5 Tactics For Improving Media Buying with Audience Insight
If you’ve been to any of my presentations or read many newsletters from the past year, you have probably seen/heard mention of SparkToro. It’s no secret I love the platform – but one of the questions I continually get is: how do you actually use it? How do you make it work? Why do you think it provides such a competitive advantage?
Before I go further, I want to clarify a few things: (1) I have absolutely zero financial interest in SparkToro. I don’t even have an affiliate code. (2) While Rand and I talk relatively frequently (1x per quarter or so), he doesn’t know this article is coming, and he’s had no influence on it. (3) The only products I write about are the ones that I actually use.
With that out of the way, let’s get into it.
Why do I care so much about Audience Research/Insight:
Anyone who is actively working in digital media will tell you that our collective, haphazard movement toward a more “privacy-friendly” web – from iOS14.5, iOS15, iOS16 & iOS17 to the rise of ad blockers, private browsers, restrictions on cookies and more – has resulted in three things:
- Pixel/Tag-based Audiences are less reliable – There’s a reason why virtually every ad platform is moving to a Conversions API and/or offline conversion upload: pixel-based audiences simply are not as robust or reliable as they once were. We’re seeing upwards of 30% leakage from tag-based remarketing ad sets into prospecting across platforms – and this is far from uncommon across both eCommerce + B2B marketing groups. Put simply, the technologies marketers + ad platforms/networks used to rely on to create audiences are not nearly as effective as they were 3, 5 or even 10 years ago.
- Third-Party Data Erosion – There’s an entire cottage industry that has evolved around consumer / user data – from the credit bureaus to data brokers to ad networks (withmany more in-between). These groups function much like investment banks assembling mortgage-backed securities: they acquire multiple different data sets, layer them together using various primary keys (usually a name/zip code, email or phone number), then sell the “enriched” data set to ad platforms or DSPs, who proceed to further enrich them with their own data. The end result? Many of the “programmatic” targeting options that are available to advertisers. While this may seem well and good, the reality is that 3P data is notoriously unreliable in the most nefarious way: it seems accurate!
How is this possible? Three things to consider: (i) all data decays at different rates – our interests, locations, job, salary, behaviors all change over time, but not at the same time or on the same timeline; (ii) privacy restrictions have limited the dataavailable to collect, which has resulted in some datasets being updated using inference (read: very-expensively-modeled guesses) instead of observation and (iii) layering inaccurate/wrong data with accurate data creates inaccurate data.
- Major Ad Platforms Have Opted For Size Over Granularity – Back in the day (I feel old just writing that), major ad platforms like Meta and Google provided thousands of hyper-niche targeting options for advertisers. As crazy as it sounds, it was possible to target Facebook Ads based on a users’ specific income/net worth tier (not HHI decile, but specific values: $100,000-$124,999, $125,000-$149,999, $150,000-$249,999, etc.), the value of their primary residence, the number of lines of credit they had open (1-9, 10+), when/how much they paid for their car, the specificairlines/class they flew, the types of causes they donated to, the types of news they consumed, and so many others. But, in the last 6 years, virtually all of those have been removed across Meta, Google, TikTok and others. A major driver of this was the push to privacy, but that was far from the only one: platforms wanted bigger audiences, because larger audiences translate into more bidders for more placements, which means higher CPCs/CPMs – and ultimately, more revenue for digital ad platforms. From an advertiser perspective, this means less control over where, when and to whom your ads show, ultimately resulting in lower efficiency, less relevance (which is why “targeted by creative” has come into vogue) and less insight into the underlying distribution of results.
If we were to sum up all of this, it’s simple: we’ve traded insight for automation and ease. Most advertisers are (at best) flying blind, or (at worst) flying with mis-calibrated instrumentation. Platforms are more opaque about where our ad dollars are going, less granular in how we can control those placements, and more unreliable in terms of audience-to-reality fidelity. The (predictable) response from many marketers was to embrace “broad” targeting (suboptimal), then use creative and/or 0P/1P data to hone in on the “right” audiences.
Marketing, Meet Georg Wilhelm Frederich Hegel
I’ve always had a weird affinity for Hegelian philosophy, and it’s served me quite well in situations like this. If micro-targeting was the thesis, and our current broad-bender is the antithesis, then what follows – and where the incrementality/opportunity lies – is the synthesis: a new proposition that incorporates elements from both the thesis and the antithesis.
My hypothesis is that the synthesis will be cross-platform, audience-driven targeting. The enabler for this is audience understanding – and that’s where SparkToro (and platforms like it) come in. I’m under no illusion that this “new” paradigm will last (nothing lasts forever), or that it won’t generate its own antithesis (it will), but for now, this is where the opportunity lies.
Cross-Platform, Audience Driven Targeting
Given the current technical and regulatory realities facing the digital space, we marketers simply don’t have the capacity to micro-target audiences the way we once did, and the data platforms we previously relied upon – for many reasons – are no longer as accurate or capable.
But ultimately, it’s still possible to do what we want to do (reach a specific audience in the places that audience is most likely to both frequent and trust) – we just need to change how we think about the problem + leverage alternative data streams (like click-stream data, public social data, and SERP data).
Let’s imagine we want to reach an audience of senior marketing leaders/executives (CMOs, VPs/EVPs of Marketing, etc.). Ordinarily, we’d probably head over to LinkedIn Ads, type in some job titles (and maybe some industries/skills), load up some ads, cross our fingers + press “go.” Maybe we’d also do some programmatic ads on well-known marketing publications like Ad Age or Ad Week, and maybe – if we were really determined – we’d try some Meta Ads (targeting CMO + related job titles) or YouTube ads. Each of these is an example of a commonplace marketing tactic, which limits the impact. Just about anyone can execute these campaigns with relative ease (leaving aside the sophistication and quality at which they are done, which do play material roles in the efficacy of those tactics).
But true opportunity isn’t found by doing the expected; you don’t catch more fish by sitting on the same pier as 100 other fishermen (which is what you’re doing with LinkedIn); you catch more fish by setting up shop at a pond or stream or pier that (i) contains a ton of fish and (ii) very few fisherman frequent. So, let’s talk about how to do that with Sparktoro (and some examples):
Tactic #1: Find Hidden Gem Websites
Given that our (collective) ability to micro-target individual users has been massively curtailed, what’s the next best thing? Go advertise on the places your target audience is disproportionately more likely to visit AND that other advertisers tend to glance over or not think about.
So, while it’s certainly true that about 90% of CMOs will visit the Washington Post on a monthly basis (as an example), CMOs are a comparatively small part of the WaPO audience (~60M people/month) – making the cost/CMO reached staggeringly high.
Where the advantage comes is in finding the niche sites where CMOs dramatically over-index:
This is a treasure trove of high-quality placement targets – as well as plenty of good, old-fashioned marketing ideas. Marketing Brew, for instance, has an entire suite of partnership/advertising opportunities that are surprisingly affordable.
CampaignLive.com, ChiefMarketer.com, ContentMarketingInstitute.com, MarketingDive.com and SearchEngineJournal all have both advertising options plus thought leadership/guest posting opportunities (yes, that does involve talking to your PR/content teams) – and are all cheaper than the Washington Post. If you’re buying ads programmatically, about 50% of the sites on this list are available via Basis and/or TradeDesk.
If you wanted to go even further, publications like Digiday, and AdWeek are all “interest” targets available in Meta Ads manager – great for part of an interest stack audience.
Tactic #2: Make + Place Better YouTube Ads
A few weeks ago, I wrote about YouTube Ads being some of the most valuable impressions on the Web – and I stand by that claim. Many of the comments I received from that issue focused on two themes:
- How do I find the right YouTube channels to target?
- Where can I learn about how my target audience actually consumes information?
This is another place where I use SparkToro. For the same (“CMO” in bio) audience, here’s an initial list of channels that this audience is likely to subscribe to:
My typical process for this is to run 5-10 of these searches, targeting different variations or combinations (“Advanced Searches” in SparkToro parlance), export each one, then combine those (ChatGPT/Gemini is helpful here) to find the commonalities. Those YouTube channels – the ones that (for example) people with VP Marketing, CMO, Chief Marketing Officer, Head of Growth and Director of Marketing all frequent – are the ones that I’ll (1) include as initial placements for YouTube Ads and (2) use for creative research / honing in on my audience’s natural content language.
But, as anyone who has run YouTube ads will tell you, nailing the placements is only half the battle. The other half is creating the type/style of content your target audience prefers – and for that, we need a combination of (i) scouting the channels identified above and (ii) understanding how the people this audience tend to follow (i.e. influencers) structure their content. Here too, audience research tools like SparkToro come in handy. This is using the same search above:
Most people in marketing likely recognize some of these individuals: Jay Baer, Brian Solis, Joe Puluzzi, Rand Fishkin, Ann Handley, Avinash Kaushik all have well-known personal brands built over decades. But interspersed in this list are some lesser-known people: Ann Tran, Matt Heinz, Christopher Penn, Heidi Cohen and many others. And (in my view) the smaller, less-well-known people are often under-estimated fonts of inspiration.
The simple reality is that anyone who is relatively new to the marketing world and is quite influential likely has achieved their success via audience understanding. Put simply: in my experience, the smaller “micro-influencers” are more likely to have a proverbial “finger on the pulse” of your audience.
In this case, I’ve identified about 50 “influencers” for my “CMO” audience – and just as with the channels above, I’ll repeat this search for various other job titles/bio descriptions, then find the commonalities across those 3-5 audience sub-segments. The output of this is a short list of influencers/thought leaders who (i) are extremely likely to be disproportionately influential among my target audience and (ii) whose channels/socials I can use as a basis for understanding:
- How my audience prefers to receive information (stories, short snippets, whiteboards, explainers, case studies, demos/walkthroughs, etc.)
- The vernacular (are there certain terms or jargon that each of these influencers uses? Do they refer to certain tools/processes in certain ways?)
- Are there commonalities/trends around content length, production quality, narrative arc?
- Which brands, if any, are sponsoring content from these creators (hint: if none, you might want to consider reaching out)
I’m going to be brutally honest: actively watching 20-50 videos from 5-10 influencers isn’t exactly a dream afternoon, but the outcome is often worth it. On the off-chance you don’t have that kind of time, we’ve successfully used the ChatGPT 4.o with web search to analyze various channels and provide summaries of the tone, style, length, production quality and themes:
This is an example from Paid Media Pros, which was a commonly-found channel (and one I happen to adore) across this audience. As someone who has watched probably every video Joe and Michelle have ever produced, I can attest that this summary is exceptionally accurate; as an advertiser, each piece of information here is helpful in shaping how I would approach creating content for a similar audience.
While it is 100% true that most people could get to a similar list and analysis given sufficient time, one of the primary reasons I’m so bullish on audience research platforms like SparkToro, especially when combined with a tool like ChatGPT, is the insight + creative velocity it enables. I’ve already written about how more brands should make more ads – but this is only possible if you can spot the trends, understand how your audience is evolving and capitalize quickly. That’s where this shines. As Jeremy Irons said in Margin Call, “There are only three ways to make a living in this business: be first, be smarter, or cheat – and I don’t cheat. And although I like to think we have some pretty smart people in this building, it sure is a hell of a lot easier to just be first.”
In that same spirit, having the right audience research tools makes being first (or at least, close to first) significantly easier.
Tactic #3: Podcasts!
In last week’s issue, I spent an inordinate amount of pixels on the (previously) untapped power of podcasts – particularly as it relates to younger/more-likely-to-be-influenced audiences. Three readers all responded with the same question: how do I figure out which podcasts are worth investing in – either via pitching (to be a guest) or sponsorship/programmatic ads?
Once again, here’s where we use SparkToro (yes, this is one of the primary tools our PR team uses to identify podcasts to pitch). Using that same audience (CMOs), I was able to find 129 podcasts – ranging from major shows with thousands of listeners (the GaryVee Experience) to more niche shows like Tech Marketing Trends, For Immediate Release, Party Like A Marketer and CMO Convo.
As with above, rinse & repeat with different audience sub-segments, find the commonalities, short-list the results, and voila! A curated list of podcasts that your target audience is disproportionately likely to stream.
The next big question: what do I do with this information?
Well, head over to AudioGo, or Spotify Ads, or Amazon DSP, or Acast, get some audio ads made using the same formula you used for YouTube above (AudioGo charges something like $10 per professionally-recorded ad), then select either the specific podcast placement (available on Acast, Amazon, Basis, Tradedesk), or the right podcast list (Spotify + AudioGo) / Audience List and you’re off to the races.
At the same time, send your curated list over to your PR/content team and see which one(s) they can secure. For my money, there are few things I think are more impactful from a thought leadership standpoint than a podcast – not only are podcasts insanely flexible (I bring my microphone, headphones, travel light and 4k camera with me on every trip), they’re schedule-friendly: I can do 95% of podcast episodes in the middle of the day, with no disruption to my other meetings/obligations. It’s the most CEO/CMO/Executive-friendly thought leadership format, bar none.
And the best part? Once you’ve recorded that podcast episode, it becomes a marketer’s gold mine – one podcast can easily become:
- 2-3 1,000-word articles / blog posts
- 5-8 Threads
- 10+ LinkedIn / Instagram posts
- 3-5 Reels/Shorts
- 1-3 Webinars/Decks
- 8-15 Reddit posts/comments/responses
That’s a month’s worth of content from a one-hour conversation, that – in and of itself – is insanely valuable for reaching your target audience. Do 1-3x of these per week and there is ZERO excuse for not having an over-abundance of content.
Tactic #4: Reddit Ads
I’ve said – for over a year now – that I think Reddit is primed to become one of the most interesting expansion channels for brands over the next 24 months. Reddit’s prominence in search results has surged 6-fold in the past year:
(Image courtesy of Marie Haynes)
Over the past year, we’ve explored far more on Reddit than we had previously in response to this (it was always a fit for some niche brands, but most were patently terrified of Reddit) – but the challenge was always finding the right subreddits for our audience.
Once again, SparkToro has come to the rescue. We can use the same audience targeting (in this case, the CMO audience I’ve used throughout this article), and quickly surface the subreddits that are likely to be relevant:
What I particularly appreciate is the membership of the subreddit alongside the affinity – often, it’s the smaller (10,000 – 250,000) communities that are the most valuable. If you’re running Reddit Ads, you can target these specific subreddits; if you’re doing thought leadership, I’d focus my efforts on participating in/commenting on posts in these communities; if you’re interested in both, all the better.
Reddit is one of those niche cases where this tool is stupidly valuable, simply because most digital marketing / PR tools have woefully under-developed Reddit insight tools, and Reddit is an absolutely massive place with some very dark alleys – ensuring you’re starting in the right place is the easiest way to raise the floor of your Reddit efforts.
Tactic #5: Topics + Content Ideas/Gaps
Finally, there’s the million-dollar question: across all of these platforms, what specific topics should you focus on – either in your ads, pitches, articles, videos, podcast episodes, et cetera?
Traditionally, our approach to identifying the right kinds of content followed the self-scouting process I outlined here – we’d review our GA4/GSC for interesting trends/topics, scour our competition’s blog, socials, etc., list everything we found, then use a tool like Moz or BrightEdge to identify what’s appearing in SERPs related to those topics (and where gaps might exist), then refine it all down into a nice list of topics that fell into a few buckets:
- Sea of same-ness: A topic where there are many articles, all of which regurgitate the same general concepts/sub-topics. It’s likely most content is stale or not up-to-date, providing an opportunity for us IF we have a unique POV or take.
- Blue Ocean: There’s very little about this specific topic. This typically happens when something is just coming to market, or before a platform/tactic/topic emerges as a point of interest in our audience’s zeitgeist.
- Marina Topics: As the name suggests, these are the topics where high competitive density/where everyone has parked their proverbial boat. These are often incredibly competitive SERPs where ranking is insanely difficult, and where the reward for ranking often does not justify the cost of doing so. Our solution to many of these topics is to advertise on the informational queries to validate whether it’s even worth pursuing.
- Wrong Port: There are plenty of topics that might be of interest to our audience, but where we (or our client) does not have the domain expertise or experience to add value for our audiences. These are topics that we might use as audience segments in Google Ads (Custom Segments → add keywords), but where we will not create content or attempt to advertise.
Candidly, doing this on an ongoing basis is a LOT of work – but it was worth it because it provided the foundation for a wonderful, holistic marketing strategy. But, just this past week, SparkToro rolled out a “Topics” feature that is 🧑🍳.
Here’s an example, using that same CMO audience:
This is a veritable feast of content ideas – and as someone who knows this audience quite well and communicates with many CMOs/VPs of Marketing on a daily basis, I can attest that these ARE the topics they’re interested in learning more about. Many of the topics listed here relate directly to the pain points keeping CMOs up at night:
- Are we using our marketing automation effectively? How can we get more from our existing automations with ad costs rising?
- What is next in marketing? How can I be ahead of the curve and position our brand for continued growth?
- How do we become more data-informed? What tools, processes, skills, etc. are out there that could help my team better understand what’s working and communicate that to my fellow executives?
- How can we better share the bottom-line impact of our efforts with the broader organization?
- Should we be investing in influencers? If so, how? What else can we do to make the case that this is a worthwhile investment?
These are ALL conversations I’ve had in the last month with senior-level executives at companies across dozens of industries, sizes and scales. And it is incredibly easy to draw direct lines from these topics listed in the SparkToro report to (at least) 4 of those 5 pain points.
But what takes this from helpful/insightful to truly, mind-bogglingly valuable are the additional data points included:
- Affinity – This (as with Podcasts, YouTube, etc.) is a predictive measure of the relevance and level of interest of your audience to the topic. Basically – a higher affinity = a higher probability that your audience will be interested in the topic.
- Saturation – Measures the relative quantity and quality of existing content on the web that covers this topic. The higher the saturation = the more difficult it will be to stand out (unless you have some truly novel perspectives or approaches).
- Prompt – One of the things we’ve done with topic information (and which I covered in the self-scouting and AI articles) is use LLMs to generate outlines for proposed content pieces, so we can better understand what goes into creating that content / what sub-topics we should include / where gaps might exist. SparkToro actually integrated an LLM prompt into the product, so clicking on that little button generates exactly what you’d need to get started:
This is both insanely cool and helpful – it removes a step in the content research/creation process, and immediately helps you (or your team) understand what content will be necessary in order to create an effective piece of content (in whatever format – text, audio, video, infographic, etc.).
In addition to the “Relevant Topics” (which is extremely cool), there’s also a “related questions” (great for understanding FAQ content, or Q-and-A style content, or just understanding what your target audience is curious about:
(as it turns out, it seems like a LOT of CMOs are worried about making bad ads – which jives with the time of year).
And there’s Broad Interests, which are quite useful for getting to know an audience. In contrast to relevant topics (which tend to be specific), the broad interests are more general, and understanding them will help you better understand the big-picture themes a given audience is likely to be interested in:
What About My Customers?
The final thing I absolutely LOVE about SparkToro is the recently-added ability to upload a list of customers (either a full customer file, a customer segment, a segment of MQLs/SQLs, etc.) and run the same analyses I showed above, but using those individuals as the observed audience (instead of a bio keyword, search topic, etc.). There are some limits on audience size, but we’ve used this successfully for multiple clients.
The Upshot:
Media buyers + marketers are facing an increasingly fragmented environment – more users are turning to more sources for information. As that happens, I think more brands will diversify their marketing spend (I wrote about that here) – but the smart brands will diversify based on where their audience is. Tools like SparkToro (it is far from the only one, but it’s the one I use) help me (and our team) make smarter, more data-informed platform selection, targeting and content decisions.
There’s no doubt in my mind that digital advertising is going to become even more cutthroat over the next decade. The marketers that win will be the ones who integrate superior data into their overall strategy; while that data doesn’t guarantee a victory, having the wrong data (or, worse, relying on faulty data) can cause you to lose.
I hope some of this is helpful as you think about your holiday + 2025 marketing strategies. Give it a try – you might be surprised (or, at the very worst – you’ll have some concrete data to back-up your existing investment decisions).
Cheers,
Sam