Q: To start us off, where should brands begin when building an audience targeting strategy?
Sean: It’s certainly a balancing act. You want to scale your marketing and reach as many potential customers as possible, but you don’t want to waste marketing dollars either. A great place to start is with the audience we already know – the most deterministic, valuable customers we have line of sight with. Engage them first, then model off of them.
Then, expand your research with a focus on the human element. There is limited first-party or deterministic audience data. So we have to get to know our audience beyond those data points. What are their interest behaviors, attributes, and even psychographics? Start building upon your original data set with these insights. This can include social listening, focus group data, or other things of that nature.
Q: In the midst of the cookieless future, what are some ways to build a futureproof audience strategy?
Savannah: We’re in a really interesting transitional time. I’ve been referring to the period we kind of grew up in here at Coegi as the ‘programmatic Wild West’. We had so much data at our fingertips that we could skim through pre-built audiences and find a third-party data set we were really confident in. As we shift toward consumer privacy being more of a focus, we need to return to marketing basics. Social listening, as Sean said, is a huge one – especially with social media looking vastly different today than it did 10 years ago.
Also, simply put yourself in your audience’s shoes. If something comes up in your research – blogs they read, shows they watch, subreddits they subscribe to – spend some time in those spaces. I think it will spark some interesting ideas of different touch points you can add to your overall strategy.
Sean: I would also add that we are still maintaining a data-driven approach. Prior to the programmatic era, media decisions were often based on assumptions. Data-driven advertising helps us use quantitative data to inform who our audience really is. Now, we may be looking at a variety of other qualitative sources, but we want our assumptions to be backed by data.
I think a good example was some campaigns we did with BODYARMOR for a number of years. Obviously, athletes are their target audience in the sports drink category. But, research found that moms were actually a primary purchaser in bulk at large retail stores. So that became an entirely new audience with a different messaging strategy.
Q: How can brands best capitalize on first-party data to identify and reach potential new customers?
Savannah: First-party data collection was one of our first recommendations when Google made their announcement to deprecate third-party cookies. But there have been roadblocks along the way. Many brands are realizing that the way they set up their point of sale systems or their website was not ideal for aggregating all of their data. Especially if you have loyal consumers who use your products and are willing to give you their personal information, you want to gather all of that first-party data in one central location.
So, whether it’s a CRM system or an ACP system, make sure your data is in an area where you can evaluate it. Then, let that high quality audience determine how you experiment as you broaden your strategies.
Q: How can you apply audience data learnings across channels to bring the most value possible?
Sean: Because we experience so many media touchpoints day-to-day, we want to take a broad, holistic view when we have valuable first-party data to gain audience insights. It could be the websites they visit, the influencers they follow, their location patterns, and even heat maps to the retail chains they frequently visit. By holistically researching how these customers spend time and where they devote their attention, we can get a full view of how to engage them throughout the day.
Savannah: And that also helps us understand how our audience is responding to our messaging throughout the campaign lifetime. For example, Coegi media planners are beginning to implement a performance scoring model as a part of our measurement strategies.
Let’s use the simple example of someone in-market for a car purchase. If they’re visiting our brand’s website and looking at different models, they might still be in the discovery phase. If we know they visited the lot too and spoke to salespeople, that’s a much more invested person who’s more likely to take the next step. So it helps us retroactively look at each touchpoint and the actions that grow out of them to understand the true effects of marketing.
Q: How do you measure the effectiveness of an omnichannel audience strategy and build a test and learn approach to refine the process?
Savannah: First, we empower the full team to come together: our in-platform specialists, strategists, research team, and even clients. Have a proactive conversation about what each step of the consumer journey really means and how each step needs to be measured against our media.
Having this conversation upfront with all the correct people not only informs your setup strategy, it will also aid your optimization strategy. It can help you put together reports with really valuable insights. And overall, it leads to more successful start-to-finish campaigns that are replicable.
Sean: This approach also powers our measurement strategy and learning agenda. As we are laying out the strategy, we make certain hypotheses. Then, throughout the campaign, we’re proving those suppositions either correct or incorrect and making pivots. The test and learn approach allows us to iterate on an ongoing basis to drive performance.
Savannah: And there’s an added value of being honest and transparent – having those real conversations with teams and clients upfront. Often, our instinct is to want to always be the expert in every piece of our campaign. This gives us an opportunity to say, this is our expectation, these are our benchmarks, but let’s plan for what to do if this doesn’t work.
Q: How do you balance human intuition with AI modeling to identify your next best customers and refine your marketing strategy?
Sean: We have to understand our audience and be respectful to the sensitivities of their data. It really comes down to putting guardrails around AI machine learning – simple things such as frequency caps and sequential rotations of your creative messages to tell a story.
Is cost-per-click or click-through-rate really driving growth for your brand? Or are you simply capitalizing on consumers that were going to purchase anyway? It’s a combination of understanding the human element, putting guardrails in place for machine learning to respect our customers, and then implementing a rigorous measurement strategy.
Q: How do you avoid alienating customers with ad oversaturation and build a roadmap to long-lasting customer relationships that grow over time?
Savannah: Well, I love what Sean said about making sure your audience is seen as a human. One of the easiest ways to do that is to think, “what annoys me?” For example, when I get the same connected TV ad 400 times. What turns my view of a brand off and what can we avoid in our strategy? As you’re putting together tactics, think of the things that personally rub you the wrong way and be sure to avoid them.
Sean: I think it’s also important to regularly refresh our customer database so we don’t forget about lapsed customers. We’re going to approach them differently than our most loyal customers. Understanding the nature of our first-party audience is another way to communicate with them effectively.
Live Listener Q&A
Q: How do you build a customized user journey without feeling invasive or creepy?
Savannah: Creepiness is obviously subjective, but for me, where I have felt that line was crossed is when I am getting a super personalized message from a brand I’m unfamiliar with. This speaks back to maintaining and nurturing your CRM list.
I may have bought a product from this company years ago and they slipped my mind. So when I get that really hyper-targeted search banner ad or those t-shirts on Facebook with my name on them for some reason – those things are typically when the red flag goes up. They feel more invasive than a personalized email from a company who I’ve purchased from several times.
Q: For a brand in the startup phase, how do you begin to build an audience strategy?
Sean: I think a good place to start is simply your website analytics. If you’re a startup, you’re likely going to do some sort of press release. You’re going to try to get your name out there, and you may be doing some things to engage customers face-to-face. Take each of these opportunities to gather as much data as possible.
From an online standpoint, there’s always your website analytics. You can drill down to the city level or even the DMA level to find where qualified traffic is coming from. If you have multiple pages, which are visitors most engaged with? What time of day are they coming to your site?
There’s a number of signals there that can be a starting point for audience learning. If you are able to engage face-to-face with a few people, you’ll gain insights about what the consumer response may be at a larger scale. So record and leverage that critical feedback.
Savannah: There’s also an opportunity in the early days to think about creative ways to incentivize your initial customers. A common tactic is offering a discount if people sign up for your newsletter.
Q: What are some creative ways for brands to jumpstart their first-party data collection when starting from scratch?
Savannah: This is where partnerships can really come into play. Second-party data is a great place to start. If you don’t have a robust CRM list of your own, look for other businesses with high-quality data and do your due diligence to evaluate it.
You can also look at things like retail media partnerships. If you’ve done on-the-ground research of where your consumers shop and what they’re interested in – you could go to Target’s Roundel, for example. Maybe you know your audience is in-market for parenting items. Look at those retailer audiences and see if there’s a unique way to reach them there.
Sean: You can also tap into your creative executions in some cases. For instance, on Meta, someone who watches a video all the way through can be put into a remarketing bucket. Then you can perform lookalike modeling off of that group. You can do the same thing with programmatic video and there are other types of creative formats that allow you to gather first-party data.
Q: What is your number one tip for audience segmentation and relationship building?
Savannah: Simply don’t forget that your audience is made up of people. Each member of your audience has a unique relationship and journey with your brand. Any opportunity you have to segment your audience and deliver different messaging at different stages of their journey is a great way to strategically build meaningful relationships.
From there, it can inform the channels you execute on. It can inform your creative messaging. Overall, it lays a really solid foundation from people who are new to your brand to the loyal customers you’re working to build.
Read Savannah’s Cookieless Targeting 101 article here. For more tips on consumer research, listen to our podcast episode, Research Done Right.