further calculations (the equations) and rules behind what we call today KPIs and predictive models. The interesting information to glean from the short story above is that you can measure the pump-up power, the power of the laser output, the frequency of the laser wave, the outside temperatures (which was as mentioned this lovely cool and stable 18 °C in our case), and the one inside the tiny microlaser. All of these are certainly time-dependent and were measured on a femtosecond timeline — that’s it. Some people expected a complex set of hundreds of attributes that needs to be measured — well they were wrong, to say the least. We are talking about four to five time-dependent core measures and four Maxwell formulas, that’s all there is to it. This allows us to simulate an intricate physical experiment, and to predict the picosecond switch-off-on effect of semiconductor micro-lasers, due to transient hot carrier effects.

The other information that you would learn from the story is that:

A Gedankenexperiment could help postulate hypotheses to understand the rules defining the real world.

In most cases, there will be an equation to simulate the real world in various ways, which we now call “predictive models”. An example of the equation would be the Maxwell Equations in my thesis case. In terms of marketing (what you’re here for!) I will explain the model in the following next chapters.

Everyone should keep in mind that the first initial gut feeling and thoughts about the expected behavior of a complex system, more often than not, aren’t true. It takes time to understand the hidden rules in the system you focus on. The easiest way to get behind the miracles of complex systems is to reduce the complexity as much as you can. Reduce it to the fundamental core and, step by step, add in the isolated facts and dimensions later on so you can finally reach a theoretical copy of the real world.

Another thing for you to know is that the mythos of super-computer predicting in real-time isn’t true. In today’s world of artificial intelligence and machine learning, computer power is always limited. Independent from the newest innovation, the latest predictive models will consume the maximum available power for the moment to get at least a really short view of the future.

Finally, painful but true, there is always somebody out with a more powerful brain than you, which forces us to think twice to understand his thoughts and theories.

Applying business engineering to marketing

In the meantime, just six years after leaving KIT, I started a new job as a Project Manager in the software industry. To me, it was obvious that physics and some business experience wasn’t nearly enough to have a great understanding of the market and business models of big enterprises. I thought for a while and came up with a new plan. I was going to gain more experience and develop the knowledge and skills I already had by learning with a hands-on in international business management. I did some research and found that an executive MBA looked like the perfect fit for me. It would broaden my existing knowledge within physics, software development, and complex business processes. This aside, let’s bring the focus back on fundamental learning which will be applied to marketing in the following chapters.

The famous University of St. Gallen teaches us that enterprises can’t be changed in the classical way of “this is my new To-Be world, let’s move on there.” Instead, to run a successful transformation you have to follow the way of business engineering (Österle & Winter, 2003) to change your business so that it’s running parallel to your daily work.

It is only this that guarantees your cash flow will remain ‘alive’ as well as earn the money for that necessary change. In comparison, this is like switching the engine of the plane while it’s up in the air and traveling over the ocean. And of course, once you’ve started, there’s no going back!

Business engineering is analyzing and shaping the three worlds of ‘As-Is,’ ‘To-Be’, and the realistic transformation roadmap. This is completed using different levels of strategy, processes, organization, and platforms, as well as underlying data models. All of these are flanked by the soft disciplines of agility, change management, and the injection of new competencies to your existing organization.

Let’s have a look at the dance of change (Senge, 1999) in modern marketing departments. Historically, most modern marketing departments are built up by creative people in sneakers and raw denim blue jeans, and they’re dealing with campaigns, creatives, impressions, clicks, and GRPs. All of this is done through a marketing supply chain consisting of headquarters’ marketing personnel, plus their teams, local country responsibilities, and a final regional execution via agency networks on various media channels. Apart from the section of the organization focusing on paid advertising campaigns, it’s normal to see other departments involved with communities, events, sponsoring, trade promotions, retail, and sales support. Make sure you don’t forget the holy communication teams dealing with press releases and partner networks; often, they’re responsible for the enterprise-owned mobile apps, webpages, bots, and newsletters.

But market-led strategic change (Piercy, 2001) takes care about the view and emotions of your customer (B2B) or consumer (B2C) and this is why isolated departments that work on islands of proprietary processes, platforms and data-models no longer work, and ultimately, in the long run, will fail.

Enterprises that can engineer and transform their different departments in a closed-loop engine to plan, prepare, execute and control all paid, owned, and earned marketing activities and touch-points will gain a competitive advantage (Porter, 1985). This type of system is too hard for your competitors to copy; it will also aid you in staying ahead of your current competitors by designing perfect customer experiences and emotions. As a result, you direct your focus away from optimizing the single sale, and instead re-direct it into putting the customers at the forefront of your strategy and daily operations. It is an outside-in approach that burns down the classic setup of different departments and their regional country sub-teams, including the historic distinct processes of advertising, sales, and customer care.

The topic of transforming your marketing isn’t a “Should I do it or not?” question — yes, you should do it, no matter what. The only question a modern CMO has to ask himself is when and how does he want to enter the new world of intelligent data-driven marketing. It’s a given that in most cases, this journey will be long and arduous. If we look at the theories of disruptive innovations from Clayton Christensen (Christensen, 1997) these rules can easily be applied to the complex setup of big brands and their marketing departments. Instead of getting leaner and removing the historic ballast, high sophisticated frameworks of KPIs, and data science, there are new marketing clouds and a vast range of media channels like bots and programmatic activations, creative optimizations, and predictive models that get introduced. This all ends up in a ridiculously over-engineered setup. Smaller, leaner startup businesses, who are based on clear and easy processes, will enter the market and will do better than the giant corporations. This is because of their quick, agile ways of attracting their customers. Christensen calls this “imprisoned.”

After 23 years of helping dozens of global clients optimize the way they use marketing, it’s the perfect time for me to apply the knowledge I’ve gained and sketch out a general ‘best practice’ for closed-loop campaign management and intelligent data-driven marketing.

In various workshops, I’ve faced the challenge of explaining this to CMOs as well as implementing this in their departments and teams — a more challenging feat, I must admit. This is why I’ve reduced how difficult and complex marketing can come across as and teach only the core elements. This helps business owners and their teams understand it better, as this way removes the jargon and doesn’t confuse them. Months ago, I began to design this holistic marketing universe and break it down into tiny elements, which I’ve called ma.tomics. For me, they are the smallest elements of the marketing world. Each of them contains individual bits of information about data-driven marketing, and this can all be ‘stitched’ (or put together) to gain a better understanding of the complicated subject.

The first part of this book will explain what I’ve mentioned above more in-depth and applies physics to marketing with one fundamental Gedankenexperiment. This is to reduce the complexity we see today in all marketing departments. In the second step, we will see a model 2.0 that will help to bring the dimensions and constraints modern CMOs have to deal with back.

I will use these initial thoughts about marketing later on in this book, using physics, disciplines, and frameworks to explain marketing from a physicist who has more than 20 years of experience in marketing.

The second part of the book explains the ma.tomics frameworks which give you, the reader, a clear list of interdependent atoms to be implemented within your transformation program to change your company to a fully data-driven organization.

The book will end with the last part how to transform your current way of doing marketing. From Mad-Men towards a Math-Men marketing operating model.

The imaginary marketing lab, playing with data and thoughts

Some readers may be familiar with the way famous Albert Einstein explained his general theory of relativity — some may not, so I’ll explain to anyway. As this topic and framework is far too complex for an easy experiment in a laboratory, he started playing around with thoughts in a Gedankenexperiment about dark elevators. This is quite common in the world of natural science; Schrödinger’s cat gained some publicity — even outside the world of physics.

One of the most popular stories is the final exam of Danish first Nobel Prize winner Nils Bohr (Wikipedia, 2017). As a young physics student at the University of Copenhagen, Nils was once faced with the following question in one of his exams: “Describe how to determine the height of a skyscraper using a barometer.”

He replied, “Tie a long piece of string to the barometer, lower it from the roof of the skyscraper to the ground. The length of the string plus the length of the barometer will equal the height of the building.”

This answer angered the examiner, who then decided that Nils had failed immediately. However, the student appealed because the answer was indisputably correct, and the university appointed an independent arbiter to decide. The arbiter decided that the answer was indeed correct, but that it didn’t display any noticeable knowledge of physics. To resolve the problem, the decision was made to allow the student six minutes over the phone to explain his answer. For the first five minutes, the student sat in silence, his forehead creased in thought. When the arbiter pointed out that time was running out, the student replied that he had several relevant answers but couldn’t decide which one to use. He then began to explain:

“Firstly, you could take a barometer up to the roof of the skyscraper, drop it over the edge, and measure the time it takes to reach the ground, but that would result in damaging the barometer.

“If the sun is shining, you could measure the height of the barometer, then set it on an edge and measure the length of its shadow. Then, you measure the length of the skyscraper’s shadow, and thereafter, it’s a simple matter of proportional arithmetic.

“If you wanted to be highly scientific, you could tie a short piece of string to the barometer and swing it as a pendulum, first at ground level, then on the roof of the skyscraper. The height of the building can be calculated from the difference in the pendulum’s period.

“If the skyscraper has an outside emergency staircase, it would be easy to walk up to it and mark off the height in barometer lengths.

“If you wanted to be boring and orthodox, of course, you could use the barometer to measure the air pressure on the roof of the skyscraper and the ground and convert the difference into a height of air.

“But since we are continually being urged to seek new ways of doing things, the best way would be to knock on the janitor’s door and say, ‘If you would like a nice new barometer, I will give you this one if you tell me the height of this building.’”

In his youth, Nils played as a goalkeeper in soccer. On one occasion, his team was playing against a German side, with most of the action taking place in the German half of the field. Suddenly, the German team counterattacked, and a spectator had to shout to warn Nils about the opposition coming towards him, who was using the goalpost to write down a mathematical problem.

Right, enough of the stories and physics: it’s time to enter into the wonderful world of marketing by taking a look at our virtual lab, and setting up an ‘experiment’ to see how marketing works and how it can be measured.

Gedankenexperiment 1.0

The Marketing Universe — our inertial system of reference

So, we are finally beginning our marketing and physics journey; we will start by stripping down the marketing universe to its basic core.

One of the first things we need to consider is that even with the advent of intelligent and predictive models, data-driven platforms, etc. the underlying process hasn’t changed in any fundamental way over the last few centuries!

In the early days of mad-men marketing, the way they ran their marketing campaign by compiling and starting a campaign with a huge bout of creativity and innovation, then collect the impressions (how many people saw the advertisement once). Then, they just hoped that the result was an increase in sales, which would mean the campaign was a success. All the gained insights could be easily compiled on one sheet of paper for each campaign. Intuition and ‘gut feeling’ were the predominant feelings they wanted to instill into potential customers through their most successful channels.

Over the last 20 years, with the advent of the digital marketing space, it has become more common to go beyond impressions, GRPs (gross rating points — this equals how many times you reach someone in a pre-defined group) and market-shares. The new trend was to add more additional measures to the ecosystem. The most prominent in trend is the “click” and “like” that has been linked to multi-layered marketing and sales funnel. For decades, the conversion rate was the central KPI. For the first time, e-Com business models could attribute marketing activations to direct user actions. The underlying process didn’t change, you still had to collect data to input them in a nice spreadsheet and analyze the figures to find the most promising activation channels and creative assets. The main difference, however, was the replacement of the mad men’s ‘gut feelings’ by hard facts managed by math-men.

With the advent of audience-driven marketing platforms, we entered the world of intelligent data-driven marketing. This means we steered away from simply counting impressions and clicks/likes and these measures with events caused by individuals. Modern marketing clouds enable you to follow your prospects, customers, or consumers through how they navigate within the digital and partly non-digital ecosystem. This collection of audience-based data allows us to make the customer journeys a priority within our marketing and sales funnel, and to segment our target groups individually. Quite often, activations are automatically based on trigger events to push your target audience to the next level in their customer journey. The structure and the amount of used data have changed while counting the impressions of a campaign; they end up in a single aggregated number. Counting the single event of getting an impression easily can break through the barrier of millions of data points per campaign.

If the process hasn’t changed, we can imagine that this will in the future. A stable process or system of reference is perfect for every physicist; it allows them to strip down the complexity to the core and start thinking about the ‘what-if’ scenarios in a theoretical world. Welcome to the world of Gedankenexperiments.

The original intention of my Marketing Gedankenexperiment was to explain to some of my younger team colleagues about the underlying principles of marketing and the KPIs frequently used in this domain. These KPIs include impressions, GRPs, clicks, likes, and conversions. I realized that it also helped me to reduce the jungle of measures, KPIs, and mystic formulas, and provides a clear and easy model to explain the fundamental way marketing works.

Starting with Audiences and Campaigns

It all starts with the selection of a group of people we can reach with any possible interaction we can think of. Let’s call this our universe and it represents the number of people in our virtual laboratory.

We may repeat our experiment or have multiple ‘houses’ in parallel, so we should give this closed space a name — let’s go for the campaign. This means a campaign defines a virtual space (like a house, for example) in a certain amount of time. The door of this house is open and everybody interested within our defined marketing universe can see inside, and access it without issues. To sum this up, we have a huge group called the universe, which is our starting point and a sub-set of people within our closed space (the campaign house) that we defined. This strategic audience of interested people is now the starting point of our next steps.

Plan the activations via different impressions

If we want to interact with our audience, we need to have a way to reach them first. In my virtual lab, a bunch of balls looks like a good way to do this. This means before starting anything, we need to purchase a box of balls that we want to throw out to our audience. At first, there are costs involved for sure, but you can’t get balls for free. This is only for the moment though, as it will certainly change later on as we get further into our experiments.

For the next steps, it’s all about setting a goal. Now, there are two options when it comes down to this. The first is to ‘hit’ as many individual people as possible, whether the ball makes contact with their head, shoulder, front/back, or legs. We are solely interested in reaching them and getting their attention. The second option is to throw the ball in a way so the people can catch the balls. It gives them the motivation to perform the pre-defined action of “catching a ball.” Regardless of the goal that’s chosen, in both cases, we need to measure how many people we reached. In the second case, we also need to count the amount of successfully caught balls.

Gaining a return on marketing invest

So far, this has cost us a lot of money and we haven’t seen any cash flowing back into our pockets. Therefore, an innovative solution to this would be to begin selling baskets — it might sound bizarre but stay with me here. Since people are attempting to catch the balls, they will need a basket to put their balls in. So, start selling them, say in the corner of the house, then place a short disclaimer on each ball on where they can buy baskets — this is a clear call to action. As we have set a price for each basket, we get some money flowing back into our pockets. This is a return on our investment from our costly balls, which is caused by the customers trying to catch a ball and finally deciding that they should get a basket, and follow the disclaimer (the message on the ball) Let’s find a better name, something related to marketing, for our disclaimer — what about creative (or something a bit more general) branded content?

Be more precise — take advantage of segmentation

As we proceed with our campaign by throwing the balls, it may become clearer that there’s a difference between the success rate of caught balls with people of different height. Taller people are more likely to be hit in comparison to those who aren’t taller 1.70 m. So, it makes sense to build two targets: the “tall” and “medium-sized” people in the house. Let’s name them “tall” and “medium” segments.

The same principles can be applied if our goal is to reach as many people as possible by throwing the ball. If we cut the floor of our main audience into two sections — with the first half directly in front of us and the second half towards the back of the room. Based on this “front” and “back” segmentation, we can tailor our throwing technique. They would be soft throws in the front and straight long-distance strikes for the second segment further away from us.

Close the loop — the advantage of data analysts

To stay on top of the game and avoiding losing insights, we add a second player to the game who doesn’t throw any balls. Instead, they are responsible for counting the balls, money, and people we reached — I’d like to call him a real data analyst. Each number in the described game above is carefully collected by him. On a whiteboard, the analyst will update the numbers of his findings every minute (latency).

On the whiteboard, there will be a table; the rows are used for the fundamental classifications that I like to call dimensions, and the second part of columns is the ongoing updated part of the measures we’d like to count in our lab.

Dimensions:

The name of our house ®campaign

The people attracted by our house ®audience

The way we clustered them ®segments

The fact that this number is a plan or the real thing ®plan/actual

Measures:

The possible number of people who can enter the house ®universe

The size of each segment ®segment sizes

How many balls we have in our basket and how much we paid for them ® (planned) impressions and costs

How many we have thrown so far and the correlated cost of the balls ® (actual) impressions and costs

How many times we hit someone ® gross reach

How many people we hit ®net reach

How many times someone has caught a ball ®conversions/actions

How many baskets we sold ®sales amount

Our data analyst is a really smart guy! He always takes a photo of the board just before he updates the numbers at the end of each minute. The photo is then printed out and stuck on the wall behind the whiteboard in a linear sequence. At the end of the experiment, it’s obvious that most of the measures have a strong time dependency on each other. While the universe looks stable, the amount of gross, net reach, etc. is different on each picture in our time-sequence.

Well, that’s it as we can’t measure much more in our lab. I’m sure you have a few more things in mind but for the sake of easiness and clarity, I’d like to leave them here and open a second, more advanced, Gedankenexperiment 2.0 later in this book.

To summarize this experience, there are only two hand-full of measures and a very limited set of dimensions to describe a stable inertial system of reference for marketing. It’s important to mention in this summary that it isn’t possible to measure things like shares and the cost per ball because it’s all part of an overarching framework on top of our core experiment that uses the measures to go beyond pure reporting of the things we “see.”

From measures to KPIs

Therefore, the next step is the need to add additional brainwork into the game and combine the measures we have to, more powerful KPIs.

Data-Driven Marketing: Measures, KPIs, and Dimensions

If you have followed the Gedankenexperiment thoroughly, then you’ll see it’s clear that we have three distinct types of things we can use on our whiteboard to stay on top of what happens in the lab.

First up are the things we can measure and the second is our classification — the dimensions. They tell us when and where we have measured these values. Finally, the real asset is the key performance indexes (KPIs) we’re now able to calculate out of the measures.

Curious? Good! Now, let’s focus on the calculation and do some of the brain work I’ve mentioned at the end of the previous chapter.

Share of …

I’d like to start with some basics — often it’s much easier to compare shares of the total universe and frequencies instead of dealing with the actual number of people we reached (our net reach#) and the number of times we are hitting someone with our balls (gross reach#). The portion of the whole group allows us to deal with scales per hundred and, for this reason, comparison becomes much easier. The trade-off is that often, the reference group for the calculation isn’t obvious and may allow room for interpretations. So, should we report the percentage of the universe, the audience group of people attracted to our house, or the final tactical segment of tall people? Most of the time it’s the first one, but make sure you cross-check whenever you start dealing with percentage values.

Since we now have the same value in the form of a percentage value and the absolute value, we need to find a way to distinguish them, which is why I’ve added a # for absolute numbers and a % for the percentages of a defined total reference.

Besides the shares, we can also bring two measures into a relationship. In most cases, we will end-up in some kind of frequency.

With the measures universe#, net, and gross reach# we can calculate:

Marketing Pressure

If we don’t care about the fact that one person may catch more than one ball or if more people catch just one ball, we can then define another KPI that express some kind of marketing pressure we put on our universe. Again, we’re going to run the challenge to first decide which universe is the correct one, similarly to our % calculation above.

This new KPI helps us to compare several iterations of our campaign and we’ve called it “gross rating points (GRP).” It’s calculated by the following metric:

Cost per …

Now, let us factor in the cost and combine this with the number of balls (impressions) we bought, as well as the success we’ve seen in the form of sold baskets. To deal with lower numbers (normally for huge amounts of impressions per thousand (per mille) is used), this eliminates three digits of zero at the end of the impressions.

We will name the two new KPIs cost per mille (CPM) and cost per action (CPA). For CPA, we should always add the dimension of the action we are referring to.

Conversion rate