ArticleWhere Is Paid Search Going… & What Do We Do About It?
After speaking with 50+ people (in-house, agency, brand owners, SaaS, operators) across a trio of speaking engagements (two at SMX Munich, and a keynote at SMX Paris), it is abundantly clear that one topic is front-and-center in everyone’s minds: where is paid search going?
Even more recently, Sam Altman (he of ChatGPT fame) weighed in on the topic on the Lex Friedman podcast (you can view the relevant excerpt of the 2.5+ hour interview here), where Altman made several interesting points/claims: (1) he’s interested in making a better system for search, not merely another Google; (2) ads create a suboptimal user experience across the internet (not just on Google, but on Instagram, Facebook and informational sites); and (3) people are willing to pay for a better, non-ad-influenced experience, citing ChatGPT’s $20/mo model as evidence. Boiling it down: Sam Altman thinks there’s a future where paid search is no longer part of the overall online ecosystem.
All of this leads to an interesting question: Is there a future for paid search?
In an effort to not bury the lede, I’m going to jump straight to the conclusion, then talk through a few key points that marketers should keep in mind as they think about where search is going and how to prepare for that future.
Is Sam Altman Right? Is Paid Search Done?
I think the strongest argument against Sam Altman’s “subscription” concept is simple economics: paid search, as it currently exists (so, before any improvements from LLM integrations) is a more effective + more efficient value capture mechanism than any subscription model could ever hope to become. To move away from ads would cost tech giants hundreds of billions a year – and likely result in their businesses collapsing. That’s….not something they’re going to do.
To illustrate that in a more concrete way, Google’s total revenue for 2023 was ~$307.4B (source); of that, $175.04B of that is from Google Search (categorized as “Search & Other” on all financial statements)(source). Google represents ~60% of total search, including retail media (~$115B est. 2023 – Insider Intelligence), other platforms (Microsoft, TikTok, Instagram search), etc. Put it all together, and global “search” advertising revenues can be roughly approximated at ~$292B +$28B MSFT + Meta + TikTok + YouTube (ads on SERPs only, estimated) = $320B USD for 2023. I think the above is relatively close to accurate, while erring on the side of being conservative.
If paid search were to be replaced with a subscription-only model, somewhere in the neighborhood of $20/mo/user, and if every US + European Union/UK resident over age 14 (~720M people) immediately subscribes at that price ($20/mo), the resulting revenue would be approximately $172.8B, or 46% less than current search revenue. While a 100% subscription rate is nowhere close to realistic (least of all because a whopping ~50% of Americans have less than $500 in their savings accounts), I have disregarded APAC, Latin America, the Middle East & South America, so the end point is probably in the vicinity of realistic. There will be cases where people have multiple subscriptions, and there will be corporate memberships, and all of that probably adds another $35B to this total – which is still ~$110B short of 2023 search ad revenues. In the end, that’s simply not a sufficient amount of revenue to sustain a product of this caliber – and so, either (a) search as we know it ceases to exist or (b) ads return. (a) isn’t viable, so (b) it is.
Not only does the above model not make economic sense for the platforms providing search as a service (a whole new kind of SaaS!), but it also doesn’t make sense for market participants (businesses, organizations & individuals), who now have (1) limited insight into what factors are driving recommendations (this isn’t going to be altruistic; it’s going to be heavily influenced by LLMOs (Large Language Model Optimizers – the next evolution of SEOs) and (2) limited capacity to reach their target audience.
If Sam Altman’s vision were to become reality, this isn’t just bad for platforms, it’s a death sentence for small + startup businesses, too. Without ads (how many businesses go from 1-100, once they’ve exhausted friends/family/immediate contacts), those businesses will either (a) fail, (b) turn to PR (another gatekeeper), (c) use traditional advertising or (d) attempt to “optimize” their way into LLMs in short order – via legitimate (or illegitimate) tactics. None of those avenues is as scalable, capital-friendly, transparent or as effective as current digital media advertising.
All of this leaves aside another simple reality: different advertisers in different regions derive different values from different searches – a plaintiff’s attorney in Texas (for instance) might value a click for a specific, niche injury term at $5,000 (yes, I’ve seen CPCs as high as $9,000 for certain hyper-specific, wildly-lucrative case type terms), while a small eCommerce seller in Jacksonville, FL might value a click at $2, an accountant in Fairfax, VA might value a click from a potential new client at $50, and a SaaS platform operating in the US and Canada might know that they can profitably acquire new subscribers paying $125.60. These are just a few examples that illustrate the underlying point: different users, conducting different searches, have different values to different businesses/organizations.
Search advertising captures this value. Fred Vallaeys (one of the original Googlers who helped develop Quality Score) wrote in his book, Unlevel the Playing Field: “Google’s Flagship, state-of-the-art bidding strategy, Maximize Conversion Value, operating at optimal efficiency with perfect information, will deliver exactly $0 risk-adjusted profit.” Put another way, modern ad auctions are absolutely masterful at capturing surplus value from the organizations receiving it (the businesses doing the advertising), while encouraging more of the value creation by not charging the users.
In a post-ads world, the only way to capture that value would be to pass it onto the subscriber.Will ChatGPT dynamically adjust my subscription price based on the content of my searches? I don’t think any of us want a random $1,000 invoice from ChatGPT because we searched for a plaintiff’s attorney after being in an accident, or a $500 bill because we were looking for flights, hotels + rental cars to visit a family member or go on vacation – but, in this model, that’s the only mechanism available to capture the value created from providing the information to the subscriber.
Eliminating search advertising would result in the platforms (the search engine, or the LLM, or whatever) that are creating this value (by directing users to specific businesses, organizations and individuals) losing out on the vast majority of it. Creating value, without sufficient capture, is a recipe for disaster.
I believe – and placed a substantial bet – that Sam Altman is wrong. Ads play an essential role in subsidizing the free, open internet that we all enjoy (and rely on for our livelihoods), as well as ensuring that brands can connect with their target audience to help them discover new, innovative and differentiated products. At the end of the day, that’s what advertising has done for 500 years, and what it’ll do for the next 500: help people find what they need, want and desire – whether or not they know it. Formats will change (more on that below), platforms will change, mechanisms will change – but the notion that advertising can be supplanted by subscription is fundamentally misguided.
As long as there’s a need to find information on the internet, there’s a need for search. And as long as there’s a need for search, there’s an economic need for search-related advertising.
So, Where Is Paid Search Going?
For much of the last ~20 years, search has been defined by bundling: Google absorbing a progressively larger share of total search – from shopping to flights, hotels, local businesses, apps, etc. The end result of that has been Google as the single, dominant player in online search (a claim the US DOJ and UK CMA have both made at various times). This has resulted in search becoming synonymous with Google – both in the minds of consumers and advertisers.
But there’s ample evidence that the era of bundling in search is over: retail media now accounts for over $100B in annual ad spend (the vast majority of this is search); Amazon has become the largest shopping search engine; YouTube and TikTok are capturing progressively larger shares of consideration/research/exploratory searches, with each platform handling billions of searches each month; even Meta, Reddit and Quora are seeing growth in searches on their platforms. All of this is happening within the context of ChatGPT, Gemini, Claude (which has out-performing GPT-4 according to a March 4, 2024 release) opening an entirely new mechanism for users to find answers.
This should hardly come as a surprise to anyone who has been paying attention to consumer behavior and trends: search, as it was implemented by Google – was always a flawed way to access information. Not every search needs 10 blue links. Google has tacitly conceded this point with the rollout of progressively more SERP features, each one a confession that traditional web search is not the appropriate mechanism to respond to classes of queries.
Unfortunately, in the minds (and budgets) of most brands/marketers, neither Amazon, nor Meta, nor YouTube, nor TikTok are considered “search” platforms – despite the fact that each provides the functional equivalent in certain contexts.
All of this leads to a single, inevitable conclusion: the future of paid search is broader than Google.
For PPCers, that means four things:
- Search is Full-Maze: historically, search has been a bottom-of-the-funnel, end-of-the-maze tactic well-suited to demand capture. Today’s users exhibit more information-seeking behavior than ever before, and finding that information is more difficult than ever before (thanks, in large part, to information pollution). PPCers need to think broader + more holistically about the role search plays in information discovery.
- More Complexity: the days of search = Google (and maybe Microsoft) are coming to a close; PPCers who only know Google will find themselves on the outside looking in, sooner rather than later. Today’s users will search on Reddit, TikTok, YouTube, Meta and more – so if you don’t have the ability to parse through each of these platforms, you’ll miss progressively more of the user journey (along with a ton of opportunities).
- More Conversational + Natural: one major change I’ve already observed in our search terms report is the rise of longer, more conversational queries. The easiest way to visualize this is to sort your paid + organic report by word count + time – I’ve seen this distribution shift toward longer queries over the past year. From a paid search standpoint, this represents an opportunity to capture incremental value from queries other marketers are missing.
- New Skills Required: the combination of more complex search journeys + more automation in search platforms means that PPCers need to get better at both managing Google AND using other platforms. The days of being able to micro-manage Google are over; the role of PPCers has evolved from lever-puller to marketer….which is something that may scare a lot of PPCers.
This should come as welcome news for most PPCers: (paid) search isn’t going anywhere. Paid search is bigger, broader and more complex than most marketers believe – all of which means more opportunities for the marketers + brands willing to rethink how search works. That also means that many organizations will need to fundamentally rethink their budgeting and capital allocation process: after all, customers don’t search based on your P&L or departmental silos – if your paid search team can’t coordinate with your retail media, paid social and YouTube teams, they’re going to miss opportunities.
Let’s go one step further: these distinctions (paid search, organic search, YouTube, paid social, programmatic, retail media) will collapse in the next 3-5 years. Changing user behavior demands it, and automation + platform interoperability enable it. The brands that embrace this change will be the ones who win tomorrow.
LLMs, ChatGPT & Search
SGE has rolled out to ~20% (link) of SERPs in the US as of February, 2024 – with little sign of slowing down. Google’s latest LLM (Gemini 1.5 Pro) is exceptionally good, as is Claude 3 – but true adoption (not just users, but people using it 10 or more times per day, every day) is still not at a tipping point.
That may fly in the face of claims made in the Twitterverse/LinkedIn, but that’s reality. By most estimates, only 1% to 5% of ChatGPT users are currently paying $20/mo for Premium. That jives with Google’s announcement that 100M users (out of ~2.5B) have subscribed to Google One (and some subset of them are using it for Gemini 1.5).
I firmly believe Google will roll out multiple ad formats unique to SGE in the next 12 months; we’ve already seen SGE-powered gifting recommendations, shopping ad units integrated into SGE results, one-stop-shop deal destinations, and new filters. I would expect, based on what we’ve seen so far, reasonable inferences from current ad unit functionality, and Sundar Pichai’s comments during Google’s most recent earnings call, that we’ll see multiple new ad formats:
- A LSA-type ad unit, integrated directly into SGE. This makes a ton of sense, and aligns squarely with Google’s desire to act more as “agent” vs. “engine”
- Some type of lead form / lead generation unit for non-LSA businesses, also included within the SGE
- More recommendation-type ad units – there’s a precedent for these with the SGE-powered gifting recommendation ad units rolled out in Q4 2023, so I’d fully expect to see Google lean into this, and potentially even create a campaign sub-type (similar to what they’ve done with video ads) for advertisers.
Note: these are predictions – I might be wrong!
All of this to say: SGE is firmly entrenched in search, and will be rolled out to a progressively larger share of SERPs over the next year. That means more (and more interesting) opportunities for paid search – not fewer.
So, What Do We Do?
I’ve spent a considerable amount of time thinking about how PPCers should respond to these changes, and I’ve come up with four things you can do:
1. Build a Remarkable Foundation:
More than ever before, a brand’s foundation is critical. That’s more than a logo or a tagline; it’s a philosophy. The brands that win tomorrow will be the ones that:
- Win with remarkable experience + emotional resonance, not price + discounting
- Acquire customers based on value, not discounts
- Build long-term, stable, mutually-beneficial relationships with their customers
- Obsess about incremental impact and unit economics, not ROAS
In many ways, this is old advice – but somewhere along the way, it got lost. What’s old is new again – it’s true in sneakers and in search. This starts with how we keep score: moving from vanity metrics to business/organizational metrics:
But it doesn’t end there.
As I wrote a few issues ago, the online obsession with discounting is a cancer that should be eradicated. The brands that win in tomorrow’s more diverse, more complex, more competitive search ecosystem will be the ones that can acquire customers without resorting to discounts, gimmicks and other BS.
2. Get Brilliant At The Basics:
The bar for remarkable always rises. This is true in sports (an ordinary play in today’s NBA or NHL would’ve been a non-stop highlight reel in the 1980s), and it’s true in marketing:
There are still PPCers out there (and I ran into a few) who have deluded themselves into thinking they can out-trade machines. And, maybe they can for an hour, or a day, or even a week. But I look at marketing investments over months, quarters and years – and there’s no-one alive who can out-trade a machine over that horizon.
Today’s marketers are better off fighting WITH (read: alongside) machines, not with (read: against) machines. Doing that means embracing a role as strategist and general, not lever-puller.
Automation has removed many of the controls PPCers have historically used to direct machines, which has pushed optimization to the data level. The importance of exceptional data has never been higher, and it will only increase. Engines now have the ability to act on all the data in their view – so if you’re only using engine data (i.e. Meta or Google audiences, interests, etc.), you don’t have an advantage. The advantage comes from layering engine data with your business + customer data – thereby giving the machine the data it needs to find alpha in an increasingly competitive market.
If you’re curious about how that manifests itself into ad accounts, here’s four ways you can begin to make this transition:
- Segment only when Necessary: the days of hyper-segmentation are long-since over; a modern SEM account structure uses segmentation only when the incremental return from differentiated messaging/landers exceeds the cost of data fragmentation.
- Fuse Active + Passive Targeting: keywords are an old, rusty lever; modern SEM looks more like a DSP, with active and passive targeting levers. Audience-based search targeting is the future – and I wouldn’t be surprised to see Google embrace the role of Search DSP to pull in more retail media placements, either.
- Think Guardrails, Not Guidance: The best PPCers I know operate from a principle of “eyes on, hands-off” – they structure an account well, do plenty of research, continue to feed strategic insights – but they don’t try to day-trade. Modern SEM accounts are barges, not speed boats.
- Solve For the Right Thing: use data passback (offline conversions, CAPI) to focus engines on the one thing that matters to your business (not just a sale, but contribution margin; not just a lead, but a qualified opportunity; not just an app download, but an actual engaged user). This is a concrete example of integrating your business data (qualified lead, contribution margin, engaged user) with engine data (conversions, audiences) to help engines find alpha.
3. Embrace The Role of True, Human Creativity:
While the future of AI / LLMs is bright (and we have barely scratched the surface), the harsh reality is that most of these tools are currently used to enable mediocrity, not create something truly remarkable.
In a world where generative AI has enabled anyone, anywhere to create something that’s perfectly adequate, the value of something truly human and truly remarkable is staggering. From a PPC perspective, there’s a massive opportunity to stand out with remarkable, meticulously crafted, emotive content – because most brands are just pumping out “meh” – and quantity does not drive quality.
But to do that, you’ll need to do the work – that means conducting exceptional user research (hello, Sparktoro), understanding what search engines think is remarkable (this is a great use case for ChatGPT or Gemini – give it 5-10 articles that rank for a given topic, and ask it to summarize the commonalities + differences between them), and connecting those data points to your brand in unique ways that add value for your users. In other words, creating great content – the thing Google’s been saying to do for a decade.
4. Zig When Everyone Else Zags:
I’ve often said, both at conferences and to our team, that the default state of everything is failure. It’s true for businesses (if you don’t do anything, your business will fail), relationships, ad accounts, you name it. It’s the business equivalent of the 2nd law of thermodynamics (don’t tell me you didn’t pay attention in chemistry or physics!): entropy (disorder) always increases – unless we act to reduce it.
The work we do each day in ad accounts or on businesses can reduce entropy – but not all work is created equal. The relative efficacy of the work we do is impacted by both our skill AND the differentiation of that work in the marketplace – which is just a fancy way of saying that if everyone else is doing what you’re doing, at the same level you are doing it, the ability of that activity to reduce entropy (read: failure) will approach zero.
But, if you do things that everyone else does not do, or you do things with sufficiently greater skill/proficiency, the surplus value of that activity will be higher – and higher surplus value = more positive impact.
The easiest way to implement this is to win what’s next:
That’s (obviously) easier said than done – but it tells you what you must do. Find undervalued attention, find emerging platforms, do things that don’t scale and that aren’t talked about by the Twitter/LinkedIn Ad Bros.
That’s where the incrementality is. That’s how you fight failure.
There’s a big, bright future for marketers who embrace these principles, build the skills required to succeed and are willing to rethink how they work.
For those that aren’t, these changes are the equivalent of a meteor….and we saw how well that ended for the dinosaurs. But advertising, like life, will find a way (to quote Jeff Goldblum).
Until next time,
Sam
