I came to marketing by happy accident. I followed my heart and walked away from being a technologist and towards being a marketing technologist. At the time, the word ‘marketing’ was an odd fit for someone who majored in applied physics, as I did. Today, it makes plenty of sense. In the almost twenty years that have elapsed since I became a marketing technologist, marketing has changed significantly. It has become one of the most technology dependent disciplines in which one can engage.
It’s not just that marketing has become technology driven. Marketers also find themselves at a place where accountability is paramount, where they must rely on data and analytics to extract insights and deliver results more quickly, where technologies like personalization and programmatic advertising are upending even modern approaches to engagement. Saying simply that marketing has changed is an understatement. Marketing is experiencing an evolutionary moment.
Any article you read identifying the qualifications for the “new” marketer lists characteristics like “digital-savvy” and “data driven”. There’s no argument from me on those. It can only help for marketers to be technologic ally and data savvy. More skills are needed, however.
There’s an old Saturday Night Live sketch where Chevy Chase, as a presidential candidate in a debate, remarks, “It was my understanding that there wouldn’t be any math.” Truthfully, marketing has always required some math. What’s different is, in many ways – with its growing reliance on science, technology, engineering, and mathematics (STEM) – marketing no longer simply needs marketers. To achieve maximum effectiveness, marketing needs marketing scientists.
Marketers and scientists do very different things. Marketers live to reach customers, to drive them to some action and, ultimately, to get them to purchase. Scientists’ goals usually revolve around experimentation that leads to incremental advances and discoveries. There is an understanding that even small steps forward can ultimately result in disruptive achievements. As is clear from concepts like A/B testing, marketing has been adapting some of the experimental approach to achieve goals. Still, more is needed.
What’s missing is an adherence to scientifically thinking about the processes involved in marketing. In the evolutionary metamorphosis being witnessed, marketing is transforming from a primordial business discipline to a 21st century, science driven iteration of itself that requires the approach of a scientist.
Aside: I’m going to be honest with you. I’m biased here. Again, I majored in applied physics. I read magazines like Science and New Scientist as much as I read marketing media. I like science so my thinking here may be rooted there. That being said, you’ll find my thinking sound and, depending on your degree of marketing induction, appropriate for where marketing is headed.
I assembled a panel discussion recently on effective marketing analytics. The panelists were marketing analytics practitioners from Fortune 500 brands in New Jersey. During the Q&A period, one of the attendees asked the panelists about the best tools for getting started with marketing analytics. One of the panelists responded that finding success with marketing analytics isn’t about tools. Rather, it’s about asking the right question at the outset. And, that brings me to the scientific method.
The Scientific Method
Scientists live by a proven methodology for experimentation widely referred to as the scientific method. Wikipedia defines the scientific method as a body of techniques for investigating phenomena, acquiring new knowledge, or correcting and integrating previous knowledge. Different sources delineate slightly different phases of the scientific method. I believe the most intuitive approach for marketers involves these steps:
- Ask a Question
- Research the Question
- Develop Hypotheses
- Conduct Experiments
- Validate Experiments
- Perform Data Analysis
- Extract Insights from Analysis
- Communicate Results
- Conduct Additional Experiments
Let’s go through these step-by-step.
Ask a Question
Marketers (truthfully, everyone) have a tendency to focus on how to accomplish things – which touchpoints, which tools, etc. In fact, how things get done is almost always very dependent on what you want to accomplish. Having a Lamborghini, for example, may allow you to drive super fast but that does you little good if what you need to accomplish requires crossing an ocean. Getting started with the scientific method means asking a foundational question about what you and your team want to learn/achieve from your efforts. This question guides what you do and can always be returned to as you think about how to achieve your goals. Examples include, “Will this campaign achieve a 12 percent lift in sales?” and “Do post-purchase exposures to our brand impact customer loyalty and improve life time value?”
Research the Question
Answering the question should occur during the experimentation and analysis phases of this methodology. It is important to be armed with information that allows you to make educated decisions about how you approach the “experiments” you’ll use to answer your question. Scientists are interested in approaches that have been tried by their peers in other places. They want to be aware of contemporary techniques and tools for conducting their type of research. Moreover, they want to have some expectations about what is realistic given their resources. The same should be true for marketers.
As you consider your question, do some discovery among your team, among your peers and in trade magazines to assess what others in your field are doing along the same lines. You should absolutely consider doing some primary research that helps you establish a baseline so you have something to compare your results to. Additionally, now is a good time to give some thought to the tools and techniques you should have at your disposal. It is important to think about them in terms of the timing of your program as well as the human and financial resources at your disposal. Your peers may have had more or less time and capital available to them when they ran their campaigns. Your technical approach should be mindful of what is achievable with the resources available specifically to your team.
I would argue that more than a few high school students have been turned off of science by having to create a hypothesis in a science class – only to be told by a teacher that their hypothesis was somehow inadequate. Let me assure you that 1) I won’t judge your hypothesis and 2) your hypothesis should be simple. At its root, the only requirements of your hypothesis are that it be related to your primary question and it should be testable. Testable means, at the end of the experiment, your hypothesis should either be proven or disproven.
For marketers, one way to think about a hypothesis is that it represents a way to consider how well individual components of your marketing mix worked. For example, if you intend to use email as part of your campaign, then your hypothesis might be, “A series of emails delivered over the course of three weeks targeting customer who have purchased a product within the last year will contribute to a 4% jump in existing customer purchases.” At the end of the campaign, you can consider whether email had the affect you expected and, if not, what you might do to improve its impact.
It’s probably best not to tell your CFO but the truth is, every campaign marketers run is an experiment. Some campaigns are tried and true and are all but guaranteed to deliver some amount of ROI – just like the old vinegar and baking soda volcano science project. The novelty of that volcano project wears off pretty quickly and it’s been a long time since it won 1st place in any science fair. There are just too many other more contemporary demonstrations of scientific phenomena to draw peoples’ attention. Marketers face a similar challenge.
There’s a place for what you know works. Given today’s changing marketing climate, there isn’t a lot that always works. Success demands trying new things. For marketing scientists, this means building campaigns that test the hypotheses you develop in ways that push the envelope on what you may have done before. The type of standout accomplishments marketers seek to achieve require going beyond the marketing version of the science fair volcano and developing tactics that can deliver both strong results and insights for progress.
The concept of agile marketing is spreading. It is derived from the agile methodology used by modern developers. Back in the old days of computer programming, computer developers followed the waterfall model – a series of project phases that ran from inception and discovery through to launch and testing. Waterfall allots plenty of time to prepare, build and test the project. The abundant time spent also introduces significant risk as the requirements of the project can change over the duration of the project’s execution. Agile methodology radically alters the approach to executing projects.
One of the foundational components of the agile approach is the idea of iterative execution. Rather than developing software and collecting feedback once the application is completed, agile insists on building a series of increasingly complete iterations. Each iteration is presented to customers, feedback gathered and that feedback is integrated into the next iteration. In other words, the experimental validation is built into the process. That may sound challenging and it can be when starting out. In practice, however, an agile approach frequently leads to leaner, more timely, more effective software. It means the same thing for marketing.
Agile marketing means making incremental experimentation and testing a fundamental part of your process as you run a campaign. Rather than sending out your emails to your whole list at the same time, you might A/B test subject lines and/or your copy. You could try sending your direct mail packages in batches with different envelopes/mailers to test groups. Then, judge response rates before deciding the ideal configuration for the next batch. It can be tedious and require more attention than simply send-it-and-forget campaigns but the payoff can also be higher. It’s worth the effort.
The key to the incremental gains that lead to disruptive discoveries for scientists is the willingness to go through the cycle of performing an experiment, checking the results and adjusting the parameters of the experiment to improve outcomes. Certainly, many marketers do some of this already. It hasn’t become a fundamental part of the discipline yet – though things may already be headed in that direction.
One of the most critical dependencies on this cycle is results validation. Lots of factors can impact campaign performance. As you consider your campaigns, it is important to gauge how well your results align with accepted normal values. If normal response rates are 7-12% for your type of campaign and you’re seeing 20-25% response rates – regardless of how awesome and your team are – it begs the question, why are your response rates so high? Data validation encourages you to confirm your assumptions, verify your list, review your creative, double-check your calculations, and ensure proper delivery. After all, errors missed now negatively influence your data analysis and, ultimately, your conclusions. Now is the time to validate so you can avoid the risk of moving forward without questioning the integrity of your data.
Perform Data Analysis
We live in a magical time. We’ve never had a lack of data. What we didn’t have was the computational power and tools to analyze that data. That has changed and, seemingly, analysis tools abound. As I mentioned before, however, tools and technologies should only be examined after identifying the questions you need answered.
For analysis, the questions are rooted in what you need to know to make intelligent decisions about the execution of your campaigns. More to the point, your questions in this phase relate back to make determinations about your hypotheses. You want to answer questions that relate to whether your hypotheses are being proven or disproven and whether there are actions you can take to improve your outcomes. Once you’ve done that, you can turn to tools and techniques that allow you to analyze your data and make correlations to your execution.
Extract Insights from Data
Analyzing data and extracting insights are two very different things. Anyone can analyze data and tell you what the data says. Just like with people, however, what the data says and what it means may not be the same. Extracting insights means turning the analysis of your data into meaningful knowledge that influences how you perform future experiments.
For marketers, knowing some percentage of your audience took an action during a specific timeframe is valuable. It’s why we focus so much on metrics like response rates, churn rates, click-thru-rates and conversion rates. There is lots of information to support those concrete values. What’s less clear is what that information means in context (i.e. is the rate good or bad compared with others). Additionally, what are the less than obvious determinations that can be made by considering that information as part of a set rather than disparate data points? Most importantly, how does the information change what you do next?
Communicate Your Results
I’m a big fan of stories. When I think about presenting information, I think of it terms of telling a compelling story. “What’s the story I want to tell,” is what I ask myself. It’s because conveying information has to be about something relevant to the audience. The story you tell and how you tell it is especially important to achieving the greatest value from the scientific approach.
There are three keys to communicating you results effectively:
- Know Your Audience
Occasionally on the Internet, in doing some research, I run across a scientific article that looks interesting. I read the abstract/summmary, my interest is piqued and I decide to read the full article. Eventually, I realize I’m in over my head. The article just was written for me.
When presenting your results, you must tailor your presentation to your audience. Your senior executives likely want different details than the peers within your team or in the audience at the panel discussion you’ll be participating in next month. To ensure you convey what your audience wants to hear, avoid making the same presentation to everyone and put yourself in the shoes of your audience so you’re thinking more concretely about their needs. This brings us to the next section.
- Only Tell What Needs to Be Told
I have a three year old daughter. She tells everything. If she knows it, so does everyone else. When you consider what you’re going to share with your audience, it’s best to be more discerning than my daughter.
Too much information runs many risks. One is, too much information can be distracting. By presenting more detail than is required, you draw your audience’s attention away from your highest priority content. They end up missing your entire point. Another risk of an over abundance of content is, more information than is necessary can lead to more questions than you may be prepared to answer. Certainly, there’s a balance as too little information can also result in questions. The key, again, is to focus on what is relevant and leave extraneous information to your notes or as appendices that support your primary delivery.
- A Picture is Worth More Than a Thousand Words
If we’re being honest with ourselves, we already know people aren’t big on reading much these days. This post is too long and I know it. People want short-form content focused on a quick top ten most creative ways to accomplish some goal. Even then, the content should be accompanied by easily digestible visual imagery. Disseminating the results of your experiments should be supported by the liberal use of images that conveys your key data points.
Those images can come in many forms. Graphs created in Excel or data visualization tools like Tableau can communicate important information at a glance. Placing a small amount of copy over a bold image allows the image transmit your big idea while offering just enough detail with the copy to tell why it matters. You might also consider using images of your creative to tell the story; frequently it can make your data more relatable. These approaches can be used individually or in tandem to make communicating your results more impactful.
Conduct Additional Experiments
Communicating your results does not end your journey with the scientific method. Sharing your data, analysis and insights with your peers, your superiors and the greater community is part of the process to help you learn and to keep you honest. It is likely that presenting your results will lead to suggestions and questions that require additional research.
The key to what makes scientists so successful is the iterative approach to their experiments. Even after sharing their results at conferences, scientists run experiments again using new experimental configurations that have the potential to reveal new observations. Again, it is this incremental progress that leads to disruptive discoveries.
For marketers, there may be a need to return to the Conduct Experiments phase to prove a formerly disproven hypothesis or to more concretely answer the question at hand. By trying your campaigns again and integrating the learnings gleaned from earlier iterations, you can discover approaches that have the potential to radically improve your marketing effectiveness.
The scientific method has played a pivotal role in the ability of scientists to repeatedly identify phenomena and discover new materials and techniques that continually improve our standard of living. It can play a similar role in marketing when used as a guiding principle for marketing processes. The natural question is, what does the scientific method buy the marketer that warrants replacing your existing processes? For me, there are three clear benefits:
- Progressive Innovation
As I pointed out earlier, science changes through incremental advances. In fact, there’s an innovation principle, advocated by the author Steven Johnson, called the “adjacent possible” that suggests innovation can occur by continuously considering the next unrealized iteration of an idea. The scientific method’s commitment to iterative experimentation is designed to foster innovation that occurs in steps. These steps are key to achieving the disruptive innovation that leads to exponential gains in performance.
- Applied Learning
Because the process is designed to build upon what happened in previous iterations of your experiment, improving performance means building upon what you learn in each successive execution. The drive towards proving your hypothesis almost forces you and your team to apply lessons you learned in earlier campaigns to increase your likelihood of success.
- Repeatable Successes
What the scientific method helps scientists do well is prove their success isn’t the result of happenstance or error. They are then able to develop models (equations) that help replicate their successes. Marketers can use the scientific method for the same purpose. Once your hypotheses are proven, you should be in a position to accurately model/explain what works for a given audience. Marketing isn’t as concrete as science so your model may prove more or less effective over time. It is effective, however, for providing your team with a proven approach for repeatedly achieving a baseline of performance that can be bolstered by other programs.
These benefits are just a subset of the reasons the scientific method is so relied on by scientists. As marketing continues a transition to a technology and data science dependent discipline, marketers are better positioned to use a more agile, iterative approach to their campaign execution. Doing so will lead to more impactful marketing that better engages your customers and leads to more value for your company.