How to Become a Marketing Scientist
As I remember it, Mrs. Cain taught us about making observations, hypotheses, dependent and independent variables, putting together a lab plan, observing what happened, taking copious notes and then drawing conclusions relative to what our hypotheses were. Were we correct? Incorrect? If incorrect, why? How could we reconstruct the experiment in a different way to test our hypothesis further? If correct, what were the proof points? What data did we have to support our conclusions?
This is the same structure we should embrace as marketing leaders, but too often, we get caught up with the shiny pennies and funny names—growth hacking, content marketing, inbound marketing, social marketing, influencer marketing—that we don’t stop to think about actually developing a strategy and testing it. These channels and tactics are NOT a marketing strategy; they’re ways in which you execute and test a hypothesis based on careful understanding of your target market, their pain points, and the product or services you’re trying to sell.
The Marketing Scientific Method
"You never know someone until you've walked around in their skin."
Ideally, before you build ANYTHING, you have identified a group of people who share a problem. Perhaps this arises from a personal issue you experienced and for which you haven’t found a solution. Once you start talking to others like you—maybe other moms, students, engineers, marketers, whatever the segment—you realize that you all share this problem and collectively haven’t found a solution. Aha! Invention, or perhaps just innovation on an existing tool or service that isn’t cutting it, begins!
“Get Curious. Talk to People.”
Continue your observations at scale. How many other people have the same problem? They may not look, talk or live like you do (this is the issue with relying too much on demographics to segment a market), but they must have the same problem and no viable solution.
When you find people with the same problem, talk to them and ask questions to draw out why they have this problem, what they’ve tried to do to solve it and why it’s not working. Take notes. Tons of notes. Code them in a way that allows you to identify patterns that will help inform a more quantitative way of surveying a larger group.
You don’t have to hire a fancy research firm to do this; indeed, it isn’t in the budget for most. This is where that network you’ve been building for your entire life comes in and where social media can be particularly helpful. Crowdsource with friends and family. Walk the streets in your neighborhood. Eschew that advice to never talk to strangers (I’m particularly a fan of talking to Lyft drivers).
When you think you have enough qualitative information to begin crafting and testing your hypothesis at statistically significant scale, start codifying and prioritizing your questions. Then sign up for free or free-ish tools like Google Surveys or Survey Monkey.
Be respectful of your audience’s time, however, and think about small rewards that you can provide. You’d be surprised how responsive people are when you offer them a free cup of coffee. You don’t necessarily need to give them $100 in exchange for 30 minutes of their time.
2) Develop Your Hypothesis
Develop a hypothesis rooted in audience observation that includes the following information.
Audience definition (aka “persona” development; for B2B marketers, you may also include definitions of target organizations/companies)—This needs to be human and comprehensive. Stating “male 25 to 35 years old who lives in San Francisco” is insufficient. You need to know everything about this guy: His name is George; he grew up in Portland going to the Oregon coast with his family every summer. He’s 32 and is a civil engineer working for the San Francisco city government building and repairing bridges. In his spare time, he enjoys extreme sports and hiking with his girlfriend of two years. He just rescued a greyhound. He volunteers at a soup kitchen once a month and his favorite place to visit is Paris at Christmastime. That level of detail. You want this person to come to life because then you can find this person and others who look/act/think like him and are likely to have the same problems for which you have a solution.
Product and service definition—What are you attempting to sell to this audience, and how will you support these customers?
A compelling reason to buy—Why are you uniquely qualified to solve the customer’s problem? If you listened carefully and crafted the right questions during the observation phase, this should be a summary of what you heard and why you believe you’re special. Yes, we’re all special snowflakes, but you need to make a compelling argument. This is the crux of the hypothesis you’re going to test. This positioning should NOT change over the course of this experiment. While your audience messaging may change, the fundamental value you provide should not change until you have data that leads you to conclude your hypothesis was incorrect.
Quantitative and qualitative goals to measure success—Here’s why many marketers fall down: They don’t begin their process with quantitative and qualitative benchmarks in mind. If you’re starting from scratch, the most logical place to focus is on the quantity and quality of leads. For example, if you have one salesperson on staff, maybe they can process 50 leads a week. So set your weekly target at driving 50 leads. At the end of the week (or maybe even daily), sit with this salesperson to assess the quality of each lead. Did you hit the right target based on the characteristics laid out? Were they qualified? Were they in the right spot in their buyer’s journey? What questions or objections did they hear on the calls?
3) Experiment and Collect Data
This is your marketing execution plan. Ideally, it will contain the following five components.
Messaging map—A set of statements tailored to the needs/wants/desires/perspectives of each persona you identified above. In this testing phase, you’ll want to develop a few sets of variants for each message, maybe changing a few words or how you contextualize things across different channels.
Prioritized channels and spend—Based on your understanding of the audience, where do they hang out and how will they be most receptive to your messaging? For example, if you’re offering a consumer product or service, does it make sense to reach them on Facebook? If you’re offering a business product, should you reach them in trade journals or at events?
The good thing about channels like social and paid search is that you can test your messaging across audience segments and messaging variants for a relatively low spend (hundreds of dollars or even less, depending on your target audience). Armed with a few weeks of data from these tests, you can look to scale or increase spend on the channels and the messaging that’s proven to work.
Partnership with sales—The sales team should be your best friends. They’re talking to your audience daily so they hold golden nuggets of information. Schedule a weekly session between your entire marketing team and key sales teammates to review analytics and qualitative insights from sales calls. These need to be sessions where everyone brings their recommendations based on what they’ve observed and where you take action for the following week. In effect, this is sprint planning.
Monitor the market—We marketers can’t operate in a vacuum. It’s important to keep tabs on competitors: what they’re saying, doing and maybe not saying to the market. Are they changing pricing? Have they announced new partnerships? Are they acquiring other customers? Who are they hiring? This should be someone’s full-time focus (or at least someone’s half-time job), depending on how competitive and established your market is. If you’re creating a new category or market, good for you for getting your first-mover advantage on, but beware of new entrants. Don’t get complacent.
Pricing and packaging—These topics can be separate experiments unto themselves but should always be included in your overall plan. Pricing and packaging can severely affect—in good and bad ways—your success. While all of your other variables—messaging, channels, sales execution, etc.—may be spot on, the reason why you don’t see traction may be due to a disconnect between pricing, packaging and the target customer’s perceived value of your offering. Take heed and pay attention to this piece. If you need help, ask your finance team to put together scenarios based on the margins that need to be preserved, the audience’s willingness to pay and the competition.
4) Draw Conclusions and Take Action
People often ask about the ideal length of a marketing campaign or plan, and therein lies the problem. Many marketers see them as finite activities with binary outcomes, much like political campaigns: There’s a prescribed length of time and at the end you have a winner and loser. Becoming a marketing scientist forces you to think differently. Your work is never complete, and there’s always room for improvement. Embrace the power of experimentation, data and action to build better businesses and drive growth in the global economy.
A version of this article was originally published on Medium - http://bit.ly/2marketingscience
Looking for the latest in product and data science? Get our articles, webinars and podcasts.