ArticleIssue #173 | The Spotlight & The Floodlight
We recently audited a brand in the healthcare space that, by any outside assessment, should have been crushing it.
They are the undisputed heavyweight kings in their space. They’ve completed more complex procedures for this niche issue than any other provider out there. They’ve changed (at least) 20,000 lives. They’ve received ~3,200 5* reviews. Every industry award there is to win, they’ve won it (usually at least twice). The face of the brand is charismatic and present, with glowing media coverage from New York to LA and back. In their little slice of the healthcare economy, they are the 800 lb gorilla. No competitor comes close – not on results, not on reputation, not on the sheer variety of proof, not on founder star power.
To top it all off, the ad account was structured and managed well. There was a logical organization, notes, good data flowing back from their appointment system. They had relevant exclusions in place, they were smart about placements, the account was updated 2-3x a week.
But – somehow, despite having literally everything going for them – the performance was pedestrian at best.
It was a puzzle. Every obvious (and most non-obvious) explanation failed.
It was not a management problem; the account was run competently, in some places exceptionally. It was not a proof problem; the company had more credible proof than anyone in the category. It was not a lack of ads – there were 800+ in the account. It was not a budget problem or a targeting problem or a data problem. It wasn’t a lack of an engaging, charismatic personality or a founder unwilling to advocate for the brand.
Then we started digging deeper into the creative. We sorted the hits from the duds – and a pattern started to emerge: the ads that leaned on the record underperformed.
Ads featuring “20,000+ lives changed” underperformed. A showcase of the industry awards was so bad it was paused. A compilation of dozens of quick-cut glowing reviews from patients did nothing. 99% success rate (something unheard-of in their sector)….barely passable.
Everything that most marketers would tell you to lead with, sucked.
The rare winners were odd: a single patient, on an iPhone, recalling her experience. An ugly-as-anything testimonial, with a few words bolded for emphasis on a background that made my eyes bleed. A quick clip of one of the other providers talking about the experience walking into the practice.
From a marketing perspective, that’s weird. None of the winners had a “moat” at all – they were generic enough that any competitor in the category could have produced a near-identical variant. Over and over again, the brand’s most differentiated asset – the staggering volume of proof + the sterling track record built over decades – was getting whalloped by ads that anyone could make with an iPhone, Canva and 20 minutes. The stuff only this brand could say was losing to the things any brand could say.
This is one of those things that I didn’t believe could happen until I saw it – time and time again. No marketing class or textbook prepares you for it, because it runs counter to everything you’re taught in school.
The specific beats the generic. The one beats the many. The story beats the statistics.
When we see it in accounts, the temptation is to chalk it up to a vague platitude – “people connect with stories” – then move on. But “people connect with stories” is not a mechanism. It is a restatement of the thing we are trying to explain – and because it explains nothing, it is useless. It cannot tell you when the specific will win, by how much, or what to do in the cases where no story is available. If you want to do that (which, I think, is far more interesting & exponentially more useful), you have to go explore Bernbach’s “Unchanging Man.”
That’s what this issue is all about.
Let’s start with something we’ve all known for a long time, but it took Danny Kahneman & Co to prove: human attention has two settings.
The first is a floodlight: a wide, even, low-intensity beam that lights up an entire field. It is what you reach for when you want to cover everything – the full roster of features, the whole customer base, every reason to buy. It feels responsible and complete, like doing the job thoroughly. The second is a spotlight: a narrow, high-intensity beam that lights up a small, defined area to the point where you can see every detail – but it leaves the rest of the field dark.
Most marketers operate on the belief that persuasion is a floodlight problem: that the job is to get as much true, favorable information into the lit field as possible. If you just have enough testimonials and proof points and features and statistics, and whatever else, then eventually your target audience will see reason and buy your stuff/request that demo/submit the lead form.
That belief is precisely why the 20,000+ lives changed ad creative existed in the account…and it is wrong in a specific, measurable, repeatedly demonstrated way: the human system that actually makes the buying decision does not have a floodlight. It has a spotlight. Nothing else.
Every dollar spent trying to scatter little bits of proof or features or differentiation across the whole field is spent on an activity that the prospect is highly unlikely to ever recognize or appreciate.
Let’s Talk About Birds
(no, not the Alfred Hitchcock masterpiece, though you should watch that).
In the early 1990s, a group of researchers led by William Desvousges ran what looks, on the surface, to be a dry exercise in environmental economics. They asked 3 groups of people how much they would pay to keep migratory birds from drowning in uncovered oil ponds (if you’re unfamiliar, oil ponds are blobs of oil sitting on the surface of the water that birds mistake for actual water). The only thing that varied across the groups was the number of birds at stake: 2,000, 20,000 or 200,000.
The willingness-to-pay were $80, $78 and $88, respectively.
This is one of those so-wild-you-can’t-believe-it findings. The scope of the problem rose by a factor of 100 (from saving 2,000 birds to 200,000 birds), but the amount people would pay increased by ~$8.00 (which is just statistical noise). A 100x change in stakes produced no change in response. That phenomenon – scope insensitivity – is one of the most replicated and least internalized findings in behavioral economics. And, for our purposes, it is a concrete demonstration of exactly what happened in that ad account.
If you examine this experiment through that lens, the parallels are there. Each group was handed a bigger floodlight. The 200,000-bird group was asked to illuminate a field 100x larger than the 2,000-bird group. If persuasion was determined by coverage, on the quantity of stakes brought into the lit field, willingness to pay should have increased substantially, yet it didn’t. The extra 99% of the field was lit only in the trivial sense that subjects could read the number off the page. It generated no incremental emotive response or feeling – and feeling is what triggers the impulse to buy and the price you’re willing to pay.
This is not a random quirk about humans hating birds (though again, if you watch the Birds, you just might). Danny Kahneman replicated the finding when he asked people what they would pay to clean up polluted lakes. The result? People will pay barely more for all the lakes in a given region than they will for a small, named fraction of them.
And the same finding holds when the stakes are human lives (yep, people studied that). If you’re curious, the title was wonderfully blunt: Insensitivity to the Value of Human Life. In it, the researchers found that our perception of lives at stake follows Weber’s Law, which is the same logarithmic compression that governs how we perceive sound, weight & brightness. We do not perceive the gap between 10,000 and 100,000 the way we register the gap between one and two. The just-noticeable difference is a constant fraction of the whole, not a constant amount. Stakes that grow by multiplication register, if at all, by addition (yet another proof point that most people are terrible at math). The human floodlight does not brighten as the scope of the field grows – the light per unit simply dilutes across a larger area.
The Prototype
Kahneman’s explanation for why the floodlight fails is simple: most people can’t construct a mental representation of 200,000 birds or 20,000 birds or 2,000 birds. In all 3 cases, the number is too large to render – it’s just a number. So, faced with that challenge, our brains do the next best thing: substitute a single instance we can visualize – the “prototype”.
One exhausted bird, feathers matted with oil, struggling at the edge of the pond. Then, it prices the feeling that one bird evokes. The number behind the single bird, the 2,000 birds or the 200,000 birds is a statistic our reasoning mind can comprehend, but not a quantity our affective system can actually use. Faced with an overwhelming number, we anchor on the one bird we can visualize and price that – the 199,999 (or 19,999 or 1,999) other birds are completely irrelevant.
That one observation explains the finding: the prototype is identical across all 3 conditions. One bird is one bird. The feeling saving that one bird generates is roughly constant, and the feeling – the affect – is what establishes the price. The spotlight rendered one bird in every instance, because one bird is all a spotlight can render. The experimenters kept trying to force people to use a bigger, wider floodlight, but the people kept switching back to the spotlight.
Before we get back to applying the lessons from birds, lakes & human lives to your ad account, one more little bit of behavioral economics. For Kahneman, the “prototype” was something specific – it is not the average customer, not the representative case, not the aggregate shrunk down. It is a single, concrete, fully specified instance, vivid enough for the imagination to inhabit. The prototype of “lives we changed” is not “20,000 people.” It is a single person, with a name and a face and a story and a smile. No number, no matter how big or impressive, can be a prototype because the vast majority of people have no way to imagine it.
The other finding – and the one that more marketers should obsess about – is the claim that affect determines purchasing power/willingness to pay. What Kahneman & Desvousges independently found was that people often pay not in proportion to the good they accomplish, but in the amount required to produce a warm glow or a sense of having discharged a duty. This is called the purchase of moral satisfaction: the amount you’ll spend to feel you have done something has almost nothing to do with the birds or the lakes or the people.
What the Spotlight Actually Is
In Kahneman’s vocabulary, the spotlight is System 1: fast, automatic, affective, always running, never asking permission. In that context, “renders one bird, prices the feeling” is the definition of attribute substitution. When you give System 1 a question it cannot answer – such as, “How much do I value saving 200,000 birds?”, it doesn’t return an error – even though the scope of the question is far beyond its capability to answer. It does what AI systems do: swaps in a question it can answer (say, “How do I feel about saving this one bird?”), answers that instead, and returns the result confidently, as though it had answered the original.
The result? The valuation is a confident reply to the easy question, mislabeled as a reply to the hard one. Behavioral economist Paul Slovic called this the affect heuristic: when a judgment is hard, we consult how we feel, then answer off the feeling. Scope insensitivity is that substitution in action – most people value saving a bird from drowning in oil at (plus or minus) $80.
The obvious next step is to call the floodlight System 2. That’s…not quite right. The floodlight is information addressed to System 2, for a decision that System 1 is going to make without ever consulting System 2.
The 20,000+ lives changed ad is the predictable result of rationalizing – that process where we make a decision with System 1 (a/k/a the spotlight), then ask our reasoning, rational System 2 brain to come up with a justification for why whatever we did was a good life choice. That entire ad is a System 2 artifact, built by the marketer’s slow, careful, evidence-weighing mind.
The problem – in this case – is that the ad was designed for an audience (System 2) that doesn’t show up in the buying moment. The only system present is System 1, and it only cares about that one bird.
Ironically, most marketers – including the ones running this account – make the same type of error (attribute substitution), just in a different way. Instead of swapping the single instance for the big number, marketers swap the hard question (What will make this audience feel something?) for the easy and flattering one (What is the most impressive true thing about my client/brand?). Both parties run the same heuristic on the same ad, but substitute on different attributes. The prospect can’t comprehend the aggregate, so they feel; most marketers would rather not sit in the discomfort of engineering a feeling, so they reach for a fact. In the end, neither gets what they wanted (kind of like the ending to the Birds).
The Spotlight & The Creative
When you apply all of this to the account we began with, the puzzle pieces start to fit into place: the generic testimonial won because it handed the spotlight a single target: one person, one face, one fully rendered prototype the affective system could empathize with. In that moment, it did not matter that 20 other brands could have run an identical ad, because the prospect wasn’t comparison-shopping. To put it bluntly: the target audience was either feeling something about one person or feeling nothing about a number – and a rendered person, with a story we can imagine, beats any number asked to do emotional work every day of the week.
Once you internalize that, it changes how you approach creative moving forward.
The prototype must be the creative unit.
The job is not to communicate the aggregate, because it is unfeelable by construction. The job is to select and render the single instance the spotlight will lock onto, and then to make sure that instance carries the meaning you need it to carry. Your customer/prospect can’t feel your install base, your case count, your years in business or your market share….but they can feel one person who looks like them, facing what they are facing, coming out the other side. Bernbach understood this in his bones decades before behavioral economics became a thing.
Now – there’s been a lot of theory here, so let’s start connecting pieces back together. Here’s an exercise I’ve found helpful in understanding if you have a floodlight problem.
Pull your top-of-funnel creative and sort every asset into two groups: the ones that ask the prospect to feel something about one rendered instance (specific testimonials, reviews, etc.), and the ones that ask the prospect to be impressed by a field, a feature set, a customer count, a wall of logos, a span of capabilities.
The first group is your spotlight ads; the second, your floodlight. Label them. Both have their uses, but it’s often helpful to see just how many of your best performers concentrate in Group #1.
Is “More” Really “More”?
As you turn your attention back to your account, you’ll notice that scope insensitivity does something that invalidates every marketer’s most common optimization instinct: the urge to escalate.
If perception of stakes is logarithmic, bigger claims don’t scale linearly in persuasion. They scale sharply sublinearly, then flatten. Case in point: “save $5,000” is not 5x as motivating as “Save $1,000.” Depending on the context, $5,000 is fractionally more motivating – but past a point, it is virtually meaningless. Doubling your proof points, inflating the discount, stacking urgency, growing the number of thrilled customers from 20,000 to 200,000 – each one runs headlong into psychophysical numbing faster than the person/agency making the ads expects.
This is why that sterling brand’s strategy was doomed from the start, regardless of execution: no number was large enough to make the floodlight work, because the problem was never that the number was too small. The floodlight is the wrong tool for the job, and a bigger, better wrong tool is still just a wrong tool. You cannot escalate your way out of a category error.
This is hard to accept, because escalation feels like progress. A bigger number feels like a stronger claim, and a stronger claim feels like it should be more compelling. But, the evidence says the pull tops out almost immediately and the rest of the magnitude is wasted.
But, there’s hope: the same research that produced the flat bird curve identified at a lever that works – though few people use it. The lever in question is the denominator. A program described as saving 4,500 people from a camp of 11,000 drew substantially more support than the identical program described as saving 4,500 from a camp of 250,000. In both cases, the lives saved are the same – but the response is not, because humans don’t price the absolute good; we price the portion of the visible whole that gets rescued. A large fraction of a small, intelligible number feels like a moral victory. A small fraction of an incomprehensibly large group feels insignificant – like a rounding error…even when that rounding error is the same number of saved lives.
The fix is simple: stop presenting the win as a small fraction of an enormous, abstract whole: “$2B recovered for victims,” “300 communities served,” “millions of customers worldwide.”
That framing maximizes numbing, because it pairs an ungraspable numerator with an incomprehensible denominator and asks the spotlight to feel a ratio it can’t even conceive of. Instead, present the win as a large fraction of a small, concrete whole. A firm that says we take 8 cases a month, and this month we won this one has built a denominator the mind can hold, and the won case becomes a large, vivid fraction of a graspable 8 vs. an invisible sliver of $2B. A senior-living community that shows one resident’s actual afternoon has done the same thing, collapsing the denominator from “300 communities with 20,000 residents” to a single, specific person/room/experience.
We actually ran this very experiment in an account a bit ago. We created 2 ads: one that showed stats/features, and a second that pulled out a quote – “He Is Thriving” – with a brief excerpt from a family member. Everything else – the offer, community, CTA – was identical. In no surprise, the specific won in a landslide (75% higher CTR vs. account, 2x higher CVR, 43% lower CPL).
If you want to apply this to your account, go pull up that list of “Floodlight” ads. Find every place you reached for the largest available number and ask whether a smaller, more graspable denominator would make the same claim more emotively viable. In a surprising number of cases, scale is working against you. Shrinking the frame is staggeringly powerful, because it turns something that’s incomprehensible into something relatable, all without having to re-do everything else. Instead of going big – $2B+ recovered for victims – go small: $52M recovered for 5 Oklahomans this month. This does two things: (1) it makes the denominator real (everyone knows 5 people) and (2) it is simple enough that System 1 can do the math – “these guys got $10M for 5 people just like me, last month?”
The Antagonist
There’s one final build on this concept I want to share with you – and like the rest, it comes with some science.
Dr. Anneke Buffone (who formerly spent 7 years at Meta working on excessive use + age gating) and Dr. Michael J. Poulin published a fantastically interesting paper – Empathy, Target Distress, and Neurohormone Genes Interact to Predict Aggression for Others—Even Without Provocation (talk about a mouthful). In it, they demonstrate that empathy for a person in distress predicts aggression toward whoever stands in that person’s way, even when the 3rd party has done nothing to provoke it. The underlying principle: care for a specific person converts into hostility toward whatever obstructs the cared-for person. And what’s more – the obstacle does not have to do anything wrong – merely standing in the way is sufficient.
This is fascinating to me because it provides another science-backed angle you can add into your accounts: the antagonist structure. It starts with a sympathetic protagonist – someone your target audience can empathize with. That person should look like them, talk like them, share details that are generic but feel specific (basically, how Tarot card readers work). Then, add in a 3rd party who is simply in the way: the insurer, the legacy bank, the crappy incumbent, the status quo that has been failing people like the protagonist for years.
Put the spotlight on a protagonist, name the obstacle between him/her and relief, and the diffuse warmth the prospect feels for the protagonist resolves into directed, motivating animus against the obstacle. The spotlight – aimed well – can make your target audience simultaneously love the thing it illuminates and loathe the thing casting shadows off to the side.
If you think about it, this is exactly the playbook great brand marketing has been running for decades. Every great cause, campaign, and category-defining brand has had a villain, and they had villains long before this experiment was run.
As you review your account – specifically those Group #1 ads – note how many have a protagonist standing alone vs. how many give the audience something to push against. One of the easiest ways to turn an otherwise bland testimonial into a compelling story is to sprinkle in a bit of adversity from an easy-to-hate villain.
The Turn
Everything so far has been about how you can use the Spotlight to influence your prospect/customer.
But, marketers are people too. The same bias is being done to us, by us, every time we open Meta ads manager or a reporting dashboard.
The fact is that marketers commit scope insensitivity against our own data constantly – and the better you are, the more seductive the trap. We overweight the vivid, identifiable, last-click conversion, the green number on the report, the one event we can point the spotlight at and narrate. We go numb to the diffuse, statistical, brand-level effect we can’t precisely pin down.
In the parlance of this issue, the black-and-white ROAS is a prototype. It is the one oiled bird of the media plan, a single, comprehensible instance your brain seizes and prices – all while the vast, real, unrenderable field of brand demand, consideration, halo effects and random externalities generates no feeling at all, because you can’t visualize it.
We marketers price the feeling an up-and-to-the-right number gives us – the warm glow, the moral satisfaction, the sense of having discharged our duty to “know our numbers,” which has almost nothing to do with the birds. The 3P attribution tool that serves up a confident, decimal-pointed number is valuation by prototype, sold back to you as rigor and priced accordingly. The irony is amazing: we spend the morning exploiting prototype bias in our prospects and the afternoon falling for the identical bias in our reporting. The only difference is that in the afternoon, we mistake it for diligence.
I’ve caught myself doing it. I’m sure – if you’re honest with yourself – you have as well. And you’re probably wondering how. How can mathematically-inclined, diligent, thoughtful people succumb to the same biases as everyone else?
Well, Kahneman’s least convenient finding was not that System 1 is impulsive; it was that System 2 is lazy. Its default posture is to endorse System 1 rather than audit it, to ratify the intuition that arrives and assemble a respectable rationale afterward (the impetus for Gerald Zaltman’s famous “We are a nation of rationalizers” quip).
Locking onto a ROAS number is a System 1 activity. System 2, doing what System 2 mostly does, gave it a quick look, deemed it plausible, and signed off. You did not reason your way to that number. You felt your way to it and attached a decimal point.
This is why the issue is not a creative tip. The discipline required to wield the spotlight in your creative is the same discipline required to resist in measurement/account management. It’s near-impossible to reliably do the former while unconsciously failing the later. The key is to refuse to let System 2 rubber-stamp the substitution: demote the spotlight’s output from a decision to an input, then force yourself to do the thing your brain doesn’t want to do.
For creative, that means feeling the pull of the single prototype, trusting that persuasion lives there, then engineering the prototype on purpose – choosing it, specifying it, giving it the obstacle. That’s exponentially more difficult than simply retreating to the floodlight number that feels comprehensive but doesn’t move anyone anywhere good.
Conversely, when it comes to assessing your performance, the opposite is true: feel the allure of the ROAS number for a creative/campaign/platform, recognize that your own System 1 spotlight seized it because it was graspable rather than the full story, then refuse to allow that one number to be the reason you make a change (pause an ad, shift budget, whatever). Instead, investigate it – sit with the ambiguity. You might still end up in the same place. You might not.
The human affective system renders one bird. Your prospect’s does it, your client’s does it, and yours does it.
Your job is to take advantage of your target audience’s System 1, while keeping your own in check. Speak to the prospect’s spotlight in your creative, because it is the only instrument they will use to make a decision. Refuse to accept your own spotlight in the analysis, because it is the instrument most likely to lie to you convincingly. That’s the entire game. It’s so simple it’s near-impossible to win.
Cheers,
Sam
