FROM RACE-TO-THE-BOTTOM MEDIA BUYS TO SPY-DRIVEN CREATIVE — WHAT AFFILIATES CAN LEARN FROM AB INBEV

For years, performance marketers could build an advantage by finding cheaper traffic than their competitors. A lower-cost click, a more efficient placement, or a well-timed bid adjustment could make an otherwise ordinary campaign profitable.

But buying traffic cheaply is not the same as building a durable competitive advantage.

As advertising platforms become more automated, marketers need to pay closer attention to the creative, offer, and post-click experience that turn impressions into customers. Competitive ad intelligence can help reveal what other advertisers are testing, while AI can help transform those observations into original concepts and structured experiments.

That creates a different opportunity for affiliates.

Instead of treating ad research as a swipe file and creative production as a separate task, marketers can build a continuous process: observe the market, identify creative patterns, develop original variations, test them, and scale only what works.

The enterprise examples of AB InBev, Kroger, and Unilever offer useful lessons in treating creative as a measurable performance asset. Affiliates do not need the same enterprise infrastructure to adopt the underlying discipline.

Performance marketing once rewarded buyers who could consistently find inexpensive traffic.

Native advertising offered opportunities through inexpensive placements and curiosity-driven headlines. Push campaigns often competed through attention-grabbing notification copy. Early TikTok advertising rewarded brands that could quickly capitalize on emerging formats and trends.

That did not mean creative was unimportant. It meant some advertisers could compensate for weak creative with favorable traffic economics.

That approach becomes harder to sustain as platforms automate more of the buying process.

Modern advertising systems use a range of signals to determine delivery, bidding, and predicted outcomes. The price and performance of an impression can vary with the advertiser's objectives, auction conditions, audience, placement, and platform optimization. Consequently, there is no universal rule that competitors pay roughly the same price for every comparable impression.

The more useful question is whether an advertiser can turn its available traffic into better business outcomes.

Why creative and media need to work together

Competitive-intelligence approaches are increasingly focused on understanding the relationship between creative activity and media conditions.

For example, AdExchanger's coverage of competitive signals inside social ad auctions explores how auction-related observations can reveal patterns that deserve further investigation.

Those observations may include changes in visible creative activity, placement emphasis, geographic distribution, or available auction metrics.

They can help marketers form hypotheses about how competitors are approaching the market. They do not automatically reveal a competitor's true acquisition costs, profitability, or internal optimization strategy.

The creative side of the equation is equally important.

A useful example comes from the Unilever creator initiative discussed in Search Engine Journal's coverage of AI-generated creator networks. The article describes an approach connecting creative-effectiveness analysis with media execution, illustrating how creative evaluation can inform decisions about which assets receive attention and investment.

The underlying principle is that creative should not be evaluated in isolation from campaign results.

If an advertisement attracts attention but produces poor-quality traffic, its apparent creative success may not translate into commercial value. Conversely, an asset with less dramatic engagement may generate more qualified leads or customers.

The shift toward continuous creative optimization

AI-assisted advertising workflows are making it easier to generate, compare, and iterate on creative variations.

MarTech's discussion of AI-native advertising explores the movement toward more continuous creative optimization.

For performance marketers, the opportunity is to shorten the feedback loop between:

• What the market is showing.
• What the team creates.
• What the advertising platform delivers.
• What users do after clicking.
• What the business ultimately earns.

This is particularly relevant across Native, Push, and TikTok, where the creative format and user context differ substantially.

On TikTok, the opening moments and overall storytelling structure can influence whether viewers continue watching and engaging. In Native advertising, the relationship between the headline, image, and landing-page experience matters. In Push, the limited space makes clarity and relevance essential.

The cheapest click is not necessarily the best click. A campaign must also consider whether the visitor is qualified, whether the landing page delivers on the advertisement's promise, and whether the resulting conversion supports the economics of the offer.

There is another part of the funnel that marketers should not overlook: the post-click experience.

Research by Unbounce and Ascend2, discussed in MarTech's article on paid media ROI investment, highlights the importance of examining what happens after an advertisement receives attention. The article's reported findings should be checked against the original research methodology before being used as precise industry-wide benchmarks.

A strong creative cannot compensate indefinitely for a confusing landing page, a mismatched offer, or a slow conversion process.

The strategic shift is therefore not simply from media arbitrage to creative arbitrage. It is from isolated traffic-cost optimization toward a connected system that evaluates creative, media, and post-click performance together.

Large advertisers often approach creative differently from smaller performance teams.

Rather than treating an advertisement as a disposable campaign asset, they can evaluate creative as something that contributes directly to business results.

AB InBev is a useful strategic reference for this mindset. Its relevance here is the broader idea of connecting creative decisions with measurable commercial outcomes—not an assumption that every affiliate should copy the company's exact operating model.

Kroger offers a more specific example of creative measurement.

According to Marketing Dive's coverage of Kroger's AI-driven creative analysis, an initiative examined more than 1,900 video and image assets across Meta and DV360 to identify creative attributes associated with e-commerce conversion.

The important lesson is that creative can be analyzed systematically.

Instead of relying entirely on personal preference, marketers can examine the relationship between creative attributes and measured results.

Build a creative performance record

Affiliates can apply a smaller version of this approach without enterprise-scale infrastructure.

Start by recording the characteristics of each advertisement:

• Opening hook and headline.
• Main visual or opening shot.
• Format and duration.
• Offer and value proposition.
• Audience or placement context.
• Landing-page destination.
• CTR, conversion rate, earnings per click, and other relevant results.

Over time, this creates a useful history of which creative approaches perform under specific conditions.

For example, a certain headline structure may attract clicks on a Native network but fail to produce qualified conversions. A short video demonstration may generate fewer clicks but lead to better conversion rates on TikTok.

The point is not to identify one universal winning formula. It is to understand which creative characteristics work for a particular offer, traffic source, and audience.

Treat creative as an investment, not an afterthought

Kroger's predictive-scoring example illustrates how AI can help evaluate creative assets before or during campaign activity.

Affiliates can borrow the principle by using historical performance data to develop their own evaluation criteria.

A simple creative record might include:

These records can help a team compare new ideas against its own historical results.

However, a predictive score should not be mistaken for a guaranteed conversion forecast. It is only as reliable as the data, methodology, and validation behind it.

Build a creative balance sheet

A useful mental model is to treat creative assets as a portfolio of reusable knowledge.

The team should know which hooks, formats, offers, and visual approaches have performed well, where they performed well, and under what conditions.

That creates three practical shifts:

1. Allocate resources to creative development. Do not spend the entire testing budget on placements and bids while leaving creative development underfunded. The appropriate allocation depends on the offer, channel, margins, and current bottleneck.
2. Document creative performance. Track thumbnails, headlines, scripts, and landing-page variations as identifiable assets with measurable results.
3. Connect creative results to buying decisions. Use reliable first-party performance data to inform which assets deserve further testing or additional budget.

TikTok provides another example of why a steady supply of creative matters.

The HubSpot TikTok campaign guide discusses the value of creative variety, including brand assets and creator content. The broader lesson is that campaigns should not depend on one isolated hero video.

Automation can help distribute and optimize creative, but the advertiser still needs to provide differentiated material that fits the product, audience, and campaign objective.

The enterprise lesson for affiliates is straightforward: creative should be managed, measured, and improved like any other important performance asset.

Competitive ad research is most useful when it leads to better questions and more structured creative experiments.

A swipe file can show what competitors are advertising. A more developed research process identifies the patterns behind those advertisements and turns them into original ideas that can be tested.

That is the foundation of spy-driven creative.

The objective is not to copy another advertiser's creative or assume that every long-running advertisement is profitable. It is to understand what is visible in the market, identify potentially useful patterns, and investigate whether those patterns can work for your own offer.

Look beyond the screenshot

A competitor advertisement should be studied as more than an image or video.

When researching an ad, record details such as:

• Opening move: What appears in the first one to three seconds?
• Dominant emotion: Does the creative use curiosity, aspiration, urgency, reassurance, or fear of loss?
• Proof device: Does it use a demonstration, testimonial, comparison, or social proof?
• Offer framing: Is the message centered on a discount, bundle, trial, or other proposition?
• Format: Is it creator-style video, product demonstration, animation, static image, or editorial-style content?
• Destination: What kind of landing page follows the click?

These details help reveal how an advertisement is structured.

Some competitive-intelligence platforms also provide information about where and when advertisements appear. For example, AdExchanger's coverage of auction-level competitive signals discusses how market and auction observations can add context to competitive analysis.

Creative longevity can be one useful observation, but it must be interpreted carefully. An advertisement remaining visible for a long period does not establish that it is profitable or even continuously active. It may be reused, retested, or supported for reasons that are not visible externally.

Convert observations into reusable creative patterns

The next step is to abstract individual advertisements into templates.

For example, a vertical video might follow this structure:

• Creator-style opening.
• A surprising but supportable product observation.
• A demonstration or relevant proof point.
• A clear explanation of the offer.
• A direct call to action.

A Native advertisement might follow a different structure:

• Relevant headline.
• Editorial-style introduction.
• Problem explanation.
• Evidence or product demonstration.
• Offer and call to action.

These are creative frameworks, not formulas guaranteed to work.

AI can help turn the patterns into variations. Instead of prompting an AI tool to write a generic advertisement, provide the structure, the brand's approved claims, the offer, the audience context, and the platform requirements.

MarTech's article on sustaining creative testing without adding headcount provides context for the importance of maintaining a steady supply of strategic, high-quality creative inputs.

Build a five-step creative intelligence loop

1. Ingest competitive observations

Use advertising research tools and supported ad libraries to collect relevant examples in your category.

Record the creative, visible placement information, offer, destination, and date of observation. Keep the source and limitations of the information clear.

2. Abstract the patterns

Organize individual advertisements into reusable creative structures.

Examples include:

• Creator-style product demonstration.
• Problem-and-solution storytelling.
• Curiosity-led Native headline.
• Customer testimonial.
• Comparison-based offer presentation.

3. Generate original variations

Use AI to develop new concepts based on the structure rather than copying the original execution.

For example, ask for multiple original headlines that use a curiosity-led structure while remaining accurate, relevant to your product, and compliant with the advertising platform's rules.

4. Route concepts into controlled tests

Prepare a structured testing queue instead of publishing every generated variation at once.

Test creative in a way that makes it possible to compare results. Keep the offer, audience, placement, and other important variables sufficiently controlled to understand what may have influenced performance.

5. Feed first-party results back into the system

Once results are available, tag the creative by hook, format, offer, and other relevant attributes.

Use those results to inform the next generation of ideas. The team's own performance data is more valuable for deciding what to scale than an assumption based on competitor visibility alone.

TikTok is a particularly useful environment for this approach because creative variety and platform-native execution matter. The HubSpot TikTok guide discusses the importance of a diverse creative supply.

The result is a continuous idea pipeline: market research helps generate hypotheses, AI helps develop variations, and controlled testing determines which ideas deserve further investment.

A creative-testing system needs more than a large number of advertisements.

It needs a repeatable method for developing hooks, generating meaningful variations, and evaluating results.

AI can help accelerate production, but the quality of the process still depends on the strategy, source information, and measurement rules behind it.

Start with the hook

The opening moments of an advertisement can strongly influence whether a user continues watching, reading, or engaging.

On TikTok, this may be the first video shot or spoken line. In Native advertising, it may be the headline and thumbnail. In Push advertising, it may be the notification title and message.

Instead of brainstorming randomly, use competitive research to identify recurring approaches worth exploring.

For example:

• A question that addresses a specific problem.
• A demonstration that reveals a product benefit.
• A surprising but verifiable observation.
• A comparison that helps explain a choice.
• A customer experience that illustrates a real use case.

AI can help develop multiple versions of each concept.

A useful workflow is to:

1. Feed an AI tool your documented creative patterns.
2. Ask it to identify the hook structures and messaging approaches.
3. Generate variations tailored to the offer and platform.
4. Check every version against approved claims and compliance requirements.
5. Select a manageable group for testing.

The output should be original and accurate—not a collection of exaggerated claims or near-identical copies.

Build families of creative variations

Once a hook has been selected, develop variations that test different creative elements.

A single concept might produce:

• Different opening lines.
• Different visual demonstrations.
• Different creator or product presentations.
• Different thumbnail compositions.
• Different calls to action.
• Different lengths or storytelling sequences.

The goal is to learn which elements contribute to results.

For example, a TikTok concept might be developed into a face-to-camera version, a product demonstration, and a creator-style testimonial. A Native concept might be tested with different image-and-headline combinations while keeping the destination experience consistent.

For Push campaigns, test clear, relevant notification language rather than deceptive system alerts or misleading claims.

Use platform-aware creative constraints

Each advertising environment has its own creative requirements and user expectations.

• Native: Headlines and images should create relevant curiosity and lead naturally into the destination content.
• Push: Short copy should communicate a clear, accurate reason to engage.
• TikTok: Vertical-first storytelling should feel natural to the platform and deliver on the opening promise.
• Pops: The landing experience is especially important because the initial format provides limited creative context.

AI prompts should include the format, audience, offer, tone, length requirements, and compliance restrictions.

For sensitive verticals, avoid unsupported medical, financial, or other high-stakes claims. Do not use deceptive countdowns, fabricated testimonials, misleading system notifications, or prohibited before-and-after presentations.

Build a repeatable testing cycle

A simple creative-testing process can operate in three stages.

Round 1 — Directional testing

Launch a manageable group of concepts with controlled budgets. The objective is to identify early signals and eliminate clearly unsuitable creative.

Round 2 — Focused iteration

Develop variations around the most promising concepts. Change one or a small number of meaningful creative elements so the team can learn what contributes to the result.

Round 3 — Validation and expansion

Give the strongest candidates additional testing and evaluate them against business outcomes. Promote only the concepts that meet the team's performance and compliance requirements.

The exact number of variations and budget per round should depend on the campaign's traffic volume, economics, and ability to measure reliable differences. There is no universal requirement to launch 20–40 assets or scale the top three to five.

Connect creative signals to real outcomes

A creative-testing machine becomes more useful when its data goes beyond clicks.

Where supported, connect campaign measurement to conversion events and CRM outcomes. Evaluate whether an advertisement produces qualified leads, purchases, revenue, and acceptable acquisition costs.

For example, HubSpot's TikTok campaign guide discusses the value of connecting campaign activity with broader marketing data.

The exact integration and event-sharing capabilities depend on the platforms and tools being used. Confirm the current setup before assuming that CRM lifecycle data automatically flows back into an advertising platform.

The same principle applies to creative scoring. Predictive evaluation can help prioritize assets, but it should be validated against actual campaign results.

The goal is not to produce the largest possible number of advertisements. It is to create a system where every round of testing improves the quality of the next decision.

Each advertising channel has a different format, audience context, and post-click expectation.

The same competitive research process can inform all four, but the resulting creative and landing-page experience should be adapted to the channel.

The common principle is simple: understand what the market is communicating, develop an original concept for your offer, and make sure the destination delivers on the advertisement's promise.

TikTok — Use creative to communicate who the offer is for

TikTok creative does more than attract attention. The visuals, language, creator, and opening message can help determine which users recognize the content as relevant.

MarTech's discussion of creative as the new targeting explores how creative can function as an important signal in automated advertising environments.

When researching a competitor's TikTok ad, examine:

• The first few seconds and opening statement.
• The creator's presentation and setting.
• The problem or benefit highlighted.
• The product demonstration or supporting proof.
• The audience the message appears designed to attract.
• The call to action and destination experience.

Then adapt the underlying concept to your own offer and customer knowledge.

For example, a product demonstration may be more relevant to an audience that needs to understand how something works, while a testimonial-led concept may be more appropriate when customers need reassurance.

TikTok's automation tools can support campaign delivery, but marketers still need accurate conversion data, suitable creative, and a clear measurement plan.

Native — Study both the advertisement and the landing page

Native advertising requires attention to the full journey.

A headline and thumbnail may generate curiosity, but the landing page must continue the same message without misleading the visitor.

When researching Native campaigns, examine both the visible advertisement and the destination.

Useful details include:

• The headline and image relationship.
• The angle used to introduce the topic.
• The first paragraph and opening promise.
• The order in which information is presented.
• The placement of the first call to action.
• The use of demonstrations, comparisons, or supporting evidence.

You can then develop an original landing-page structure that suits your offer and follows the same broad user-intent progression.

Avoid copying another advertiser's branding, protected creative, or unsupported claims. A competitor's page architecture is a research reference, not permission to reproduce its content.

Push — Make the message clear and relevant

Push advertising has a limited creative canvas, so the title, message, and icon need to work together.

Research can help identify common messaging structures in a category, including questions, benefit-led messages, and timely information.

But attention should not come at the expense of trust.

Avoid fake operating-system alerts, fabricated security warnings, misleading personal notifications, or other formats that disguise advertising as an authentic system message.

A good Push campaign should make the nature of the message clear and give users an accurate reason to click.

The landing page should then qualify the visitor and deliver the information or offer promised in the notification.

Pops — Treat the landing experience as the main conversion environment

Pop advertising provides a different challenge because the initial ad format may offer little room for a detailed creative message.

The landing page therefore carries much of the responsibility for explaining the offer and helping visitors decide what to do next.

When researching pop campaigns, examine:

• How quickly the page communicates its purpose.
• Whether the offer is clear above the fold.
• How the page introduces its main benefit.
• Whether the next action is easy to understand.
• Whether the page works properly on mobile devices.
• How the page handles consent, privacy, and required disclosures.

Interactive elements such as quizzes or multi-step forms can be useful when they genuinely help visitors find relevant information. They should not be used to disguise the offer, manufacture urgency, or collect information without a clear purpose.

Connect every channel to the post-click experience

Creative performance should not be judged only by the initial click.

A TikTok user may expect a story or demonstration that continues the tone of the video. A Native visitor may expect a helpful article or editorial-style explanation. A Push visitor needs a clear destination that matches the notification. A pop visitor needs an immediately understandable landing experience.

The message, offer, and destination should remain consistent.

This is where competitive research and first-party measurement work together. Competitor examples can suggest how advertisers structure their journeys, but your own conversion data determines whether the journey is effective.

For instance, Marketing Dive's coverage of Kroger's AI-powered creative analysis offers an example of connecting creative attributes with e-commerce outcomes.

Affiliates can borrow that measurement mindset by tracking not only the advertisement but also the quality of the traffic and the resulting business outcome.

The most useful creative intelligence is not simply a list of advertisements that appear frequently. It is an organized understanding of which ideas deserve testing, which tests produce meaningful results, and which post-click experiences turn attention into value.

The next performance advantage is not simply about finding another inexpensive traffic source.

It comes from understanding the market, developing stronger creative, testing ideas systematically, and improving the experience after the click.

The examples from AB InBev, Kroger, and Unilever point toward a broader principle: creative deserves the same strategic attention as media buying and measurement.

Affiliates can put that principle into practice without enterprise-level infrastructure:

1. Research: Use competitive ad intelligence to identify visible creative patterns and market changes.
2. Structure: Turn useful observations into reusable creative frameworks.
3. Generate: Use AI to develop original, brand-appropriate variations.
4. Test: Evaluate concepts through controlled experiments.
5. Measure: Judge performance using conversion quality and business outcomes, not clicks alone.
6. Improve: Feed validated results into the next round of creative development.

This is what spy-driven creative should mean: not copying competitors, but learning from the market and using that knowledge to build a better testing system.

Stop competing only on the cost of traffic. Build a creative process that gets smarter with every test.