Broadening the context of contextual marketing
Contextual targeting is back in vogue like so many ‘old school’ marketing capabilities we relied on before data-rich audience targeting became the norm. The consumer privacy movement can take credit for that. Not knowing who is behind the screen you are advertising to may soon be the new norm for all but walled gardens who have bartered discounts, entertainment, or convenience for the ability to target you individually. Advertisers must readapt to less precise targeting for most other forms of media. The most frequently mentioned alternative for privacy-conscious marketing is contextual. It seems however that most industry conversations are focused on using contextual targeting purely to mitigate lost addressable targeting scale. Yes, there are separate discussions on using context for brand safety, fraud detection, moment targeting, and ad personalization too, but there are few industry conversations about making context a strategic pillar across your omni-marketing strategies and platforms. This is where I would like to take this blog post.
I have been fortunate to have helped brand marketers get attention and motivate sales across most marketing tactics and media channels from in-store merchandising, newspaper circulars, and billboards to e-commerce, social media, and television. In many ways, these marketing channels could not be more different. Yet, it never seemed that way to me because what mattered most was how your marketing message looked, sounded, and felt to the end customer. Regardless of the marketing vehicle, the old marketing cliché was always your goal - ‘right message, right person, right time’. And to me, that is contextual marketing.
My goal with this blog is to encourage you to broaden your leverage of context across your marketing strategies and platforms. This matters for everyone involved in your marketing strategy and tactics but is particularly relevant for those responsible for architecting your data, technologies, and processes to adapt to the generational changes our industry is going through.
These are the questions they should be asking…
What is context anyway?
To understand how to broaden and maximize contextual marketing, we need to understand what it is. It is broader than most think. It is the time, location, place, surroundings, conditions, and mindsets that influence what we buy, not buy, when, where, how, and on what terms. It includes what is happening right now, but also what just happened, what is about to happen, and what people predict will happen in the future.
Examples of context include:
Location – address, zip, DMA, state, country, urban, rural, suburban, at work, at home, at beach, in mountains, inside, outside, etc.
Place – specific building, store, grocery store, retailer banner, restaurant, entertainment venue, in parking lot, at home, at work, etc.
Time – day of week, time of day, weekday, weekend, season, holiday, morning commute, breakfast, primetime, school in session, spring break, etc.
Weather – Hot, cold, colder than normal, raining, snowing, storm, forecasted storm, high pollen, high smog, etc.
National Events – Olympics, presidential election, NFL playoffs, Final Four, Academy Awards, etc.
Local Events – Farmers market, county fair, music festival, road construction, etc.
Content – Sports, football, NFL, healthy eating, cooking, recipes, music, rap, politics, election, etc.
Device – Mobile phone, TV, laptop, Apple, Android, kiosk, digital billboard, static billboard, car radio, etc.
Life Stage – Growing up, adulthood, parenthood, schooling, working, empty nesting, retiring, etc.
Life Events – Births, marriages, anniversaries, divorces, deaths, new home, sold home, new job, lost job, etc.
Health – healthy, sick, diabetic, high cholesterol, lactose intolerant, gluten sensitive, etc.
Economics – High inflation, low inflation, gas prices up, mortgage rates down, stock market up, etc.
So much context! How do we ever make decisions? Thankfully for consumers most of this is subconscious. Unfortunately for marketers, it is not. Making matters worse, these contextual signals are not binary, and their meaning and impact can change when combined. For example, someone reviewing an online recipe at 6:30pm on a Tuesday likely already has the ingredients they need, but someone on that exact same site at 9:30am on Saturday is likely creating a shopping list. Likewise, the community impact of a local festival is likely reduced due to forecasted heavy thunderstorms.
What context matters to my brands?
Most contextual targeting solutions are BYOC (Bring Your Own Context). They enable you to target and buy media via contextual signals, but it is totally up to you to determine the context that matters for your brand. This is not as simple as beer drinkers love football. Those days are over. The consumer privacy paradigm, wall garden data fragmentation, and AI capabilities are rewriting the rules and expectations for contextual marketing.
Despite the AI hoopla, I expect your consumer and category-savvy experts will still drive most contextual strategies given how deeply embedded context is with human emotions and mindsets. I am sure the data would find plenty of inappropriate times to say ‘Had a bad day? Drink Feel Better Whiskey.’ That said, humans cannot keep up with the scale, diversity, and speed demands that are coming. Carefully crafted data and machine learning will be required to proactively discover the best contextual scenarios for your brand objectives and audiences.
This is not a simple correlation exercise either. Some context is awesome at driving immediate sales but is not great at driving incremental awareness. Other contexts may have nothing to do with your brand’s category but it is popular with your brand audience. Some contextual targeting should be lagged (baby registry, house hunting). Some require you to accumulate patterns (consistent early riser, football fanatic). And some context has yet to come (snowstorm forecasted, mortgage rate predictions). Brands need to understand these subtleties to maximize the value of context.
How will you leverage context?
Hopefully, your answer is everywhere, but be honest. Many of you are being asked to prepare your contextual capabilities specifically for Google’s privacy sandbox or upfront publisher buys to ensure quality advertising scale. The targeting mindset for many is audience ids OR context signals. The reality is that it can be an AND too. In fact, optimizing media buys and personalizing creatives is best when you know both the audience and the context.
Regardless of your immediate ambitions, I suggest you prepare your platforms to evidently leverage context everywhere which means being ready for each of the following contextual targeting scenarios.
Context only targeting – Relying exclusively on context may be required where addressable audience data is not available or affordable. This may increasingly become the norm for the open web and other non-walled garden media platforms. Contextual targeting may not be super precise, but can reduce non-audience waste relative to mass marketing or other poor quality ‘targeting’ methods.
Context with audience ids – Leveraging context even when you have an id is among the most powerful targeting approaches. It should be considered for high-value moments and high-performing personalization opportunities.
Context with segment – This is a relatively new concept where you trust a publisher, platform, or browser that this person is in a cohort of interest. This can be an IAB seller defined audience or browser cohort based on recent browsing history (e.g. Google privacy sandbox).
Do I have the data to do this?
Now is the hard and ironic part. You still need id-level behavioral data. Yeah, you heard that right. We need id-level data for id-less targeting. At least some. Brands need data to correlate desired brand outcomes with diverse media content and other contextual signals. Analytically, this is not that difficult. The challenge is securing and assembling the necessary data required and mitigating the data quality and bias that exists in most available data.
So what data are we talking about? For most advertisers and agencies you should strive for the following types of data to fuel your contextual targeting.
Bare minimum data
Expert human data – Get what you know into your data so your AI and technology automation can use it.
Survey data – This has been the historically easy button for contextual intelligence. Just ask consumers who they are, what they buy, and what media they watch. For many, this will be your baseline dataset until you enable data with more granularity, accuracy, scale, recency, and fidelity. Survey data will always fill in where other data does not exist.
Contextual classification data – Most data does not come context ready. You need to transform it so humans and AI can see patterns as consumers do. You need to classify URLs, TV programs, etc. to align with media platform categories (TV genres, IAB categories, Google TOPICs, etc.). Precise locations need to be classified into postal designations, DMAs, etc. Time assigned to dayparts, work shifts, seasons, holidays, etc.
Ideal data
Desired outcome data - Typically this is shopper transaction data that includes consumer id, time, location, and product purchased. You want this data to be as nationally representative and demographically balanced as possible using purchase panel and/or retailer transaction data.
Media consumption data – Ideally you secure broad-based media viewing data that covers most media content categories to eliminate bias and avoid blind spots. This data includes consumer id, time, content site, and publisher classifications
Ad exposure data – To understand how specific brand objective campaigns perform across context, you need to capture as much ad exposure data you can. This data includes consumer id, time, content site, campaign objective, cost, etc.
Id Graph data – You will need to link your sales, media consumption, and ad exposure together via a common consumer or household id.
Are my marketing platforms prepared?
Most brand experiences with contextual marketing have been one-off campaigns, media platform-specific, and highly manual. Few have mastered individually personalized, omni-media, or ‘always on’ contextual marketing capabilities. Beyond the internal media organizational fragmentation that hampers synergy goals, the entire mar-tech industry seems to have different contextual definitions and technology approaches. There may be some industry standardization over time, but brands are likely on their own to effectively and efficiently transform their centralized brand context strategies to dozens of varied marketing platforms.
The bulk of the complexity comes down to the bid. Many media buying platforms only provide the basics like domain, URL, device, language, etc. Most publishers, SSPs, and DSPs augment with context like IAB content categorization and geography definitions too. You are usually on your own if you want to add more sophisticated and relevant contextual signals to make campaign setup and bid optimization easier. Thankfully DSPs are adapting to these radical industry changes too and are enabling more customization capabilities. Walled gardens are adding contextual signals to their fancy new targeting capabilities too, but you will have to trust them on that .
Do not be discouraged if your marketing platforms are not ready. Remember that your platform partners are also adapting to these increased contextual targeting needs. Also remember that privacy, clean rooms, retail media, fragmentation, signal loss, etc. will require you to revamp what you have anyway. This may be the ideal time to think strategically and bake context into that new enterprise design.
Where do I start?
What! You have not started? Google told you that cookies were going away in 2020!
Turns out you have time. I suggest that you use some of that extra time to briefly step back and more deeply assess how context factors into all your marketing. Imagine you had a central knowledge source for all brand context intelligence, definitions, and taxonomies that consistently informed and aligned all your marketing. Avoid any temptation to just bolt capabilities onto legacy contextual capabilities you may already have. There may be great stuff there, think bigger. This is the time.
Ultimately, your journey to make contextual marketing a strategic pillar in our mar-tech platform will take a major concerted effort, but worth it in this new privacy-conscious ecosystem. Below is a summary of some of the suggested actions mentioned in this blog.
Align corporately on your contextual marketing strategy, goals, and investment
Design and centralize your brand context data, definitions, taxonomies, and analytics
Identify and source necessary context, purchase, and media consumption data
Classify brand objectives and contextual moments in your historical behavioral data
Determine contextual correlations for each brand objective and audience
Map your proprietary contextual definitions and taxonomy to each media platform’s unique contextual capabilities
Assess and quantify the cost-benefit of leveraging context for media buys and creative personalization
Experiment, learn, measure, enhance, adjust, etc.
Hope this helps. Let me know what you think. Reach out if you would like some help on your unique journey. Good luck!