Why is it so vital to understand your customers and develop robust customer data? And ultimately, why did I feel compelled to write my book, Data-Driven Personalization, now?
After all, we’ve been talking about the power of digital marketing to help marketers connect the dots between marketing impact and marketing activity for over 10 years. Yet, even as customers’ expectations have shifted—research we conducted for the book found that nearly 9 in 10 customers expect brands to engage with them based on their shared histories—most marketers (48%) say their organizations are not yet collecting enough data. This is your opportunity to build a competitive advantage and break through with your audiences—by being the most relevant brand to them, based on their needs, wants, and motivations. To do that, you need to know more about them than the next brand.
But what happens when you don’t have enough owned data? That’s where this edited excerpt from my new book comes in:
In an ideal world your organization would make the necessary investments in high-quality customer research without hesitation. But I recognize that in the real world our organizations have limited budgets and time, and it can feel like a step backwards to say that we need to research our audience in order to move our marketing strategy forward. So how can we fly the metaphorical plane while we’re building it? Let’s take a look at three different approaches that you can use individually or in combination with each other to strengthen your personas and make strategic decisions on the types of data that will be most useful to collect and use in your personalization efforts.
1. Leveraging Third-Party Research
Although conducting your own first-party research allows you best control of the questions asked of your audiences and the way the research is analyzed, it’s also often the biggest investment of time and resources to produce. Therefore, many organizations may elect to use third-party research as a starting point. How do you select research that is aligned with your organizational needs?
In the classes I taught at the City College of New York and Columbia University, my graduate students would take a similar approach in their assignments: Developing personas and customer journey maps, without the ability to conduct real-world first-party research. Each semester, I would inevitably have a few students who would start with age cohorts and bring back data about the behaviors, needs, and interests of those age cohorts.
For instance, they might find reporting on the technology preferences of Gen Z or the financial considerations of Boomers. While there are, of course high-quality research pieces that delve deeply into the needs of these specific age cohorts, such as the analyses from AARP Research (a client of both Convince & Convert and Media Volery) or reports from Pew Research, this reporting doesn’t necessarily align with the best ways to segment your audiences. Unfortunately, this kind of data may be too broad or non-specific to be valuable for shaping your customer segments and personas. Your audience segments will often relate to purchase behaviors, changing life stages, and/or attitudes, rather than age or other demographics. So, start with data that’s highly pertinent to your audience’s needs and use case, rather than making assumptions based on demographics.
To that end, I recommend leveraging industry research: What research has been conducted within the specific industry in which your brand operates? Industry-specific technology brands often publish reports based on their owned data. For instance, if your organization is in real estate, you might leverage the home-buyer reports that Zillow has produced. Trade industries are another good resource for reporting. If you are in the events space, you might leverage the data produced in conjunction with the Professional Convention Management Association (PCMA, a Convince & Convert client).
Many vendors provide reports specific to industries or even geographies. For instance, you may consider reports from IBISWorld, which offers an in-depth overview of a variety of different industries. Or if you have physical locations, you may consider purchasing reports from Esri, which offer detailed demographics and psychographics for populations in specific geographies.
Finally, you should also leverage digital audience insight tools that allow you to understand audience cohorts based on the topics they discuss and even the websites they frequent. For instance, Sparktoro and Audiense Insights are two tools that identifies the online and social media behaviors of audiences, based on content they already engage with. I’ve used tools like these to help clients identify the skills and interests, geographies, content consumption patterns and demographics of their likely audience—based on people who follow certain accounts online, visit competitive websites, and/ or discuss certain topics online. (These tools can also be used for social listening purposes, focusing on your brand’s own audiences.)
While there are costs associated with sourcing high-quality, highly relevant third-party research to develop out your audience research, segmentation, and personas, it’s absolutely worth the time and effort. This research grounds your audience work in the specifics of what your audience values, what channels they frequent, and who they trust. It helps you to be more strategic in what data you collect moving forward and how you analyze the behavioral signals from your customers throughout their relationship with your brand.
2. Capitalizing on AI and Large Language Models
While research will provide you with a great deal of direction and allow you to add depth and specificity to your personas, you may find that it still leaves questions unanswered. Furthermore, you may prefer to be able to dialogue with your customers to learn more and test hypotheses further. In the next section, we’ll look at how we can build content that helps to fulfil this roll, but short of that, what else can marketers do to get more in-depth insights into what our key personas think, feel, and do?
With the increased accessibility of AI tools built on large language models (LLMs), we all have the ability to leverage the wisdom of the crowd. Because LLMs are built upon scraping reams and reams of human language and, by their very nature, predict what are the likely words related to each word provided, they are good mimics of people—pulling from the language that people use when describing specific situations. These tools can be good brainstorming partners when you give them relatively detailed prompts or inputs.
My friend Andy Crestodina, founder of Orbit Media and an expert in content marketing and technologies for marketing, walked through the process of using AI to support persona-development in a blog post on his company’s website, “How to create AI marketing personas with 8 powerful prompts” (2023). In it, he suggests feeding ChatGPT (or a similar chat-based AI tool) the basic parameters of your customer segment (such as job title, business category and mission, company size and geography, and business objectives) and then asking ChatGPT to provide the goals, pain points, and decision criteria of this persona. The benefits of this approach are that, because LLMs have read thousands if not millions of pages of content from people who fit your parameters, it can provide a relatively cogent description of the goals, pain points, and decision criteria that people like this have described online. Crestodina then takes this approach further by using the tool to ask questions about the words that this persona might use to describe their marketing challenges and solutions. He also uses Bing to suggest resources that this persona might use in their research process.
All of these exercises allow you to gather more detail that approximates the types of responses that customers like yours might share based on their unique situation. (While it’s not perfect, it can give you more depth and specificity to your thinking about your customers’ needs and motivations.) Once you have your persona revised, you can even input it into your AI and ask it to answer questions in the guise of this persona. With that capability, you have a feedback partner that can give you insights into how to make your marketing messages and content more relevant to your core personas. Ask your AI persona to rewrite descriptions or headlines to be more appealing to that persona. Ask it to give you feedback on key questions your customers may need answered. But remember: Just as with the third-party research, this is an exercise that gives you a starting point, not the official personas you should use forever and ever. Instead, treat the personas you refine with AI as hypotheses to test with real customer feedback.
3. Testing With Live Customers
That brings us to our final category of testing: Engaging with customers in the real world and learning on the fly. Because working with live customers is potentially costly in terms of resources and potentially risky in terms of reputation, testing with live customers should only be done once you have a strong starting point. That being said, testing should be an ongoing part of your customer segmentation and persona-building process. As I’ve said before, organizations must continue to revise these frameworks over time, as circumstances and audiences change.
What does testing with live customers look like? If you’ve ever stumbled upon a new restaurant in your neighborhood, you may have wandered in to try it out and you may have been told that they’re in the “soft opening” phase. This is a period when the restaurant is open, but not yet promoting their opening, so that they can test their menu and their staff and work out the kinks. Similarly, there are many software companies that launch new product offerings in “beta” to gather feedback prior to the official launch.
Beta periods can also be a useful time to gather customer feedback on messaging and offers. Many brands test out various messages, offers, and pricing models using experimentation or multivariate testing tools to measure which versions resonate best with customers.
But beyond just putting messages passively in front of customers and waiting to see which ones resonate, your organization can also take a more proactive approach. Consider paying for survey panels to give feedback on specific offers or messaging. With the accessibility of platforms like SurveyMonkey and Cint, recruiting panels that fit your audience criteria is more affordable than ever. If you want to further engage with customers who are specifically interacting with your brand, you may consider offering discounts or gift cards to prospective customers in exchange for participating in surveys or focus groups. Using exit intent pop-ups (those windows that pop-up on a website as your mouse moves out of the window) to showcase single-question surveys can also be a useful way to get feedback on specific questions that may be on the minds of your team. Lastly, ask for feedback after your interactions with customers. While many of us are aware of NPS surveys as a customer experience feedback metric, incorporating open-ended questions that encourage customers to share their feelings can give you more insights that help. (I would recommend only putting one question in front of customers at a time to increase participation and show respect for your customers’ time.)
In addition to testing through surveys or feedback mechanisms, how customers engage with content is also a valuable mechanism for learning. In the last two decades, content marketing (the providing of valuable resources for free to customers to build brand affinity and trust) has become the backbone of many digital marketing programs. With good tracking and measurement, your organization should be able to see what content pulls in traffic to your website, as well as what traffic results in conversions to actual sales.
Creating content—on social media, on your website and landing pages, in emails, and in other channels—based on your assumptions about your core audience’s needs and motivations is a useful way to test what truly resonates. Using your personas and an experimentation or multivariate tool, you can test various aspects of your content, analyze what resonates, and refine your assumptions based on what performs best. Things to test include:
- headlines
- sub-headlines
- calls-to-action (CTAs)
- meta descriptions and/or short summary statements
- value propositions or framing
- types of appeals
The goal is to understand what your customers care about and respond to by seeing what they engage with, how they engage with that content and how it impacts their relationship with your brand. For instance, perhaps your brand is a commercial bank looking to understand the priorities of your professional Millennial core customer. You may test content about building a budget to save for a new home, building a budget to save for a dream vacation, and building a budget to make a career change stress-free. By seeing which of these value propositions resonates with your audience most strongly, you may be able to develop additional content for social media, for ad campaigns, and for further nurturing those customers.
Want to learn more about these concepts? My book, Data-Driven Personalization, explores why and how marketing leaders need to implement a data-driven strategy that puts the customer at the center of how they build marketing. Not yet convinced? I shared some of the key frameworks of the book in Orbit Media’s Wine & Webinar series—click here to watch that session.
This edited extract is from Data-Driven Personalization by Zontee Hou ©2024 and is reproduced and adapted with permission from Kogan Page Ltd.