Skip to Content
Article

What’s Next In Paid Media?

by Sam Tomlinson
September 3, 2024

Unlike summers past, the past three months have been jam-packed with changes impacting the ad ecosystem and broader ability of brands to market to their audience(s): from Meta rolling out its much-anticipated third-party conversion optimization, to Twitter (er: X) placing inventory through Big G, to a radically-altered Presidential campaign (for those of us in the US) and an economic rollercoaster (for pretty much everyone).

Regardless of your feelings on any of those changes, the reality is that they are upon us and their impacts will echo across the entire digital media space – from the costs we pay, to how we identify + reach audience(s), to how we evaluate the relative efficacy of our efforts. The challenge then becomes: what do we do? How do we gain an advantage?

And for that, I have three ideas:

1. Do More Audience Research

Audience understanding has always been the “X-Factor” that separates mediocre from remarkable. In an environment with more competition, higher CPMs, and more brands than ever before vying for attention, it’s gone from “important” to “imperative.”

I’ve worked with hundreds of brands and reviewed the work of hundreds of agencies. The common thread that unites the high-performers is an unfailing desire to better understand their audience – not just before the sale, but throughout the entire customer lifetime. Great brands and remarkable agencies evolve with their audience, not in response to them.

The form that research takes varies from brand to brand, agency to agency – but here are some of the specific types I’ve observed to be consistently effective and impactful:

Keyword Research:

Keywords may be an old, blunt lever – but old + blunt still works. At the heart of more than a few failed PPC campaigns is a failure to evolve their keyword list in response to audience behavioral changes.

The most successful Google Ads accounts are the ones that integrate search data from many sources (Search Terms Report, Paid & Organic Report, 3rd Party Tools, Customer Surveys, Competitors, Google Search Trends) to identify anomalies, untapped opportunities and areas for exploration AND have a structured process to introduce those new KWs into the account, validate their performance (or lack thereof) and properly organize (add/exclude/promote/demote) those terms within the account/campaign.

Customer Insights: 

Your customers/clients are an endless well of insight – but for that well to be useful, it must be tapped. We’ve consistently found success with a simple, automated process that any brand – B2B or B2C – can use:

  • Identify the 3-5 essential answers: This might be what other brands/competitors the customer considered before going with you; it might be what they searched for; it might be what set your product/service apart; it might be what feature/functionality/service they wish you’d offer.
  • Create a simple, one-to-one email from a senior member of your team: This works best if it comes from someone who is perceived to be a decision-maker – a C-suite executive, an owner/partner, a VP at the lowest. Leave room for personalization or specification – “First, I wanted to thank you for trusting us with your ___” or “I hope you’re enjoying your [product]….” It may seem trivial, but small details often lead to disproportionate gains in both response rate and response quality. From here, make a simple transition to the purpose of your email; something along the lines of “I’m obsessed with continually improving the quality of our products and ensuring every customer is delighted….” usually works. Then ask your questions – bullet points are preferred. This structure keeps the user focused and ensures questions aren’t missed. Close out by asking them if there’s anything that hasn’t been perfect or that your organization could do better (not part of the email per se, but you’ll be amazed at how many little issues you’ll uncover. Fix those and you’ll earn significant credibility from your customers).
  • Ask Users To Reply to the Email: Let’s face it: forms are impersonal. No-one wants to take surveys (and the people that like surveys are often not representative of the broader population). Asking for a simple reply to the email avoids all that – and makes it significantly more likely that your audience will respond.
  • Google Sheet, Zapier + ChatGPT: A little automation goes a long way. First, filter all response emails into a folder (Gmail/outlook). Then, configure a Zap to port the data from the email to a Google Sheet. Depending on the number and format of questions, you can likely then run a formatting script inside of the spreadsheet to separate out the answers. And finally, set up ChatGPT’s plugin (or a Gemini plugin) to run on the sheet to help surface insights. 
  • Schedule + Run: Most brands can get remarkable levels of customer insight by simply sending 1-3 of these emails per day. Pick a person from your customer file at random (excluding those who have already received a request) and send. If you’re a service-based business (B2B) with a limited customer file, I’d suggest adjusting the process to follow up with new clients 2-4 weeks following their onboarding.
  • Don’t Forget The People Who Didn’t Buy/Choose You: The insights you can gather from the prospects/potential customers who did not go with your brand are every bit as valuable as the ones from customers. A few small tweaks to the above email (“I see that you were considering us for [X] or you were considering [product Y], but have not yet made a purchase….”), a separate tab in your google sheet, and voila! You’re in business. 
  • Use the Data: The problem most organizations have is too much data and too few uses for it. This program is intentionally structured to avoid deluges of data – by sending only a few emails each day, there’s an opportunity for your team (or you) to respond to the ones with legitimate issues and address them. This makes your customer insights data doubly valuable: not only does it lead to a virtuous cycle on the marketing side, but it also acts as an early warning system – alerting you (or your team) to potential issues before they fester into something that could threaten a customer relationship.
Third-Party Analytics:

It’s no secret that I’m a proponent of audience research tools like SparkToro (no affiliation, just love for the product they’ve built). For the money, there is no better tool for understanding the sources of influence driving your target audiences. We’ve used SparkToro to inform media buys, to develop lists for PR initiatives, to identify podcasts for sponsorship, to identify + recruit influencers, to find content inspiration, and to craft targeting for YouTube, Meta + Google campaigns.

While this may be an unpopular opinion in some circles, broad-only is not the most efficient media strategy. Creative alone is not targeting, it’s “Zuck take the wheel.” The more data you can share with Meta, Google, TikTok, YouTube, LinkedIn, etc. to help it serve that creative to the right audience, the better off you’ll be. Smart people + smart machines = better outcomes. Note: I’m not saying that people can out-trade machines; I’m saying that machines with good information will out-trade machines with no information, and it’s our (media buyers’) responsibility to ensure that our machines have the data necessary to make smarter, more informed bids.

If you want to thrive in today’s environment, do more audience research. Not only will it help you craft more compelling, more effective creatives, but it’ll also help you place those ads where your audience is most likely to see them.

Most brands are woefully under-invested in their audience research – relying exclusively on intuition, gut feel, and subjective memory. And, for some, that may be sufficient. But, with platforms taking away control levers, costs going through the roof, and the rise of “privacy” initiatives, that’s no longer good enough. The organizations who invest in and build this capability now will be the ones that reap the rewards in the weeks, months and years to come.

2. Find Under-Valued Attention

It is said every year, but Q4 this year is going to be different. Not just because there will be more brands placing more ad dollars than in any year previous; not just because there are economic headwinds and increased competition and the lingering effects of inflation. All of that is true. But it’s incomplete. 

The 2024 election is also happening. It will be the most expensive election in the history of the world. Billions – likely $20B+ – will be raised and spent between Presidential, Congressional and State-level campaigns. Millions of hours of volunteer time will be used on everything from phone-banking and texting to door knocking. This isn’t just a case of ad markets getting more expensive (but they will); it isn’t just a case of disposable income declining (but it is); it isn’t just a situation where a large chunk of the population will have less available time and mental bandwidth – it’s all of the above, all at once.

Basically: a massive segment of the US adult population is going to be bombarded non-stop by election-related stuff for the next ~10 weeks – and a not-insignificant number of them will volunteer, donate, phone bank, tune in and post about it. Some brands will (inevitably) be caught in the crossfire. The rest will just see conversion rates dwindle while CPMs spike (every marketer’s favorite thing).

Now, to state the obvious: political campaigns aren’t known for their proclivity for innovation. A majority of political ad dollars will once again go to linear TV and direct mail. But a growing share of what’s left will be deployed via digital ads, concentrated primarily on Meta + Google Search (with smaller buys on X). Between normal brand Q4 spending + campaign advertisements, CPMs will spike, likely to never-before-seen levels. This impact won’t be evenly distributed, with battleground states (PA, GA, WI, MI, NC, NV, AZ) and states with highly-contested US Senate races (NV, AZ, MT, PA, MD, VA, WV, FL, TX, MI, WI) seeing the highest impacts. If you’re a regional brand that focuses on NY State or the Dakotas, you’re in luck; if your core customer base is in Georgia or Pennsylvania, gird your loins. 

But even in what will be the most expensive digital ad market in history, there will be under-valued attention. There will be impressions that trade at far below their value:

  • Twitter (X): Since Elon’s takeover, X’s advertiser base has collapsed. I recently launched a campaign with CPMs at ~$0.48. Clicks were ~$0.07. Conversion rates (subscription/sign-up) were 1.25%. The net-net: X produced sign ups at a cost per of ~$5.60. The same concept on LI was ~$50 per sign up (~$30+ on Meta). Quality was comparable. I’m not saying X is the dream platform, but when it’s a fraction of the cost of Meta pre-surge, it’s worth considering post-surge. 
  • YouTube: Unlike X, YouTube’s advertiser base isn’t hurting – in fact, it’s expanding at a rapid clip. But the rate at which watch time is growing is out-pacing advertisers budgets, resulting in significant value-buy opportunities. If you’ve had success on IG Reels, or TikTok, you can be successful at YouTube shorts. If you have the ability to put together an effective Meta ad, you can make an effective YouTube ad.
  • Podcasts: Podcast ads, without data, are difficult. But – armed with solid audience research – can present a wonderful opportunity. If you know a major segment of your audience listens to a specific podcast, the costs of sponsorship can quickly become less-than-onerous. A majority of the failed podcast campaigns I’ve seen are not a fault of the ad format or the podcast, but rather are a failure of audience research. It doesn’t matter how great the ad or the format if the right audience isn’t listening. But that failure can be your opportunity – especially if the show in question is struggling to gain advertisers.
  • Linear TV: Yes, it’s true: most political ad budgets go to TV. But political campaigns are often noncommittal on their spend – they make reservations, but don’t always follow-through on the buy, especially if a race appears lost or won. For savvy advertisers that maintain some budget optionality AND have robust audience insight (to know which channels/shows to track) this presents incredible opportunity. Now’s the time to make friends with the ad reps at both the national and regional/local levels. Without fail, every election year we receive dozens of emails and offers for last-minute ad slots following campaign cancellations. If you already know which channels + slots your audience is likely to watch (Sparktoro, again!), you can often score massive (50%, 75%+ off) discounts on high-quality inventory by simply being able to move fast.
  • LinkedIn: This one might seem odd (especially since LI has a reputation for being the among the most expensive networks around) – but LI is often a great value buy when the cost of every other channel (read: Meta & Google) spikes. Why? Because LI is largely immune from political ad and BFCM ad pressures. Put another way: relative to a “normal” Meta or Google, LinkedIn ad costs are expensive. But, relative to a Presidential Q4 Meta or Google, LinkedIn ad costs look pretty attractive.
  • Newsletter Sponsorships: Just like Podcasts, newsletter sponsorships are down bad this year; and just like podcasts, that means opportunity for brands that have the data and the resources to be discerning. 

These are just a few concrete examples of where under-valued attention can be found; based on your business, audience and resources, I’m sure you can find more. Every savvy media buyer or advertiser plays some form of moneyball: capitalizing on the value others miss. That’s all we’re doing here – finding value that everyone else has missed.

3. Leverage CRM + Offline Data

It has been 3.5 years since Apple rolled out iOS14.5. In the months following it, Google (Offline Conversion Import, Enhanced Conversions) and Meta (CAPI) unveiled a series of solutions designed to ensure that advertisers were able to (at the very least) track conversions from ads. 

To date, adoption of these tools is limited at best. Only about 13% of Google Advertisers using Optymzr (so, a particularly savvy sub-set of all Google Advertisers) are using Offline Conversions. I don’t have a ton of data on CAPI usage for Meta Advertisers (it’s relatively high for Shopify sellers, but that’s certainly not a representative segment of all Meta Advertisers), but I can’t imagine it’s much higher.

The simple fact is that data is your most valuable targeting lever across these two platforms.

There’s a material difference between optimizing for lead forms vs. optimizing for qualified leads – particularly in competitive verticals. There’s a material difference between optimizing based on actual retained purchases (i.e. purchases plus upsells less returns) vs. optimizing based on initial completed transactions. Most brands have too much data – to the point where they don’t use it. It just sits there, gathering dust and losing value.

The most common reason why brands don’t deploy this data to inform their marketing? They don’t know how. Getting anything to work in Salesforce (seemingly) requires a PhD or a rocket scientist; HubSpot is temperamental and more than a little clunky; Microsoft Dynamics is just an exercise in misery. But you don’t need a CRM to utilize these features. For less than $20/mo and a Google Sheet, you too can have offline conversions:

  1. Create a workflow from your website CMS (or form provider, like Gravity Forms) to Google Sheets.
  2. Add some columns for “qualified” or “not qualified”, type of service, type of product, whatever.
  3. Create a trigger to send an Offline Conversion whenever the relevant set of custom columns fits your fancy – i.e., send a “Qualified Sales Call” whenever the “Qualification” column is changed to “Qualified” and the Service Type is “Enterprise” (for example). 
  4. Shift your primary optimization action (Google) or Optimization Event (Meta) to the offline conversion.
  5. Be a hero.

To be exceedingly transparent, setting up offline conversions from website to Google Sheet to Google + Meta takes me about an hour (including troubleshooting). It’s likely to take most marketers a bit longer than that, especially if they’re unfamiliar with it. But this isn’t an all-day development affair; it’s a Ridley Scott film. It’s not overly complex, it just requires a little perseverance and a willingness to do the work.

What I can share – without equivocation – is that the impact of shifting from website-based conversions to offline conversions is staggering. It does take time for the ad platform to re-learn (a new optimization event will trigger a learning phase), but once you exit and the campaigns ramp up, the impact is remarkable. In one account, qualified lead rates went from ~11% to 62%. In another, calls booked increased by a factor of 4. Ad platforms are phenomenal at giving you exactly what you ask for – so if you’re only asking for form fills or people who make a single purchase, that’s what you’ll get. But if you use the above workflow to ask for more – qualified leads, attended demos, customers who purchased and didn’t return (or bought more!), subscribers who didn’t churn – that’s what you’ll get.

It won’t happen overnight, but better data is strongly correlated with better outcomes over the long term. Use your customer data – the best, most robust, most differentiated data you have – to help ad platforms work more efficiently for your business (or your client’s business). 

Each one of these is a relatively small thing. But together, and combined with diligent management, clever strategy and robust ad production, they will help you future-proof your marketing efforts.

That’s all for this week! I hope you all have a wonderful, restful, relaxing, Labor Day weekend!

Cheers,

Sam

Related Insights