The Discovery Economy
Being known is no longer the same as being found. People rarely set out to research a brand. They are shown things, in a feed, in a search, in a logistics flow, in an AI answer. The brand's job is to be present where intent surfaces, not where the marketing team is most comfortable.
The language you will use this week
Six terms. Read the term, predict the definition, then flip. Discovery economics has its own vocabulary because the old one (reach, impressions, awareness) was built for a world where you decided when the audience saw you.
Three modes of discovery. Search is declared intent. Algorithmic feeds infer intent from behaviour. AI answer engines synthesise intent into a single response with citations. Each mode rewards different content. A brand designed only for the first usually disappears in the other two.
Original diagram, after the AI-search and platform-ecosystem discussions in Clement (2026), Chapter 5 and the discovery framing of DataReportal (2025).By the end of this chapter you will be able to
- Distinguish awareness, search, algorithmic discovery, and AI-answer discovery as separate modes with separate rewards.
- Map a brand's current discovery footprint across the three modes and name where it is absent.
- Explain why closed-loop ecosystems like Naver, Kakao, and Coupang resist link-graph SEO playbooks.
- Recognise logistics, delivery, and interface design as discovery channels, not as operations.
- Describe the citation economy: how being referenced inside an AI answer differs from ranking first.
- Apply the FOUND rubric to audit a brand's discovery readiness before launch.
4.1 From awareness to discovery
For most of the twentieth century, the goal of marketing was awareness. Buy a billboard, run a TV spot, repeat the message often enough, and eventually a percentage of the audience would remember the brand existed. Awareness preceded interest. Interest preceded action. That sequence no longer matches how people decide.
The launch of Google in 1998 fundamentally changed discovery. Marketing shifted from broadcasting messages to appearing at moments of intent. Twenty-five years later, intent itself has fragmented. It now arrives through search bars, scroll feeds, map apps, messaging shortcuts, voice assistants, AI chat, and logistics filters. Each of those surfaces is a discovery channel, and each rewards a different kind of presence.
Being known is no longer the same as being found
A brand can be deeply famous in a category and still be invisible at the moment a customer is deciding. The discovery economy treats presence at the intent surface as the unit of marketing value, not aggregate brand recall.
4.2 Three modes of discovery
The three modes shown in Fig. 4.1 are the dominant intent surfaces a modern brand has to think about. They are not channels in the old sense. They are contracts: each one rewards a different kind of content and punishes the wrong kind quickly.
| Mode | Intent | Rewards | Punishes |
|---|---|---|---|
| Search | Declared, keyword | Relevance, authority, page speed | Thin or duplicated content |
| Algorithmic | Inferred, behavioural | Native format, novelty, re-watch | Recycled, cross-posted material |
| AI Answer | Synthesised, conversational | Clarity, structure, evidence, trust | Marketing copy without facts |
Most strategies fail not by being absent from one mode, but by treating all three the same. A blog post optimised for Google often performs poorly when re-cut for TikTok. A TikTok script rarely lands as an AI citation. Discovery requires translation, not repetition.
Where product discovery starts. For a young segment, the search bar is now barely a third of all discovery starts. Social feeds nearly match it, and AI answer engines, near zero a few years ago, are already material. A brand present on only one surface is missing most of the market.
Illustrative classroom dataset, one segment, 2025. Shares of product discovery sessions by the surface they start on. See Dataset 4.A below.| Starting surface | Share of discovery | Mode | What it rewards |
|---|---|---|---|
| Search engine | 38% | Search | Relevance, authority, speed |
| Social feed | 34% | Algorithmic | Native format, novelty, re-watch |
| Marketplace app | 19% | Closed loop | Reviews, price, delivery promise |
| AI answer engine | 9% | AI answer | Clarity, evidence, citable facts |
Objective two of this chapter is to map your own footprint across the three modes and name where you are absent. The grid below is that map, worked for a small skincare brand. Build your own underneath, and be honest about the empty rows.
| Mode | Where intent arrives | Present today? | Native content we hold | The gap to close |
|---|---|---|---|---|
| Search | Google, Naver search and shopping | Partly | A thin product page, no reviews | Seed reviews, build a Naver Smart Store |
| Algorithmic | TikTok, Reels, YouTube Shorts | No | Cross-posted ads that flop | Make vertical, native, creator-style video |
| AI answer | Perplexity, ChatGPT Search | No | No structured, citable facts | Publish specific, comparable ingredient data |
4.3 Algorithmic discovery
Algorithmic discovery is the largest and least understood mode. The audience does not say what they want; the system infers it from the videos they finish, the posts they save, and the accounts they follow. Behaviour replaces keywords as the input. For brands, this changes what good content looks like. Posts now compete with the audience's own attention history.
The first feed designed around inferred intent
TikTok's For You Page broke the social-graph assumption that defined the previous decade. Users do not need to follow anyone for the system to learn what they want. The model watches behaviour at a sub-second resolution and adjusts after every swipe. For brands, this rewires the rules: paid reach is no longer the way in, and follower count is a weak signal. What matters is whether a single video earns enough completion to be served to a second cohort.
The cultural consequence has been significant. A Korean indie skincare brand can reach a Tokyo skincare audience overnight without translation, without a media buy, and without being known. Algorithmic discovery rewards the artefact, not the brand behind it. That is freeing for new entrants and uncomfortable for incumbents.
Smith, B. (2021). How TikTok reads your mind. The New York Times. nytimes.com. Brand mark used for editorial reference under educational fair use.What algorithmic discovery actually rewards
A brand re-uploads its 30-second TV ad as a vertical TikTok and gets very low completion. Which response best reflects this chapter?
C is correct. Algorithmic discovery is a contract: native format wins, recycled material loses. Cross-posting often signals to the system that the content was not made for it. Section 4.3.
4.4 Closed-loop ecosystems
In Korea, platforms like Naver and Kakao dominate discovery and communication. Search behaves differently. Messaging is commerce-enabled. Ecosystems are closed rather than link-based. Western SEO strategies often fail when copied directly. Global platforms do not erase local behaviour; they sit on top of it.
Discovery, evaluation, and payment without ever leaving the page
A Korean shopper searching for a moisturiser on Naver rarely sees a clean list of brand websites. The result page is a layered surface: Naver Shopping price-comparison tiles, sponsored Smart Store listings, Naver Blog reviews, Knowledge-iN questions, a small map result if the brand has physical retail, and a video panel. Most of the journey happens inside this surface. The link out is the exception, not the default.
For a foreign brand entering Korea, the implication is concrete: maintain a Naver Smart Store, seed credible Naver Blog reviews, register with Naver Place if relevant, and treat Naver Shopping price tiles as a primary creative surface. Doing the Western thing, driving paid traffic to the brand's own .com, leaks most of the discovery value back to Naver competitors who chose to live inside the loop.
Statista. (2025). E-commerce market in South Korea: Naver Shopping share. statista.com. Brand mark used for editorial reference under educational fair use.Closed-loop ecosystems exist outside Korea too. Amazon's product detail page, YouTube's in-platform search, Pinterest's saved-pin journey, even Spotify's "More like this" panel all behave as loops. The diagnostic question is the same in every case: at what point does the user have to leave the surface to act? The shorter that distance, the more discoverable you are.
The six tests of discovery readiness
- We have a defensible presence on at least one surface in each of the three modes (Search, Algorithmic, AI Answer).
- In our primary market, we are present inside the closed-loop platforms our audience uses daily.
- Our content is native to the surface it sits on, not cross-posted from another medium.
- Our logistics, pricing, and delivery filters are themselves discovery surfaces and treated as such.
- We have at least one piece of content explicitly designed to be cited by an AI answer engine.
- We can name the discovery loop we want (see, save, return, buy, repeat) and where it currently leaks.
4.5 Logistics as discovery
One of the most under-recognised shifts in modern marketing is that operations have become a discovery channel. When a delivery promise appears in the search filter as "tomorrow," that promise is doing the same work an ad used to do. The customer's hesitation drops because the friction has been removed up front, not after the click.
Rocket Delivery as a discovery surface
Coupang succeeds not because of aggressive promotion, but because the experience is predictable. Discovery, ordering, payment, delivery (로켓배송 Rocket Delivery), and returns follow the same logic. The Rocket badge inside the search result is, in marketing terms, a credibility cue that performs the same job a creative campaign performs elsewhere. The system reduces uncertainty in advance. That reduction in effort builds trust faster than advertising ever could.
For brands selling on Coupang, the lesson is concrete: opting into Rocket fulfilment is not a logistics decision; it is a discovery decision. The brand becomes findable in a way that a brand on standard delivery quietly is not. Logistics is now creative.
Coupang. (2024). About Coupang: Rocket Delivery. aboutcoupang.com. Brand mark used for editorial reference under educational fair use.4.6 The citation economy
Traditional SEO assumed people searched with keywords and that visibility depended primarily on ranking. That logic is weakening. AI-mediated search systems increasingly interpret questions, summarise answers, and recommend sources directly. Users receive synthesised responses rather than lists of links.
Being found is no longer only about ranking high. It is about being trusted enough to be referenced. Chapter 4 · 4.6
The answer engine that made citations the product
Perplexity's interface treats the cited source as a first-class object. Each answer shows the small set of pages it leaned on, numbered inline. The economic consequence is that the discovery payoff has shifted from clicks to citations. A brand whose product page is the third citation in a comparison answer is, for that question, the brand the user trusts, even if they never click through.
The defensive implication for marketers is uncomfortable: most existing SEO content was written to win the click, not the citation. Citation-friendly content tends to be specific, structured, and willing to compare honestly. Marketing copy without facts gets ignored. Optimisation without substance is fragile.
Perplexity. (2025). How Perplexity sources answers. perplexity.ai. Brand mark used for editorial reference under educational fair use.Conversation as the new query
OpenAI's web-connected ChatGPT shifted a meaningful slice of "how do I" and "compare X vs Y" queries from search results into multi-turn chat. The user does not refine a query string; they refine the conversation. For brands, this means the discoverable unit is no longer a page; it is a fact stated clearly enough that the model picks it up and repeats it correctly.
OpenAI. (2024). Introducing ChatGPT Search. openai.com. Brand mark used for editorial reference under educational fair use.What citation-friendly content looks like
A coffee brand wants its origin sourcing page to be cited by AI answer engines. Which change most increases the chance of citation?
C is correct. Answer engines favour content that is directly addressable, specific, and structured. Marketing copy without facts gets ignored. Section 4.6.
4.7 Designing for discovery
Discovery is no longer a metric that gets added at the end of a plan. It is a design constraint that shapes the product, the content, and the operations from the first decision. Three working principles follow from this chapter.
The collapse of the funnel. Discovery, consideration, and purchase that once took weeks can now close in one session, so the old staged funnel no longer describes how attention converts.
Original diagram, after the discovery-economy argument in Clement (2026), Chapter 2, and DataReportal (2025).The discovery loop. See, save, return, buy, repeat. The two most common leaks are the link out (asking the user to leave the surface they discovered you on) and checkout friction (forcing a new account at the moment of action). Naver Shopping and Coupang win by closing both.
Original diagram, after the closed-loop ecosystem framing in Clement (2026), Chapter 2.7 and DataReportal Korea (2025).| Principle | What it means | What it asks the team to give up |
|---|---|---|
| Native, not repurposed | Build for the surface where the intent arrives | The efficiency of cross-posting |
| Inside the loop | Live where the audience already transacts | The pride of "our own website is the destination" |
| Cited, not just ranked | Earn the right to be referenced | SEO content written only for clicks |
The chapter argument, in one sentence
Which statement best summarises Chapter 4?
D. The chapter is one argument with three faces: intent arrives through different surfaces, each surface rewards different content, and operations are now one of those surfaces.
· Worked example · The FOUND rubric
Before a brand calls itself "discoverable," run it through FOUND. Five quick checks, scored 1 to 5, totalled out of 25. Below 15 means the brand is famous in its own meeting room and invisible in the moments of intent that matter. The worked card audits a small Seoul roastery preparing to enter Tokyo. The blank card is for one of your own.
FOUND: Findability, Obviousness, Usefulness, Novelty, Density. Score each, then total.
Now do one yourself. Pick a brand entering a new market, or your own project, and run it through FOUND. If you score below 15, write one sentence explaining what would have to change to push the lowest-scoring letter up by one.
Tick the score, write one line of justification, then total.
Test yourself before discussion
True or false. Answer first, then read the explanation. If you miss more than one, revisit the section noted before continuing to Chapter 5.
-
In the discovery economy, being a famous brand guarantees you will be present at the moment the audience is deciding.
Being known is no longer the same as being found. Brand fame can coexist with absence from the intent surface that matters. Section 4.1.
-
Algorithmic discovery rewards artefacts native to the surface; cross-posted TV-style ads tend to underperform.
Recycled content reads as legacy advertising and is downweighted quickly. Native, novel, re-watchable wins. Section 4.3.
-
Korean ecosystems like Naver and Coupang behave as closed loops where most of the user journey happens inside one app.
Link-graph SEO assumes the audience leaves the platform to act. In Korea, they usually do not. Section 4.4.
-
Logistics and delivery filters can function as discovery channels in their own right.
A "delivery by tomorrow" filter is doing the work an ad used to do. Operations is now creative. Section 4.5.
-
AI answer engines tend to favour content that states specific, structured facts over content that is purely emotive marketing copy.
Citation requires substance the model can lift and attribute. Optimisation without substance is fragile. Section 4.6.
· Reflection
Discussion questions
- When was the last time you discovered a brand you ended up buying? Which intent surface did it appear on?
- Why do you think repurposed TV-style ads tend to underperform on algorithmic feeds, even when production quality is high?
- How does Coupang's Rocket Delivery function as a discovery channel rather than a logistics service?
- For a foreign brand entering Korea, what is the cost of ignoring Naver Shopping and Naver Blog?
- Has an AI answer engine ever decided a purchase for you? If yes, what made the cited source feel trustworthy?
- Is it ethical for a brand to engineer content specifically to be cited by an AI? Where is the line between clarity and gaming?
- Which of the three modes (search, algorithmic, AI answer) is the brand you study weakest in, and why?
- If discovery now happens inside closed-loop ecosystems, what is the long-term value of a brand's own website?
- How would your discovery strategy differ for a luxury brand versus a discount challenger?
- Will AI answer engines eventually compress search and discovery into a single conversational surface? What would brands have to give up to win there?
· Chapter summary
Discovery has fragmented into three intent surfaces that behave like separate contracts: search rewards declared intent and authority, algorithmic feeds reward inferred intent and native artefacts, and AI answer engines reward synthesised intent and citable substance. A brand designed only for one of those surfaces is increasingly invisible on the others. Discovery requires translation, not repetition.
In Korea, closed-loop ecosystems compress the journey further: Naver Shopping and Coupang resolve discovery, evaluation, payment, and delivery inside a single surface, and link-graph playbooks underperform there. Logistics has become creative, because a delivery promise inside a filter does the work an ad used to do. And the rise of citation-driven AI answers means visibility is no longer about ranking first; it is about being trusted enough to be referenced. The FOUND rubric (Findability, Obviousness, Usefulness, Novelty, Density) is one way to keep that judgement defensible before launch.
· References used in this chapter
The full bibliography is on the References page.
- Clement, M. (2026). Digital marketing: An integrated, project-based approach (3rd ed.). Independent practitioner publication. Chapters 2 and 5.
- Coupang. (2024). About Coupang: Rocket Delivery. aboutcoupang.com
- DataReportal. (2025). Digital 2026: Global Overview Report. datareportal.com
- Harvard Business Review. (2021). How Amazon uses data to drive a culture of experimentation. hbr.org
- OpenAI. (2024). Introducing ChatGPT Search. openai.com
- Perplexity. (2025). How Perplexity sources answers. perplexity.ai
- Smith, B. (2021). How TikTok reads your mind. The New York Times. nytimes.com
- Statista. (2025). E-commerce market in South Korea. statista.com
- Think with Google. Moments of intent. thinkwithgoogle.com